Post any screenshot,picture of anything that you are working on with A8 , it can be anything like a model, a 2D level design, a texture, a game etc ...
(It's a standard thread in lot of 3D artist and some 3D engine forums, i just wanted 3DGS to have one, it can be interesting)
The screenshot is made on a Mac running OS X 10.IMNOTALLOWEDTOTALKABOUTIT, but it compiles with Gstudio as well (beside of the stuff in the LFUnsupported folder). Just in case screenshots of code are allowed too =/
@JustSid: yes code screenshot allowed too, or tools screenshot also It can be 3D Work, 2D work, code, 3D game WIP , music, anything realated to a 3D project with A8 !
@Progger : i really like the game test you are making, reminds me some really cool arcade games that was made some ages ago !
Just took the shot yesterday... Been working on the lighting...
Something a bit different from my usual work...
It's a model of the "Consciousness Grid" (Done in 3dsMax for a new type of game.)
You know... The grid that envelopes our planet...
The geometry is very compicated but at the same time, very simple... Eeach sphere crosses its adjacent spheres and creates the "Flower of Life" pattern in 3d...
In its geometry, is many, or as some say... all patterns...
@Rondidon : Just one thing : put ShowREAL instead of ShowREEL Great video, i saw in the video some 3D character manipulation in WED, perhaps yu could have used another lot more detailled character model than a basic one ? Anyway great.
because it looks strange there too. rippled glass in a clean roomlike that doesnt really match. I ithink it would look cool with a much less dense set of ripples. (I still like them, i just think they need a very small frequency, like 3 or 4 ripples across) I've always thought that about the windows in portico tho. never mentioned it before this.
Mr. Schwarz... I have never seen horizontal ripples, only vertical, and I only noticed them in one shot. That's why it seems a bit odd... As I said, very nice work..
Actually, it's not a bathroom. Have you really never heard of Valve's "Portal"? Michael Schwarz's game is heavily inspired by Portal -- perhaps even a parody (I'm not sure on the details), and that's why he has the rippled glass in his game.
TerraSame, I can understand if you've never played Portal, but have you seriously never even heard of it? What, have you been living under a rock for the past three years?
Dark Inferno, You are right and I agree... My chat posts have been deleted, as much as I was allowed... I encourage others to do the same... This post will be deleted shortly also... Regards
i still cant see why one would remake a game like this, idk, whats to gain ?
Actually, making such a game makes a whole lot of sense to me. That way, you can find out more about game mechanics, how all that neat stuff is coded and get to know what it is that makes the original you're trying to rebuild for yourself such a great game. So there's a ton of knowledge to gain, and IMHO that's worth a lot.
that and sometimes its fun to make your own version of a game. i remember when i was a kid redrawing mario levels. by the time i was 14 i remade 2 full levels of my own with exact mario bros 3 physics and images. they were fun as hell.
You have a third party client that acts as run loop for the server and updates it over the wire, this has several disadvantages: First, your system isn't autarkic anymore and you gained a lot of "single points of failure". The second problem is latency, which you always have to deal in an network environment, but yours is unneeded. The third problem is concurrency and this is also the thing you will feel in labor tests (the other two points are most likely only noticeable in a real world environment). You will face many concurrency issues once you are logged in with your client, this goes from wrong updated labels to cheat possibilities for users.
The second big mistake is the choice of PHP, while PHP is great for dynamic websites, its unbelievable slow and costs too many performance. What you do is something like "hey everyone, please DDoS my server. kthxbye" and it won't take many clients to take you server down.
The third big mistake is scalability, one day, you just can't scale up anymore without facing really bad bugs, mostly due to the external run loop. You also might want to reconsider the choice of MySQL, while it is great and well tested (someone ran into a possible problem long before you), PostgressSQL for example might be well more suited.
There are many many more mistakes in your design, even if its just a rough "lets do it this way and look how we implement it in detail on the fly" (which is also a big mistake), but I'm not going to list all of them for a project that will be canceled in a few weeks so if you want to here more or a actual proposal for a new infrastructure with protocol, write me a PM and tell me what you are ready to pay for it.
In the meantime I have to stay with my conclusion: Looks like the worst and inefficient "protocol" ever.
Starting new projects is easy... but finishing them - is not your strength. Tell me: Why should I bother giving you advice when the project dies in a few weeks anyway? Beside that, I posted numerous advices about multiplayer and programming in general. I also wrote a pretty big library which uses lots of advanced stuff so what do you want to tell me? That I'm just saying things without showing that I can do them? Isn't it more like YOU are the one who always says "I do this epic game now" without ever finishing one of them?
@Devon : i like a lot your Zelda remake Indeed i think you'll gain big experience in programming and making cartoon stylised textures and level design ! Keep it up !
I am only making this game as a learning corner - and mostly for fun. As a child, I enjoyed Zelda3: A link to the past and always wanted to play another game like it... but that's impossible to find. So why not make it myself?
Mainly @ Joozey: All the characters will be in a chibi style, inspired by the original game - Thats why link looks a little fat.. hes chibi style XD
I will not be selling this game at all - its simply just a fan game, that will be free to play when it is finished; Like i mentioned its to learn how certain aspects can be achieved.. then I will make my own game.
Just wondering... has Sid ever finished a game? He sounded like it...
Yes, he actually did... He even finished three games and some serious applications for this new shiny iOS device. The first one was the famous gulli:adventure (I guess you can still find it on gulli.com) with which he won an iPhone 3G and a short trip to vienna. The second one was 5mb robot (can be found in the AUM resources) The third one is the successful and infamous iAtoms which you can currently get for free on the App Store (to be honest, I just wrote the multiplayer part and the graphics engine for the game, Slin wrote the single player part)
And two of my recent serious applications are Moneybag and Trafiken which both can be found on the App Store as well (the second one is a pretty cool swedish traffic app).
Oh and then I wrote a small Open Source kernel called NANOS ( https://github.com/JustSid/NANOS ) and the best 2D engine for iOS devices (its still! the only one with shader support ( https://github.com/JustSid/Void-Dead ). And I still have this pretty unsuccessful (for unknown reasons) library called Lite Foundation. But all of these aren't really games and there is an even greater list of finished small projects... Oh and I have a list of abandoned projects as well, but not as long as the one from Esper: A Minecraft recipe browser for iOS (Notch told me to stop it. Meh), Die Händler (A7 game, you can find it here I guess), Miltun (A5 game) and iAtoms v2.
Well then, congrats for finishing these. I still dislike the way you try to make fun of him. If you think you're in a position allowing you to give advice, do so. If not then your posting is a waste both of your and our time.
Sorry ^^ We'll stop. Sorry Sid. To get back on topic, I'm working on: (Tadaaaa!!) fixing CSiS.
Almost all bugs either don't happen on my setup or they are ones that only happen every 30 minutes or so. That means I have CSiS running all the time, and every 30 minutes (on average) I get a tiny bit of information about one crash. This sucks... I seriously need to think about a way to get more crashes.
In the meantime, we are adding a damage effect when health < 30:
@Germanunkol: Did you use Acknex wait(x) ? If not, you could try to implement a logger for detailed runtime information and a stacktrace on crash(last one isnt possible if you use Acknex wait(x) feature)
Rakscha, sadly, I use wait(1) all over the place, yes... I'm hoping LiteEditor's debugging functions will help. Is there another way to get acknex to render?
Good news though: We've fixed a few of the more pressing bugs today. Not bug free yet, but much, much closer to it.
Realspawn: well done on the models. But a normalmap wouldn't hurt. Why not make a highpoly version and then bake normals? I think especially the barrels would look much better. Also, what about skins?
It is triggered everyframe. This way you'll have to implement your own entity management if you dont want to use waits, but it allows clean stacktraces from your on_frame event down to the function that failed.
mh..just noticed there's a on_message for the winmessages to, nice^^
Hm well, I would have had to know this 4 years ago... ^^ I don't think I'm going to redesign the entire game for this now. It might be very useful, but it's not really practical at this stage into the game. Also, I believe to have fixed most of the things. I just found on_ent_remove... I find this one very interesting as well... ^^
Ok, I apologize - I don't have the device here yet, it is a procedural pattern. The red-orange-yellow-green color change indicates low-to-high heart variability, so to say, if everything is green, you are in an absolutely relaxed state.
The circles are target circles. The theory is: if the user remains his stress/relaxation state constantly (by breathing, meditating, etc.), the light-worm is circling. And the better your heart variability is, the bigger gets the circle. So, the goal is to "draw" bigger circles with your heart then the target circles. 256 is just my initialization radius ;-)
This project is no replacement, it is just another one in development for that device. "The Descent" reached now milestone 1 and when this here is done, we proceed to the next milestone.
@HeelX A surival horror prototype using this concept would be awesome. The only lightsource would be the one your character carries and it´s color and intensity changes the way you implemented it in your firefly project. And only hiding or a lot of light could save you from getting killed through enemies...
i ran some tests to fake GI using the last shadeC beta, basically each light was two light sources, its certainly not great but it beats the direct lighting by miles ... as for speed, wasnt bad, plus my pc sucks so i assume it acceptable for most
I am currently finishing G.R.U.N.T.S. and working on three other game related projects, two music related projects, two video related projects and the ongoing project of having a life with a fiancee and friends. i dont sleep much lately...
top: direct lighting bottom: secondary non shadow light placed to illuminate previous shadows
ahm, at the moment i was just testing the possibility so all i did was the other light at the same position as the first, this ofcourse means that this light would shine through walls but i think it showed me that correct placement could achieve something good
also i think boh_havoc has GI planned, its already in but commented out, havent heard from him but it sounds good
@darkinferno: Rendering a second light is quite inefficient since the same effect could be achieved with one or two extra lines in the 'shadowed light'-shader, but it looks quite good. You should request this as a feature for ShadeC.
Well, I'm currently working on the Lite Foundation threading stuff (again) to make it even better when it will be available with the next LF update. Prepare for delicious code.
@darkinferno: Rendering a second light is quite inefficient since the same effect could be achieved with one or two extra lines in the 'shadowed light'-shader, but it looks quite good. You should request this as a feature for ShadeC.
i know and GI is expected in shadeC, i was just running tests
Most time sits in the tile editor which now features a versatile layering system and soon tile events coupling. This game is supposed to be top down, I just suck at drawing proper perspectives .
Update on InnerFlow. You may notice, that the visuals changed for some reasons. Everything is rendered in A8. In realtime. I didn't photoshopped it, it is pure A8 awesomeness. And hell - I love the HDR shader! Beware, bubbles and lightbeams ahead
P.S: the number in the upper right corner is not the fps. It's an indicator how well your heart performs.
Yesterday I've had a look at Slin's deferred lighting and got an idea for spotlights with arbitrary start and target cone widths (without additional views and the like):
Then I realized they cannot cast dynamic shadows that way, what makes them pretty useless... :<
Slin... What ya got there? The terrain layout looks pretty good! Except, it looks like you have over-lapping or "Striping" at the edges. If that is "Striping, try changing the verticie count of your terrains so the count is as described in the manual:
"For avoiding unchunked mesh stripes, terrain_chunk should be a multiple of 8, and the number of terrain mesh squares in x and y direction should be a multiple of the terrain_chunk value."
In other words... a multiple of 8 plus 1... So, for example... 32 verticies in the x direction plus 1, 32 verticies in the y direction plus 1. So the terrains would be 33 x 33 = 1089 verticies total. Then set your chunk value at something like 8 or 16... Just experiment... That would make the terrain an even number of rows and the striping will go away and the game will run much faster and the edges will look better when textured while running it through the engine... Just my thoughts... And it does look good!
@Spike: your level looks good, some RPG action (by seeing icon mouse) ??
@Realspawn: When i tried it, i also thaught of using the simple and fast physic car demo from AUM100. Great beginning for your game, you'll just have to make great looking carts !
Working on a template/code contribution that should be finished soon, it will feature a stable PhysX car, traffic lights and traffic AI that will interact with traffic lights, stop signals and other cars while follows a path created in WED.
Sculptris WIP : alien ! For some Dead Space project, i mean levels on some space stations and bad aliens to fight, something fun and action with RPG elements.
Sculptris is great , any time you can go on details or back reducing polygons, it's really free sculpting without care about detailling level !
Wow...HeelX, your pics remind me at really bad days in the past. The time between the semesters, when you think that some stuff could be a spiritual expansion of your human neuronal feelings. Love and peace bro. I think you listen tooooo much the doors
Wow...HeelX, your pics remind me at really bad days in the past. The time between the semesters, when you think that some stuff could be a spiritual expansion of your human neuronal feelings. Love and peace bro. I think you listen tooooo much the doors
"some stuff" that begins with an L, has a S in it and ends in D? wait, i think i ruined the whole spelling outloud so no one knows what it is, convention.
I don't take drugs. I have Dr. Pepper! Huh, by the way, Enter the Void is a great film!
Thanks for the nice comments
Sid is right, both games I am showing here are for the same client and in the "games" you control the "player" (in InnerFlow) and/or some gameplay related values (aura in FireFlight) with your heart. I will certainly add soft shadowmapping, but for now the other areas and RIGHT NOW a custom level loader (gnarr!!!) and entity-recycler (grrr....whoarwhh!!!) have much more priority. And then I post to the future forum.
Great 3D model my personnal taste : -You should color the clothes pants in a different color, caus the color is too near to the body color ? - perhaps remove the grey part on arms ? looks strange !
You just have to make some level now , and we'll have an original indie FPS with original ennemies.
thanks.Im trying to figure out how to get the leather in my head on him. been experimenting. Still needs a lot of work on the shoes, claws, eye, also he has mutton chop like facial hair that sticks out, you can see it if you look close, that has to be detail normal mapped and painted.
And then....
I am going to add details of drool hanging down (all normalmapped of course), and shackles and chains all around him (if im not too lazy) Im hoping to make him wield my axe i gave out in aum 100
EDIT:
I think im pretty much don with the diffuse texturing. of course when i make the low poly one, and bake this tex to the proper uv map im going to touch it up in gimp, but any crits and comments before I retopo it would be much appreciated.
Sorry, I did not get the sense of this thread. Is this a project thread in the project thread?? Don´t get me wrong. It is cool to see that some guys are still working with Acknex but I don´t understand the above.
Sorry, I did not get the sense of this thread. Is this a project thread in the project thread?? Don´t get me wrong. It is cool to see that some guys are still working with Acknex but I don´t understand the above.
This thread about projects in project thread. Main idea of this thread to show each other your projects and to prove that 3DGS isn't dead yet. lostclimate@ model looks really great, where will u use it? Or will it be for sale? Will you animate it as well?
well this thread allows us to get kinda realtime snapshots of what we are working on instead of long drawn out posts. admittedly i overdid it a little and probably should have made my own. will probably do that soon.
@3run: both. i will be selling source art from projects im working on, I have quite a bit of it made now, (want to have a decent set before i start sales)
yes it will be animated. I'll be manually taking frames from videos i capture of myself so that i can get a mocap type look.
Sorry, I did not get the sense of this thread. Is this a project thread in the project thread?? Don´t get me wrong. It is cool to see that some guys are still working with Acknex but I don´t understand the above.
alpha, i've read some of what you've posted and wondered wtf, but a thread that is titled "What are you working on" and youd don't know what it's about???
are you looking for your usual forum argument or really not understanding what the thread of "WHAT ARE YOU WORKING ON" thread is really about?
"WHAT ARE YOU WORKING ON" : Is for any 3D work people are making and that is towards A8, it can be Programming , tool, 3D art, in game or experiment with A8. ----------------------------
One rule , should be to post only one or two shots each time, and not a lot in some few days. It's like a brief and quick snapshot (or flash preview) on what you are working on !
It's a thread where people want to see lot of different things going on , and not a WIP of a project only on several pages ! ---------------------------
@losticmate : It's ok like that, but next time you could post only one or two screenshot each time like all people Or you could post when your model will be final or when it will be great differences on your progress ?
And you can open for lot of lot of feedback and put 5 or 10 or more screenshots each time in few days, perhaps open a new thread dedicaced specificaly to your model, to improve it and show your progress each time ?
I'm not against, but i think it's better to keep this thread like a real snaphot preview of what people do. Some 1 or 3 screenshots each time , and when things have progressed a lot or new things have been done !
I like the "G" in the title.. nice ... the last shot is creepy :-) but I guess in that room you don't need a flash light... nor are "they" illuminated...
Too many. There's a reason I'm not showing the wireframe
I'm having a hard time remembering to turn off polygon adding when sculpting, so there's always too many polygons flying around. I might show a wireframe later on, when I've cleaned the mesh a little.
You should try some more contrast, as well as some more variety in that cave texture blending over a second texture or something like that. Are you btw using a triplanar projection, it kinda seems so and I couldn´t imaging unwrapping that cave to be fun . Other than that it kinda reinds me of trine. You should probably just make it a real game after that heart thingy is over .
Thanks for your comments - for now the cave texture is sufficient, I will consider using a multitexture shader in later milestone, if I get the contract for that I am heavily relying on the HDR shader and colors, its harder to get it right than I thought :-) I also had to extend my visual settings framework so that different maps can determine their own settings. The HDR shader and all visuals are very sensitive and the settings for the cave produce a total overglow in the valley scene.
The unwrapping was done automatically by Cinema, so no triplanar projection I will certainly talk with these guys about distributing an "offline" game version of this, maybe.
Thanks. There are still many issues like overlapping clouds and blur being independent of the distance to the camera, but I will try to fix those. It runs with > 760fps, but I assume the impact on the framerate will be too high for a game that is no flight simulator. The sky cube tool idea is a nice one!
Btw. 5 months ago I've had an idea for postprocessing fur, but I lost interest (and did not know how to create a nice fur outline):
Working on a procedural lava texture in Cinema4D, so that I can generate different texture instances (cold, hot lava) to be combined in a multitex-shader later in Gamestudio. In a lava scene.
Translates maybe into a HLSL shader in the future, it uses a relative easy technique, like noise mixing and colorizers...
Years ago, I have had the obsession to make crap and get very fast a deal with a publisher. This worked for some of my finished projects in the past. These days, I dont have the push to work on this route anymore, so I went back to the style of my early projects. I just take long time and work out animated bitmaps. These are all windows, needles and a chaotic script sturcture with overloaded panel elements. No "bitoperations" and strange drugs like that. And tell you what... chaotic scripts are working very well.
I think you cand find some free ones on internet, you just have to make polygon cuts, deform vertices, than put apply some dirty rust paints on the texture.
@HeelX: Really good ! does Cinema 4D incorporate some texture generator?
Kinda. You can compose any arbitrary material by composing and blending everything procedurally together. I used e.g. procedural noise textures to build the (cold) lava texture and I am coloring it by a color gradient, dependent on the noise value, to achieve the lava effect. It is really awesome!
Thanks. I've been spending a lot more time on my normalmap detail (high poly bakes, with detail overlay's lots of editting ), and i think its paying off. Its going to be a grenade hopefully
well its going to be in the player view close up for at least a second, i'd guess (actually not sure how they are going to handle the grenade system) I'm also hoping to use this stuff as portfolio work.
For the first website, that's pretty good. You're doing a lot of things right: simplicity, intuitive layout, clear design, readable. The "Nerds in Space" does not look good yet - too much aliasing (I cannot spell the word, I never remember how it's spelled). Too much ragged edges there. Also, I don't like the background. But that's just personal taste. It looks too much like a cold concrete wall.
Well, I cannot increase the contrast much more, because it would be black/white then... (Or did I understand you wrong?) - also, I think that the contrast in the text fields is high enough, no?
Yeah, the background has that noise because it looked too clean without that noise - what would you suggest? =/
The logo is just a placeholder(as well as the ad-blocker things)
(Also, is it secure enough to save the md5 hash in the mysql and compare the hash of the entered password with that? [The password is still "post"-ed directly)
Of course you should not save the clear-text password in the database, so saving a hash is an improvement. Maybe salt the hash too for another improvement of security. However: Access to your database should be very secure as well. It's of no use if you save hashes of passwords in a database which is easy to access by any attacker. (A system is only as secure as its weakest element)
Well, I cannot increase the contrast much more, because it would be black/white then...
Why should it? The brightest part of your design is the text with a brightness of 119/255. Try to increase the brightness of the NerdsinSpaceWebsite.png image file a little and increase the contrast. I doubt that you will ever go back.
Thanks. I've been spending a lot more time on my normalmap detail (high poly bakes, with detail overlay's lots of editting ), and i think its paying off. Its going to be a grenade hopefully
looks like a lightsaber shaft to me wouldnt that be anawesome stage? battle across a lightsaber? or maybe battle across the exterior of a spaceship
I am just using dynamic lights and the HDR shader. I have to admit, that it is really great, as long as you tweak it right. I also make heavy use of fog to illuminate the scene.
For the lava scene I wrote a special shader for the lava. It involves multitexturing, turbulence, an edge glow term & simple hotness estimation and fog supression (I am discounting the fog density, so that it "glows" still in heavy fog). For the foreground rocks I added a simply orange edge glow to simulate the red light emitted by the lava glow.
The aura is composed through sprites, so that it is resolution independent. And there are LOTS of particles in the scene(s) to make it feel vibrant and full of life.
The magma eruptions are particle effects, but with several different splatter images to get a "natural" eruption (see splatter category on cgtextures).
My personnal taste : -In the green level, the grass sprites, are too much low polygon, the feeling is polygon , not grass life , i mean, why not putting lot of grass. Like in Supercan A8 game ,where the grass is just perfect Your level is not full of polygons so you can bring on more grass no ? -I find visually strange and not enought natural the grey slopes on the fisrt plan on the grass terrain , why not putting real rocks insteas
Anyway, only few people can achieve suc beautifull levels ! (even me , im'not sure at all i could do as good ! Keep it up !
Make Human export to 3D Coat (transform as Voxels) : Works like a charm ! So i have my base to begin the sculpt and detailling ! Bfore i could waste hours in good proportions and shapes.
So something to the story: Your are a mercanary, in the wasteland. 100 Years ago was the world fine. But a atowwar between the USA and japan destroyed the whole world. There are lots of mutants and other creatures. But the new goverment dont fight against the mutants. They just fight against raiders, they ignore the mutants. But the mutants are the real problem. They kill erveryone. You have to bring down the goverment, and create a new. Then you can fight agaist the mutants and create a new World without Mutants, etc.
All from the Ego perspective. of cource! Sounds intresting, eh??
Posted By: Anonymous
Re: What are you working on? - 08/22/1113:36
Originally Posted By: TheShooter
But a atowwar between the USA and japan destroyed the whole world.
so why are there stil Mutants and even a GOVERNMENT if the WHOLE world was destroyed? Which GOVERNMENT won the atomwar? Or was the GOVERNMENT new founded after the atomwar? By whom?
@Quad: Am I correct in my understanding that this allows us to access engine variables with Java Script and render them to an HTML page, which is in turn rendered onto a panel object in Gamestudio?
'cause that's awesome. Using HTML to define panel elements would be just great.
javascript is only one part of it. best part of it is it's chromium, your ui runs on a separate process, supports html5/css3/flash, sports v8 js engine. You can render "tabs" on bitmaps, which does not have to be panels, it can also be level/model/terrain. And you can still interact with the page even if it's not a planar surface.
pages can be transparent. In the above screenshot html page does not use any images except for nvidia logo.
--
Though right now it's a function by function wrapper of awesomium. I will try to use as it is for a while to see which parts are not convenient, then i will add some utility functions to the plugin.
I am not really sure about their policies about wrappers i am waiting for an answer on that from the awesomium team.
But awesomium itself is free-to use commercially for indies that made less then $100k last year.
Expect a demo soon but, although this is my main project rignt now, i do not have much time to work on it, because of work(working on a full time job) and other real-life matters.
@Sebastian: Looks good! Especially the terrain, do you use multi texture shader? If you are looking for an existing project done in Lite-C have a look at my tower defence: http://www.jofre.de/?p=175
It looks great, BUT: a) texture seams. The texture's nice, but there's seams that are annoying. b) why on earth are you using a - what - 30 sided (?) Icosphere?! Make it much, much smoother, add details! Because the ship/fighter seems like it's a small one (cause it moves/rotates so fast) it looks like the planet, too, is very small, and at this distance it looks to me like you should be able to see mountains, craters, etc. Since the planet is the one main object on the screen (and in space there won't ever be much more than that planet), make it something interesting to look at! Adding craters and the like could also increase the sense of orientation for the player. c) at 0:45 - is that a lag in the video? Why is it lagging? there's hardly anything on the screen? It just doesn't seem right that the fps is so low... Or is this simply the recording? d) One thing these type of games really suffer from is the weird movement of the shots: When you shoot, you'll see the second projectile at exactly the same place where you saw the first one just a split second before, so it often looks like they're hardly moving. To make this less bad, you can use a model/particle/sprite that is blurred on it's texture (in the direction it's moving) to give it a sense of motion blurring. Just an idea, dunno how it'll look on your game...
That said, I really like the concept. Don't make this another one of the dead ones, it could be very simple to make a great game out of this, since the mechanics seem simple enough yet very enjoyable.
a) that's the top I guess, needs more polies to have it somewhat worked away. b) You're right, I'll give the planet more polies soon. But there will be many levels with planets, craters etc. are a lot of work to create procedurally when you got a random biosphere on each planet. Still it's nice to see details, just 1) hard to achieve a convincing unified look and 2) time consuming, and I need to work on so many more parts, esp. scripting side. I'm not an artist, I just try to make a solid framework for the game, but in the meantime make it look cool enough without too much effort to not loose my motivation. c) recording lag, average fps is 60 with fps_max = 60; d) Not sure what you mean, perhaps its just a recording "feature"?
Pokemon/digimon/monster hunter type game. at first it was only gonna have slime based monsters but now theres gonna be a wide variety of monsters. Right now im creating characters and working on the battle code.
Until the end of September I'm focusing on some contract work (KarBOOM related), so I haven't finished off the ramps yet. I know it has been a ridiculous amount of time since a KarBOOM release I'm hoping to get a fairly stable version done for IGF (deadline is October 17th), so I'll hopefully have a version up here beforehand so you can all help me test it
Infinity. Dungeon Deities is a kind of persistent online red-vs-blue dungeon builder slash dungeon crawler with powerups.
And loot and stat leveling now, apparently
During the last alpha session players were split into two teams competing to complete various goals, mostly involving finding things in dungeons built by the opposing team. Completing a dungeon would steal a powerup from the player that built the dungeon. This way, players competed with each other for speed boosts, attack boosts, fireballs, boomerangs, spinning attacks, and heaps of other stuff.
The alpha closed a while back so some bigger changes could be made. Looks to me like we might get more soon
Dungeon Deities is a kind of persistent online red-vs-blue dungeon builder slash dungeon crawler with powerups.
Really? I honestly had no idea, I saw the alpha test last year and was somewhat curious but I was so busy at the time I never bothered to take a look.
It sounds awesome though! What kind of connection does Dungeon Deities require to work well? My connection isn't great but it works well enough for online games most of the time. When the next alpha test rolls around I'd like to try it!
i believe it is all managed through php, and levels upload how you do at the end (correct me if im wrong) although to me this always seemed like there would be some security issues that way.
What kind of connection does Dungeon Deities require to work well?
It doesn't need a good connection at all. I'm pretty sure it only communicates with the server to download a dungeon, to report completion of a dungeon, and perhaps some checks in between in case a player steals one of your powerups while you're playing in another dungeon. But each player plays in their own instance of each dungeon.
The actual dungeon building is done through the dungeon deities website. Tiles, monsters and so on cost gold, and gold is found in other people's dungeons, or earned when a player dies in one of your dungeons.
It was a lot of fun, and it's good to see Error014's making progress!
EDIT: Ninja'd by John. As far as I know, what he says is correct (besides potential for security issues -- I mean, I understand why he said it, but haven't heard any confirmations), but that's just from playing it and talking to Error014.
Reaaaally nice, reminds me of Final Fantasy, cannot explain why...
EDIT: The last screenshot is the most realistic scene I've ever seen in this forum.
after you said the ff thing i felt that way too. i think it is the the black and white high contrast and the feathers give a slight throwback to ff8 (although i never played too much of it)
it does look very interesting, i'd be interested in seeing it in motion since im having a hard time distinguishing prerendered from actual scenes.
Thanks for the comments! Especially to you Jibb - you somehow managed to explain the game better in one sentence than I ever managed in dozens!
To answer the few questions - yeah, the game mostly communicates via PHP. It is, at it's heart, a browsergame. This means that you'll be able to play it with basically every internet connection. My own here is limited, and DD is the last game to give trouble! It shouldn't go past a few kilobytes for HOURS OF DELIGHT (more accurately, it's in the same range as refreshing a few text-only pages for a few times). Of course, communication via PHP can be a problem as far as security issues are concerned, though nothing of that sort happened last time. I think flash-games face similar issues.
There are some security measurements in place, and we're adding more this time around*. Of course, perfect security is impossible to achieve. We're simply trying to make it so that if people are acting suspiciously for some time, they'll not be able to access all parts of the game, essentially, we'll ban them from partaking in the multiplayer-aspects of the game, where its team-vs-team. They'll still be able to play the dungeons of other people (unless we ban them completely).
Since there is nothing to gain, and no one will ever SEE if you succeed (apart from us admins), I also guess there is really no incentive to cheat. It's not that there is a high-score table for the game, or anything like that.
* - Yeah, I'm not explicitly stating them. I know security through obscurity is a bad idea.
@Error014: The game needs statistics to visually compete Imagine a big map, split in small squares. Each square represents a dungeon, when someon beats the dungeon, hist team(red/blue) optains the tile, can place an own dungeon on it and progress on the map :D(you can only enter a dungeon if its attached to one of your colors)
Goal?... mh crush enemies Castle?
After crushing, season end, big statistic table of participants, kill/death points etc and map reset for a next game.
So you could have random games(like now) and war(my idea above). The last one truly gives some competetive aspects^^
Thanks for the idea! I like it, but I don't think I'll add it.
Two reasons for that come to mind: First, while teamwork is important, you can't expect everyone to actually DO that. In other words, if a spot on the map is taken by the team, people will place their dungeons there as soon as possible, and not a dungeon the whole team agrees on. You could get around that by having people in the team vote on what dungeon would be placed next, but that'd add another layer of complexity. Second, that'd also makes unfair dungeons more powerful (this time around, the dungeon-selection should be more intelligent, and thus better suited to pick out fun dungeons most of the time, ignoring the unfair ones).
..
However, there is still a visual competition. Aside from story events, which may even include something like that (the ones we've planned now don't, but again, I do like the idea :)), there's the option to buy towers. Those boost the stats of your team. If you're level 10 or higher, you have a rare chance of finding tower blueprints in a dungeon. If you do, your team can build that tower - though they are very expensive, and it can only be build if all players in a team work together on that. You can also pay spies to bring down the enemy-towers. THOSE are shown as graphics, and that allows some degree of visual competition.
Though I do like your idea. Maybe we'll add it as a pure visual indicator, with no real gameplay consequences, or as one of the weekly events.
The monster is great ! What you can do is refine the model , with smooth (specialy on fore arms on the back where there is a strange scuplt) ! Using more the crease tool to avoid the glue effect on some parts.
I'm quite impressed by your skill to form the upper body and arms. You seem to aim for a very lowpoly model, that's why you glued the throat with the tentacle-like things, right?
@rachet-thanks for the tips rachet. im new to this sculpting thing, and i use body builder magazines and pics as references when creating muscular characters.
@pappenheimer-yes i actually make high-poly highly detailed models then create a low-poly version of that model and use screenshots of the high-res model like the one above to texture the low-poly model.It looks about the same but my uv unwrapping could be alot better .
you can follow my progress on this game at my blog Sweet onions
I'm working on an ingame model editor where the kids can modify given models, because it happend that MED is too complicating and cumbersome to 8-year old kids.
You can thicken parts of the model per mouse click...
...and you can drag single vertices or a bunch of it, as you like.
thanks guys i love how the models turned out in game like theres almost an ambient occlussion effect but theres no shader in the game.
@quad
Sweetonions (title pending) is a monster catcher game where you play a kid named Horace who has entered a fantasy make belive wonderland of his own creation. but the world he has created is hunted by an evil wizard king ( a character he loosely based off of his step dad) who wants to destroy horaces world and create a new. the games art style is inspired by Adventure Time, a cartoon in the states. and most of the monsters are based off of creatures in the series. adventure time
just wanted to show some screens of (not to go into detail) my "realistic" medieval game. One large level is allready completed. I'm still looking for a skilled scripter? If there is anyone out there, please pm me. Besides, I will make a request here in "jobs offered" in about a week.
EDIT: On second thought, the second and fifth shot do look a bit bland. There isn't enough texture variation between the hills and the plains, nor enough objects to fill the scene. The third and fourth shots still look excellent though.
@CD_saber : Yeah great screenshots ! We can do that in some minutes in Unity 3D unfortunately And it would just be basic terrain if you would post that in Unity 3D forums !
Anyway good terrain textures and good trees.
(What tool have you used to make the terrain and place the trees in A8 engine ?)
in unity 3D You already have great floor textures and free additionnal bunch of real quality floor textures ! You have really good quality trees and animated grass to start your game. Something Torque 3D is following by putting quality 3D assets and characters now , they have understood !
Unity 3D is more artist side : you sculpt , paint terrain , paint vegetation , all that in real time without quitting the editor
-----------
I also use Unity Free for a Free roaming RPG action , caus i can edit any time and directly the terrain, vegetation and texture, i can import my models and textures for 3D assets and you can see them in the editor as thumbnails ... etc ... etc ... I don't have to care about terrain shaders or even terrain coding or tools, i just have to create !
I just wanted to spot, that with A8 only few people make good terrains, not a lot are seen in showcase ...in Unity you can find good terrains made by beginners everywhere !
We won't go in another flamewar A8/Unity, this thread type already exists and lot of people will say paint them in photoshop or use another third party tool I know , i know ... When Shiva 3D, Torque 3D, C4 engine,Unity 3D, Esenthel engine etc ... have their editor, A8 is just behind and no one here cares about a real complete terrain editor integrated in A8. So i stopped talking about that, and i talked now, but i stop caus i know nobody cares about that
@Superku : Sorry but your question as started my long response
Hey there, sorry for my late reply.Here are the answers to your questions :-) Terrain work was done in earthsculptor, and was completly shaped and painted from from flat. the terrain shader is a multexture shader with normal maps.
The game is supposed to be a medieval "bow simulator", but with story in it. The aim was to make the game rough and realistic (the medieval ages weren't a nice place - therefore I use dark, brown color, a lot of fog and smoke.) The problem is not the level and model creation, but I lack serious scripting skills - for example for player movement, weapons etc. Please feel free to contact me.
Therefore, IS ANYBODY (seriously) interested in joining my "team"? I do believe, that this could turn out to be a very good game, and even commercial.
yeah its a companion side kick like character. this model is just a stand in. the real horace model will have a rigged back companion that aids him through the level by turning into wings,flash light, underater breathing apparatus, etc.
by the screenshots i posted you can obviously tell im not that good at unrapping and texturing models. so if any of you guys want to help. please pm me!
If i finish it , i could give it to community for free use ! (You could use it as bonus alternative character bonus or ennemy to beat at some footing race or climbing race or whatever ?)
@ratchet: The thighs of your models usually, esp. the recent one, are too thin, no matter if you want realistic proportions or not. Try to decrease the "distance" between them, too.
If i finish it , i could give it to community for free use !
I enjoy seeing what everyone is up to and glad you started but copying someones idea and giving it away to everyone probably isn't the best approach...
What I've meant: Post a recent screenshot of your game!
Originally Posted By: ratchet
Post any screenshot,picture of anything that you are working on with A8 , it can be anything like a model, a 2D level design, a texture, a game etc ...
@MrGuest : "Your character have inspired me laugh"
I never said it inspired me laught ? just re read what i posted before ?? Who are you to bring on false information or to make such desinformation ?? I said his character inspired me , i never said it inspired me laught ??????????????
----------------
But yes, my ideac to give to community was not so good , i understand better now Well, i'll gice it to him he can use it as second playable character or some rival , or for some two players coop mode ?
@MrGuest : "Your character have inspired me laugh"
I never said it inspired me laught ? just re read what i posted before ?? Who are you to bring on false information or to make such desinformation ?? I said his character inspired me , i never said it inspired me laught ??????????????
I copied the text and pasted it... it changed your smiley face to 'laugh'. Now get down off your high horse and keep being inspired
btw you really only need to use one '?' the others really don't have much meaning after the first.
Originally Posted By: Ratchet
But yes, my ideac to give to community was not so good , i understand better now Well, i'll gice it to him he can use it as second playable character or some rival , or for some two players coop mode ?
Sorry for the inconvenience ! I didn't know smiley could be traducted and produce such things. I just wanted to say that his character was great and inspired me instead of something else
Yes, i hope to finish the character and that he will find some use of it ! Perhaps some evil rival ?
Things done by weekend... 1. Level (idea finalization, geometry, textures) that I wanted to do for some time:
2. Got to tweaking previously created character model. Re-did mapping from scratch, added fingers, and fixed bones. I'm pretty pleased with results so far.
love that my concept model was able to inspire you to make something great! your new model looks way better than the old one but like superku said, the thighs need work.Also i really love the idea of your charecter being a rival to horace in my game. It works really well seeing that horace is in a dream world manifested from his imagination. your character could be an evil alternate reality version of horace!if you want to send me the model PM me for my email. And yes my game will have a basic melee fighting system. its a platform game so im aiming for a mario 64-ish style gameplay.
@Zeusk : Sometimes not only graphics are enought to make a good game. Gameplay and game mechanics plesure can be lot mroe important. I have finished Xenoblade on WII, believe me the graphics even on WII are artistically amazing and the RPG experience is one of the best i haven't seen after FF7 or Lost Odissey ! So yes, you can refine graphics at the end and concentrate in gameplay and entertainment !
I have finished the model retopology, i'll put colors that you could change and some more painted details no normal map. I have to do animations now ! Just PM me for the list of animations you need.
@ratchet: i agree with you 100% but i have to see the big picture sometimes in order to work on the gameplay mechanics properly. it helps me get the job done
I'm agree with bart about the rest of that hand plus I'm currently working on a horror project too. May be we could help eachother, if you think in same way please pm me.
Here's a screenshot of something not happening in the video, in case your imagination is good enough that you don't need the pictures moving all by itself, as if by magic.
And here's a link to the corresponding IndieDB-page.
Can you feel it? That warm, fuzzy feeling in your stomach? That's happiness, that this game brings to you.
@Otter : Your cartoon character is marvellous , i like it a lot (It is from "Justsid 3D character contest" ?) Some 2.5D shooting game ?
For the character i made; yes the head was cartoony somehow , bigger than body and with some adult touch, not kid. You are right, perhaps i rescale the head before animating to feet with the character style for the game of Zeusk!
@ratchet yes you should study the screenshots of horace i posted in this forum to get a better feel for the art-style im going for. and i lke the basic coloring but i think the shirt ahould be black (the oppisite of the white shirt horace wears)
and for the face he should look menacing or completely evil.
@otter
i love the anime art-style you went for with your character.
Ich arbeite gerade an einem "Minenlevel". Dort wird es Zwerge geben, die ihre schätze aggressiv gegen jeden "nichtzwerg" verteidigen. Als sie rausfinden das Sid un co. in wirklichkeit gar keine zwerge sind, (bei ihrer ingame größe wäre es gut möglich) stürzen sie sich sofort auf die helden.
.....hmmm .... irgendwie erinnern die beiden mich an etwas ....
ratchet, do yourself a favour, and get rid of the seam in the middle. Nobody accept a model ever that has a seam in the middle like an old puppet made of plastic.
Import it in blender, duplicate it without changing the coordinates, make a new uv-layout, and bake the texture of your model to the model with the new uv-layout. Twice, once for the color texture and once for the normalmap.
Second, learn the proportions of a face, copy the faces with the pencil, even the simple of the following examples: drawing cartoony faces drawing faces
here another good example ratchet if you do decide to hone your face creation skills, andrew loomis has a great method. its something ive also been trying to improve recently, and it helps if you look at people and try to pick out the underlying bone structure of the skull and how the muscle stretches across it (easier to observe in real life while someone talks because you can see how it moves while it stretches)
Thanks ! For the seams, it's 3D coat baking, i'll try Blender solution. But i made standard UV , it should work directly, But when 3D Coat re import the low poly model it makes the seams.
For the head proportions ? I don't know but ib the real life all people have really really different heads, really. Some have a big mouth , some other a bigger top head, anothers some really large head , some have little heads than body and some it's contrary like bigger head than body etc ... etc ... Perhaps observe people of any age in real life ? I don't see where the head is so wrong ?
yes there are variances, but its like they vary from the standard. to know what is abnormal you have to know where the standard line is in the first place. i rarely use standard proportions anymore, but i start with the same base drawing for every one, and then i edit from there. i know it seems silly. i used to think the same thing, that i was just going for a "different style", but it really does help
btw... i hate how blender makes normalmaps not baked in blender look... stupid nonstandard z/b channel mapping....
EDIT: screwed with some of the specular and normal settings and got something a lil more acceptable. also doesnt help that im still working on the model (shoulders are still high poly) and havent gotten to the spec map yet. also the guy himself is still pretty low detail diff.
But again this is general head, it some some people havign eyes in the borders really :), or some having them near the noze and large head, really you should observe people of any age ! But thanks for the advices
@losticmate : I have two points , personnal taste like always : -you should make some leather to support the armor parts, they look like they are just posed or floating over the model , or i'm wrong ? - He seems to smile, it is a monster or not, perhaps make the mouth more angry ? (just some normal map and color repainting ?)
thanks. your not wrong about the armor, just early still working on it :)im making a club for him right now and then im doing some leather straps for the armor. as far as the smile, his face will be rigged and im still working on an inner mouth piece for him since im modeling that seperate (seemed like it'd be a pain to do it with the character in sculptris)
its for a game that im not sure which engine im going to use for it. thinking a8/7 then was going to use torque but then i lost use of my good computer so i might be back to a8/7 with it. its more of a side project and demo that i do in between other projects, but im releasing all the art for sale as i finish it (same game i made the kobald for) I've already done some a7 engine renders, and they are turning out nice, much nicer than those blender screens
But again this is general head, it some some people havign eyes in the borders really :), or some having them near the noze and large head, really you should observe people of any age !
Nonsense. It isn't even anatomically likely or possible at all. Just have look on group fotos. Just as one of thousends: http://unity3d.com/company/ Measure the distances between the edges of the eyyes and the side of the head, between chin and head top, between chon, mouth, nose (lower end), eyes(height) and top. You will see that all of the faces within those group fotos are very close to the standard proportions. (Maybe, this foto is a bit too small to actually measure th eproportions, but there are enough fotos that are very large.)
I mentioned 'anatomy' - that's a different point to consider when modelling or painting. It is not identical to proportions, although proportions depend on anatomy.
It is a point that came to my mind when looking at your model, lostclimate. I thought about what I'm missing in your model. I'm missing details anatomy. Anatomy is about our expectations how the body of the given model 'works', it is about where we expect to see 'hints' through the skin about the muscles and bones and ligaments and joints of them, and it is about 'weights', means, that we expect that a big muscle is connected to strong ligaments and strong bones that are able to bear the strength of the muscles.
Hints of anatomy is that what comes additionally in when we are forming cartoony characters, because their forms are often anatomically impossible. The exaggeration of certain parts is compensatd by anatomical details. Additionally, there are conventions of drawing (expecting) cartoony characters, that you have to consider, too. Nontheless, cartoony characters and, even more, caricature have some different rules, too. The following picture from the wiki is a nice example:
looks nice, im currently working on a td game too only for android.wondered what your doing for describing a path? I know i struggled with how i wanted to implement it. i ended up with a very unconventional way of doing it i think that still makes the game fun, but also works better form mobile phones.
@Losticmate: For the path i already used A8 Path points system to for example make a character follow path, A8 have already simple functions like NextNode , FindPath etc ... it's easy to create on the editor and to manage by code
@dice : You couldchange the textures in the final version, they look like having too much noise perhaps ? You could put some Blurr filter on your ground textures to make them less grainny from long distance view ? Or use some Bloom filter in game to have more homogeneity on graphics, it looks like all elements and textures was not in the same ambient light.
just my personnal taste.
Anyway great game , a tower defense for 3DGS ! Keep it up.
looks nice, im currently working on a td game too only for android.wondered what your doing for describing a path? I know i struggled with how i wanted to implement it. i ended up with a very unconventional way of doing it i think that still makes the game fun, but also works better form mobile phones.
There are two types of towerdefense in my game: the one's with a preset path and the one's where monsters can decide which way they choose. For the first type I used the path tool in the wed editor... every level has its default path! The second type I implemented the Dijkstra-algorithm, where the monsters choose the shortest path to their goal. How did you implement youre way of "pathing" ?
Originally Posted By: ratchet
@dice : You should change the textures in the final version, they look like having too much noise perhaps ? Or use some Bloom filter to have more homogeneity on graphics, it looks like all elements and textures was not in the same ambient lightor not lit. Event if you don't use lights you can adjust luminosity of textures to matche better ?
just my personnal taste.
Anyway great game , a tower defense for 3DGS ! Keep it up.
Hehe, I know the design looks horrible! Ill improve it when game mechanics work correctly. I dont know much about good texturing and lightning... at this time my game only uses the standard texture pack from 3DGS I also want to implement some terrains to let the environment look and feel better... I want to do so much!!! But thanks for the hints
@Losticmate : Now the model really rocks ! It is sculptris ?
@Dice : Okey they are just for prototyping. After that go simple even very simple textures can look lot better than detailled that don't match between tehm or on contrast/luminosity I think you can find good ones on internet that match between them.
This makes me think , we should really ask for A8 to have really decent texture pack for floors and terrains ! Like the free additionnal packages of ground textures that Unity 3D has !
Sculptris playing and WIP ! Unrealistic proportion : I started from some Dino Cartoon head, than turned it more human ! I know : large heaps, no normal head (but .. hey it got eyes, mouth and nose , so it's ok for me ) ! Could be some ennemy for a casual arcade game !
even with the Dynamesh tool for Zbrush 4R2 and ClyLive tool on 3D Coat : I prefer Sculptris, i feel i have lot lot more control on scultping with the Pen Tablet, and hard surface with teh crease tools are so precise and easy to make ! For me it's the definitive sculpting tool. (If Pixologic didn't have baught it , it coould have evolved in something monstruous )
Yep, after some 30 minutes of fun i have turned it more evil ! (It's inspired by the goblin model i made also)
With Zbrush or 3D cota i would have lost lot of time making that character, with Sculptris i keep good control on details as i want by zomming/Dezooming and at fast speed. Well ... it's just my taste and personnal workflow also !
Just would need a terrain and a shotgun and make a horde of Babies Zombies like that to have fun shooting at them Like borderlands little groms !
Some awesome models that gives me lot of inspiration :
Last night I've had an idea for a very simple hitzone system (there's not much to see on the screenshot, I know):
It supports an almost unlimited number of hitzones per model, it's very precise, it supports animated actors, it does not use any additional models (unlike my previous hitbox solution) or bones, and it doesn't even use c_trace, woot!
Oh no, I've just noticed that I don't get the hit (world) position for particle/ decal effects, what makes my method above pretty useless. I guess I will write a simple hitzone system with c_trace and SCAN_TEXTURE then (I will release the source, too).
@Superku : Will it be a raycast against polygons for animatec characters ? you trace against the real animated(rotated) polygons of the characters ? Others engines do that for ages (10 years or more) ... Sad that A8 don't have that simple feature.
Yes, the hitzone system is already finished. Acknex has this feature, too, at least since A6, it's called c_updatehull and reloads the (invisible) collision mesh from a certain frame. As the function is, of course, pretty slow, you should avoid every call when possible and thus I've made some optimizations: All enemies are saved in a linked list. Before you c_trace, I iterate over all enemies and check if they are close to the ray. Then I sort those close entities by distance to the trace origin and update only the collision mesh of the first 3 (any value is possible) enemies.
looks nice, im currently working on a td game too only for android.wondered what your doing for describing a path? I know i struggled with how i wanted to implement it. i ended up with a very unconventional way of doing it i think that still makes the game fun, but also works better form mobile phones.
There are two types of towerdefense in my game: the one's with a preset path and the one's where monsters can decide which way they choose. For the first type I used the path tool in the wed editor... every level has its default path! The second type I implemented the Dijkstra-algorithm, where the monsters choose the shortest path to their goal. How did you implement youre way of "pathing" ?
Originally Posted By: ratchet
@dice : You should change the textures in the final version, they look like having too much noise perhaps ? Or use some Bloom filter to have more homogeneity on graphics, it looks like all elements and textures was not in the same ambient lightor not lit. Event if you don't use lights you can adjust luminosity of textures to matche better ?
just my personnal taste.
Anyway great game , a tower defense for 3DGS ! Keep it up.
Hehe, I know the design looks horrible! Ill improve it when game mechanics work correctly. I dont know much about good texturing and lightning... at this time my game only uses the standard texture pack from 3DGS I also want to implement some terrains to let the environment look and feel better... I want to do so much!!! But thanks for the hints
GreetZ dice
ah. i missed this. i have several nodes that have "distance from entry" value most levels have about 6 values. once a character reaches a value, they choose a random next node of the next value up. this only allows seperate paths to have a 1 node depth before remerging onto a main path but as my maps are not big at all it shouldnt make a difference just have to do some creative level design
Yes, the hitzone system is already finished. Acknex has this feature, too, at least since A6, it's called c_updatehull and reloads the (invisible) collision mesh from a certain frame. As the function is, of course, pretty slow, you should avoid every call when possible and thus I've made some optimizations
@Sodoku : strange how A6/A8 engine works, in others engines you just access the nearest triangles to the ray and perform some math detection collision between rays end triangles.
The reconstruction of some Hull is typical to 3DGS, taht's why it's slow ! In some other engines you have the general character collision shape for the level, and with raycast you can check on any polygon when the character is animated directly
Well, this Hull system should be removed in A9 ! Just do like other engines , you create a capsule independant and just "attached to your character". Any collision ray if hit a character bounding box would perform more accurate math collision detection : Ray Vs List of triangles.
------------
Well i'm also very interested in your new solution
Your Hitbix ssystem was great also , caus we could check what limb was hit and perhaps make custom reactions or lacunh custom animation (the ennemy drop teh gin when hit on the hand or foreharm for example) !
Wow this forum inspires me to find more time to spend on my project so i can show it off! XD. I really like what everyone is working on and the i finally got zbrush and all autodesk software today! (thanks to my school). so I'm free to create without the limitations of free software. i'll post some of my project related creations soon.
Contextmenu added to a button with contextbuttons having tooltips.
(Btw, we are rendering everything by ourselves in directX and use RTT to show it on screen, it's faster and you can animate any 2D Object in 3D + It allows better rendering to 3D-Objects!)
painkiller: good start (sounds are pretty good, models are nice, movement seems good), but there's a few points I think need work: a) way too much sliding around. Try to get the animations and the movement code to fit! b) you almost always use the same jump-to-the-side to avoid the attack, which works because the guys only do one attack. I know it's WIP, but the guy should at least have a sideway slash and a forward thrust animation as well. Check out some of wolfire.com's combat movies. c) I know it's a game, but people shouldn't be standing upright after having been hit with a sword once, or when having an arrow in their chest. This is something that bothers me with pretty much every RPG though. d) the jump's too high. e) it would be cool if you could make the shield block even when not in blocking stance. For example, in the video, almost every hit you get on the enemy goes through his shield from the side.
I think you can make any type of ennemy like strong, or some more quick, ro some that blocks or evade a lot, or why not some having deadly attacks hard to avoid some times !
painkiller: really nice, just as others advised, if you were used separate attached helmet/shield/weapons models to base character for better hit/blocking/injury detection, it would be really fine. moreover, some camera zooming when using the bow would be not irrealistic in this case too, even if no sniper rifle is used ...and I'm expecting more weapons, more attack/defend animations with cutting off body parts, group attacks and siege weapons too
I'm just playing with environtment settings, tweaking my mapeditor, finding out new options, to get something nice without shaders... and meeting engine bugs "as always"
painkiller: good start (sounds are pretty good, models are nice, movement seems good), but there's a few points I think need work: a) way too much sliding around. Try to get the animations and the movement code to fit! b) you almost always use the same jump-to-the-side to avoid the attack, which works because the guys only do one attack. I know it's WIP, but the guy should at least have a sideway slash and a forward thrust animation as well. Check out some of wolfire.com's combat movies. c) I know it's a game, but people shouldn't be standing upright after having been hit with a sword once, or when having an arrow in their chest. This is something that bothers me with pretty much every RPG though. d) the jump's too high. e) it would be cool if you could make the shield block even when not in blocking stance. For example, in the video, almost every hit you get on the enemy goes through his shield from the side.
Originally Posted By: ratchet
@Painkiller : great Just need a crosshair when shooting with the arrow.
I think you can make any type of ennemy like strong, or some more quick, ro some that blocks or evade a lot, or why not some having deadly attacks hard to avoid some times !
Strong base you have here, Keep it up !
Originally Posted By: sivan
painkiller: really nice, just as others advised, if you were used separate attached helmet/shield/weapons models to base character for better hit/blocking/injury detection, it would be really fine. moreover, some camera zooming when using the bow would be not irrealistic in this case too, even if no sniper rifle is used ...and I'm expecting more weapons, more attack/defend animations with cutting off body parts, group attacks and siege weapons too
Thanks, I really appreciate your feedback. Yesterday I added a crosshair and also zooming the camera when you are using the bow. Weapons are actually separated models attached by script, so virtually it's possible to have a unlimited number of weapons. I will also take into account your other suggests.
Just had an idea for "advanced" dynamic lights, see the comparison below:
The first screenshot shows a regular dynamic point light, the second image shows the light with "edge detection", that is a regular object shader. It's a very simple 2D effect and it has many limitations, but there's no additional rendering involved, works on PS 2.0 and is very fast (runs at 2200-2300fps).
Well, Superku. That's just a bit too lovely and perfect. Questions! 1> Does it rely on a sort-of hallway/room layout? (that is, will it give every scene stunning and perfect dynamic shadows, or does it specifically take advantage of 2D layouts to give that lovely effect?) 2> Will this perfect shadow with perfect penumbra be cast on other objects, such as props and characters? 3> Will you give us a running techdemo, or at least a video? 4> Will you tell us how you did it and/or make it accessible in some form to the community, even if for a price?
Hopefully soon, but as we want are working on one element after another and we get more and more ideas for those element (like more source implementations or behaviors) it takes longer than expected.
Also our DX9<->Gamestudio rendering step took/takes it time [If we will ever get multiplatforming, it would be too bad to rely on dx9]
Thank you guys! As I've said, esp. right now it has many limitations. 1) Exactly, the (2D) layout matters and thus it's only suitable for corridors and hallways with some pillars here and there. Props and characters should use stencil shadows, though. 2) Yes, it "casts" on everything, it's an object shader (see below) that you can apply on every model. 3) Here you go: http://www.youtube.com/watch?v=4b3BvHuDetk 4) Currently, I take an invisible model, place it at the corner and let it calculate a (2D) normal vector that is perpendicular to the direction vector light->corner. (I make sure that the normal vector does not point inside the level geometry.) Then in my regular point light object shader, I add the following lines:
Code:
edgelightdir = vecLightPos[i].xz-dyn_corner_flt[0].xy;
fj = length(edgelightdir);
edgeposdir = InWorldPos.xz-dyn_corner_flt[0].xy;
projrange = dot(edgelightdir,edgeposdir)/(fj*fj); //yes, the case fj == 0 is ignored right now
if(projrange < 0)
{
off = distance(dyn_corner_flt[0].xy,InWorldPos.xz)/5;
fj = dot(dyn_normal_flt[0].xy,dyn_corner_flt[0].xy)-dot(dyn_normal_flt[0].xy,InWorldPos.xz)+2;
Shadow += vecLightColor[i]*fac * (1-pow(saturate(fj/(5+off)),1.5));
}
else Shadow += vecLightColor[i]*fac;
All I do is that I check on which side of the line light->corner the current pixel (InWorldPos.xz) is (in respect to the normal vector calculated by our corner model) and depending on the result of that calculation I blend the shadow/ light.
That's brilliant. Thanks for the info, Superku! It looks perfect in action, and makes me want to make a corridor shooter just to use it. Given how many corridor games exist it has a lot of use even with its limitations.
Thanks, I'm prototyping a shooter (far from Battlefield/ COD or any realistic setting) for some time now that I plan to release after my sidescroller, that's what I'm developing those small systems for (like the hitzone system). When I work on lighting for indoor levels and continue to work on this shadow system, I will probably release a working code snippet as I did with the hitzone script.
As my father likes tanks and esp. driving them, I've made a tank game some years back (good old A6 times). This year's birthday present is a new tank game:
Here it is another quick video to show the progress in the combat system of the medieval project being developed by CD_saber and me. First I show all the current possible attack types, and then try to beat the three soldiers (and fail )
You can also see that now you can turn around your enemy while in combat mode, apart of other small improvements and bugfixes.
Ah, I see. No, it's for a client. Everything is done inside PhysX, for which I had to extend the plugin. The only Lite-C code is for enabling physics, setting up constraints and collision shapes, and applying forces, when keys are pressed.
nice project! a video would be cool. what were the biggest hurdles to overcome with physx? does everything work like you want already? (i would have chosen newton for something like this. :p)
answered.. sry for my absence.. was a little bit stressed ^^" totally forget about that... becoming a father changes your live completely.. stress enough...
nice project! a video would be cool. what were the biggest hurdles to overcome with physx? does everything work like you want already? (i would have chosen newton for something like this. :p)
I'll ask my client if I am allowed to post a video. I am not yet finished, I will post more details about it tomorrow - we are going to replace the raycast wheels with solid ones, I will inform you if this went well, or not
@CD_saber I really enjoy watching these images. Looks really nice . You and Painkiller works together on this project? If yes this is great combination beacuse i really like his fighting system, and it would be interesting to see it in this level .
@CD_Saber : Great , looks professionnal. On first screen , i don't know why , but rocks and buildings seems too much illumintade compared to the rest of level ! For example the buildings could have some bigger shadow or SSAO to avoid that "cutting" edges : bright Building posed on dark terrain ???
But it's only noticeable on first screen, and for an indie project that's already amazing level !
@ratchet. very good start.i dont do much hand painted texturing, but it always fun to see it in other peoples work
@saber pretty good so far. a little bare(wip i think anyways, right?), but well built :), one can easily look at those pics and see where all the little details would fit and the beauty that the levels could have when finished.
ha ha ... it's not a pure 2D painting program in fact ... so yes perhaps hard to recognize : it's 3D coat
Yes 3D coat is scultping,retopo,uv tool, but also great painting tool. It has a very usefull feature that is Texture Tile making. You choose the resolution and you can paint on the maps you want : diffuse, normal, specular. You have : -layers -brushes -masks -bézier curves, reactangles, paths -several lighten/Darken tools -simple blurr (just press Shift when painting)
I forgot , you can indeed paint in 3D on your 3D model and any of the maps (diffuse/Norm/spec).
In particular, I've transformed the previous experiment into a useable system that supports more than one corner (now almost unlimited) and objects such as the small blocks in the screenshot. The shadow (and the system) is far from perfect, but it should work fine for my sidescroller. On the downside, I did not manage to make it a shader model 2.0 effect like the rest of my game's shaders. As this is supposed to be a gameplay element in 2-3 levels, this alienates 10% of possible gamers (according to a recent Hardware Survey by Steam).
Thanks! He complains on PS2.0 that I use too many constants: "Invalid const register num: XX Max 31." I don't get how the value XX is calculated, though. When I write f.i. float dlight_block_pos_flt[20]; and change the array dimension, it stays at XX for a few more indices and then increases by ~9 (XX => XX+9) or so. Why?
Hm..don´t know, but you could encode your constants in a bitmap to work around this limitation. But that would mean an increase in shader instructions since you have to decode the information from the map in the shader... An other possibility would be to send in only the necessary block positions for a certain area if you don´t do so already.
An other possibility would be to send in only the necessary block positions for a certain area if you don´t do so already.
Yes, that's what I do. I have up to 100 corners and 100 blocks/ objects and one loop that checks which corners are in the lightrange (with some optimizations including regions to disable shadows in close-by corridors). Then up to 10 active corners and 10 blocks are sent to the shader.
@progger: i think it would be cool if you render the rear view mirror's onto the ones on the car Also, nice you got the environment map to work. Are the reflections realtime?
@ROel yes you are right i will work on this and no the reflection is not realtime but ill add this later It is only a cubemap rendered on the car. WFG Progger
Ganz gut für den Anfang. Ich würde noch 2-3 weitere Texturen (Rock / gras) verwenden um mehr Detail darzustellen. Zudem würde ich noch härtere Kanten hinzufügen damit das alles nicht so "Rund" aussieht. Ich weiß ja nicht wofür das Gedacht ist, aber du kannst locker noch 200-300 Bäume dazu setzen. Nutze die zusätzlichen Bäume als Billboards um mehr dichte rein zubekommen.
@MasterQ32 : Not so good for my taste only : I prefer to be honest ! I htink it's beta work also ? I don't say that you haven't worked or put effort on it, but it looks far from other 3D terrain in some indie engine having terrain editor and vegetation editor.
Perhaps use multitexture, better ground textures, and the zoom is too much far for us to see something also !
@rvL_eXile, @ratchet: The whole map gets 11 textures which are blended with multitextureing, loaded from the output of l3dt i could get also more textures on the map, but in this map, there are only 11 materials assigned to the terrain wait for next screenshot!
Die Farben passen nicht (Boden zu hell, Bäume zu dunkel). Bodenvegetation fehlt. Kontraste fehlen. Küste fehlt. Außerdem sieht es aus wie ein Kuchen in Bergform, auf den jemand Bäume gestreuselt hat. Du musst das alles ein wenig natürlicher gestalten . Orientier dich an Far Cry. Die Küstenlinie und die Farbanpassungen sollten im ersten Schritt schonmal für eine Verbesserung sorgen.
Hier gehts eher um den Import aus einer Software, welche mir ein solches Terrain designen und generieren kann. Man kann damit ziemlich gute Terrains recht schnell erstellen. Ich hab jetzt halt noch Seeding-Funktion je nach Bodenmaterial und den Textur-Importer geschrieben. Geht natürlich auch mit anderen Heightmaps, wenn man dazu eine Attributemap hat (eine Karte, die das Material angibt)
@MasterQ32 : It's hard to see some 10 or more textures, i just see one grey texture, with strange green things on it. Some AAA games uses terrain chunks and some 5 multi texture, each chunck can have different textures and common ones for merging (Oblivion game) !
For your screenshot : it's very very grey, it looks strange, far from some natural terrain.But if you don't target making a game and just bring tehcnology it don't matter !
You should search for better textures : try cg textures , and put real better lightening instead of flat look and light !
@Ratchet: As far as I can read german, MasterQ32 isn't going for graphics here: It seems he is programming an importer so te screenshots are just placeholders, so they are not final at the moment.
Wo ok i thought 3DGS is mosnta and could be like BF3
It can. Not straight out of the box ("look ma! I made BF3 in 5 min!"), but it can.
Originally Posted By: Harry Potter
here are some screenshots from my 3DGS project 'pax et bellum' (WIP).
Harry, does your game actually run on any modern gaming PC?
In all seriousness though those shots do look nice. Whether they look better than BF3 is debatable, but you certainly have much more detail in your creation.
EDIT: Then again, your screenshots do look much better than BF3.
We currently got the texture things work its looking realistic now but its VERY laggy cause the textures are soo big. If we reduce size its lokking shit but there is no lagg! There is any faeture dhader etc wich scale SOME textures in the far smaller and in the near bigger? Would be really helpfull!
Yeah but I added it and its still laggy. My code: d3d_anisotropy = 7; d3d_mipmapping = 4;
I think the problem is that we genare the map (texture it and add lots of plants) The finish-textured map could be exported as wmp or something so we can load a precompiled map?
What you could do is to stream your models so that you only load them if they are near the player. Thats not to hard.
Concerning the texture size problem: I did never find a solution to display a hugh map with an acceptable quality without using shaders. If you find one, please tell me
hmmm... for me the non-shader multi-texture fixed function effect (effect = "terraintex.fx";) works fine after exporting the texture alpha channels from L3DT, of course by replacing the original textures with my own .dds (DTX3/DTX5) ones...
Yes but our textures are 1024*1024 big and the models lagg me out BTW: 256*256 looks shit but i get my full fps. There need to be a func for textures to see how far they are away from camera. Would be very amazing. Than I just could use good textures when camera is in front of them. Other wise the 256*256.
Yes but our textures are 1024*1024 big and the models lagg me out BTW: 256*256 looks shit but i get my full fps. There need to be a func for textures to see how far they are away from camera. Would be very amazing. Than I just could use good textures when camera is in front of them. Other wise the 256*256.
use dds textures. With Gimp and a dds plugin you can export compressed dds textures which can contain mipmaps and they will take less memory.
PainKiller: If you use DDS textures, how is its LOD processed? As in, how does it know which mipmap to use?
Manually by shader? Auto by hardware? AckNex Engine?
If there is no "clear" answer, can you feed my some links or useful search-terminology so I can read up on it myself? I cant seem to find any explainations 'simple' enough for me to grasp its concepts...
Im trying to build a scripted LOD system (I only got A8 commercial), and I need to know/figure a way to implement mipmaps, as I dont ATM...
LOD models also work on commercial edition, you just have to name your models like this: enemy_0.mdl, enemy_1.mdl, enemy_2.mdl, were _0 is the main model and _1 and _2 are the lod models.
Looks awesome You like to make a multiplayer FPS shooter with great weaponmodels? When you are interested plz pm me Im working together with 3 friends.
This is your shot, only I used the bucket tool in paint on all white areas. I've only used it once for every connected area - those are all the same, bright white!
Your shot does look definitely pretty, but maybe you can think of something else but extremely blinding whiteness for the sky and horizon.
draw a bit with gimp for my inventory (the big panel in the middle) it's the second try and not finished over all - "working on". But what do you think? Too dreary or okay? (my Dyschromatopsie don't allow to see all colours, so some Tips if it fits woud be great )
draw a bit with gimp for my inventory (the big panel in the middle) it's the second try and not finished over all - "working on". But what do you think? Too dreary or okay? (my Dyschromatopsie don't allow to see all colours, so some Tips if it fits woud be great )
@memoli : Terrain textures and rocks are great even if you baught them or not ! For the lightening, it's a bit too strong, you could lower it a bit perhaps ?
---------------
Back to the cartoon oldier work :
I'll try pure hand painted texture based on actual normal map to see how it can look at final !
I thought at first that you'd use the same high res textures as harry does in his middle age game for your house. But then I realized it was a real photo
Here you go, not the best he'd made probably, but awesome:
Originally Posted By: Harry Potter
In my project (Pax et Bellum) I use thousands of grass models with alpha channel. And I have no problem with the frame rate, because I use mipmaps and LOD.
Originally Posted By: Joey
Your grass has an alphachannel and it does not write into the depthbuffer, which means that every pixel is rendered for every grass plane every frame. And those are blended together, which means that the previous colour has to be read and mixed with the new one for every grass plane.
That's right. But if you use mipmaps, only the smaller mipmap must be rendered when an object is far away. Rendering/blending is much faster then. In my game I use very detailed textures (8192 x 8192 pixels). To render them as fast as possible, I use DDS textures with 9(!) Mipmaps.
Here you can see a screenshot of my game with thousands of grass models (all with alpha channel). The frame rate is higher than 60 FPS even though the level is not finished and optimized now (the grass models in this test level are old models with only 4 Mipmaps, and I use no LOD).
Can someone point me to the pictures from Harry? cant find them o.O
You can find my screenshots in this topic on 22/11/2011 11:19!
I had just posted links to some old screenshots, and not pictures of my current demolevel I am working on. Because the level is not finished yet. And I don't want to publish it before it is completely finished.
The old screenshots are not perfect and they are all work in process. They are only for showing you how the game will look like when it is finished. And I promise that it will look much better when it is finished. And then I officially will present screenshots, videos and a small demolevel.
When I am designing a level of my game, I do this in 3 steps:
1st step: Making photos. Thousends of photos. Millions of photos. Currently I have 4 terabytes of photo material!
2nd step: Leveldesign by using sprites and simple 3D models. I think this is the hardest part of the work. I have to check thousends of photos to find the best one. Sometimes it is necessary to repeat the photo shooting because important photos are missing or because of blurry photos.
Then I have to combine several photos together, correct the perspective, and correct brightness and color to fit the other textures in the game. Sometimes I have to make them tileable, or remove any unwanted details, or add additional details like static shadows (smooth shadows calculated in 3dsMax).
And I need to think about the perfect level design. The level design must be historically correct, and should make the player feel like he is in a believable environment. But the level design also must be optimized for the artificial intelligence of the NPCs and for many other things (e.g. reduce texture size wherever it is possible, avoid dynamic shadows wherever it is possible, etc.).
3rd step: Replace all the sprites and placeholder models by the final models properly modelled in 3dsMax. And removing all visible seams between two models by adding semi transparent objects like small plants or dirt.
The screenshots I've shown here in the forum were made in phase 2 or between phase 2 and phase 3. Many parts of the level are only placeholder objects or sprites. And some of my screenshots were only done for testing and are not part of the final game. Or will look a little bit different in the final game.
... And because my game will be a time-travel game, here are the first ingame screenshots of the corridor of the space station 'OSS Virginia'.
For these Models I didn't use photo textures. The texture is rendered in 3dsMax. Maybe I will finish this space station level during my Christmas holidays?!
Thanks again to Superku for his reflecting floor shader.
KarBOOM update is really coming out this Wednesday for the public! It's a bit different to the one you guys have been testing -- mostly aesthetic changes. The new effect in the shots below (each a bit different) isn't quite finished yet.
I know KarBOOM has its own thread, but... being in Showcase I can't post new images mid-thread. I'll update the top images when the update is finished.
Although frantic gameplay is a must, I also recommend that you spend a LOT of time on player input. The improvements you made to the network code last time around are exactly the kind of things that I'm talking about.
I partly registered back then (2003!) to talk about Harry's cool project. Thats like 30 years ago.
@Damocles: Mir kommt es eher so vor, als ob das erst 5 Jahre her ist. Unglaublich wie die Zeit vergeht.
Mein Projekt von 2002/2003 hat aber kaum Ähnlichkeit mit meinem Projekt von heute. Da hat sich inzwischen einiges geändert. Nicht nur die Handlung des Spieles ist heute eine ganz andere als damals, sondern auch technisch hat sich in den letzten Jahren einiges getan. Hier ein Vergleich zwischen der Auflösung von meinen ersten Fotos aus dem Jahr 2002, und von Fotos aus dem Jahr 2010/2011:
Die Fotos von damals kann ich für mein aktuelles Projekt gar nicht mehr gebrauchen, da die Auflösung viel zu gering ist. Erst seit 2 Jahren habe ich eine Kamera, die perfekt für mein Projekt geeignet ist.
Auch von der Engine her hat sich natürlich einiges geändert. Meine ersten Versuche waren noch aus reiner Block-Geometrie modelliert. Ich hatte zwar relativ bald, ich glaube es war 2003, erkannt, dass Blöcke keine Zukunft haben. Und hatte seit damals voll auf Model-Entities gesetzt. Damals noch mit Kollissionsprüfung mittels unsichtbaren Blöcken. Heute arbeite ich natürlich nur noch mit Models.
Und jetzt halte Dich fest ... zum ersten mal präsentiere ich bis jetzt unveröffentlichtes Bildmaterial von meinem Projekt (habe es soeben in den unendlichen Weiten meiner PC-Festplatten entdeckt). Screenshots aus einer Zeit, bevor ich hier im Forum registriert war. Das hier waren meine allerersten Versuche mit Gamestudio: Erste Screenshots vom Projekt aus dem Jahr 2002 (klicken für vergrößerte Ansicht):
Der erste Screenshot, den ich hier im Forum gepostet hatte, war dann glaube ich dieser hier:
Und jetzt verrate ich noch das Geheimnis, wie ich zu meinem Usernamen "Harry Potter" gekommen bin (obwohl ich selbst gar kein Harry-Potter-Fan bin ). Ein paar Tage, bevor ich mich hier registriert hatte, war ich mit meiner damaligen Freundin Silvia im Kino, und hatte mir den Harry-Potter-Film angeschaut. Und weil sie damals zum Spaß sagte, dass ich dem Harry Potter ähnlich sehe, hatte ich daraufhin folgende Fotomontage gemacht (ich als Harry, Silvia als Hermine): Der echte Harry Potter
Und an dem Tag, an dem ich mich hier im Forum registriert hatte, hatte dann außerdem ein Arbeitskollege über mich gesagt, nachdem er von einem meiner Programme so begeistert war: "Der Thomas ist kein Programmierer, sondern ein Zauberer". Das war dann der Anlass, warum ich mich Harry Potter genannt hatte.
Hehe, es stimmt aber wirklich. Du siehst wirklich ein bisschen aus wie Harry Potter im ersten Film. Und Silvia übrigends auch wie Hermine.
Ich bin ja wirklich gespannt auf das Projekt, auch wenn es vielleicht noch fünf Jahre dauert. Dein Vorteil ist, dass das Spiel sowieso fotorealistisch aussieht. Von daher kann es auch noch 10 Jahre dauern. Die Grafik ist immernoch aktuell. Außerdem finde ich das Konzept so interessant. Gibt es bisher eigentlich nur Grafiken und Levels, oder steht dahinter auch schon ein gewisser Programmieraufwand bzw. Gameplay? Ich kann mich noch an 2003 erinnern. Damals hast du irgendwas von einem "revolutinären Movementscript" gesagt, dass den Spieler auch stolpern und hinfallen lassen kann. Bin ja echt mal gespannt.
@KarBOOM: Looks very nice. It`s a bit sad that there`s no way to port your game to Linux and Mac OS besides Wine and Virtual Machines. I think it could do better than most similar games in the software centers / app stores.
Ich und Thomas sehen uns übrigens ziemlich ähnlich. Nur daß ich Hochdeutsch reden kann.
Ich weiß, Bruder! Als ich Dein Foto zum ersten mal sah, dachte ich, dass wir Zwillinge sind. Ich spreche übrigens auch immer hochdeutsch. Nicht nur, wenn ich beruflich in Deutschland bin, sondern auch zu Hause in Österreich.
Vielleicht bin ich ja Thomas der mit der Zeitmaschine zurückgeschickt wurde... Ich bin verwirrt.
Ja, ich bin DU aus der Zukunft.
In ein paar Jahren wirst auch Du soweit sein, dass Du fotorealistische Spiele programmieren wirst können. Dein sechster Wasteland-Versuch wird dann fotorealistisch sein. Durch den 3.Weltkrieg (ausgelöst durch den Krieg mit dem Iran) wirst Du aber dann die Lust verlieren, solche Spiele zu programmieren (weil eine verwüstete Welt dann sowieso der Alltag sein wird). Und Du wirst Dich an die gute alte (mittelalterliche) Zeit zurückbesinnen, und wirst ein fotorealistisches Mittelalterspiel programmieren.
Zu dieser Zeit wird dann auch die Zeitmaschine erfunden werden, und Du wirst ins Mittelalter zurück reisen, um dort Fotos für Dein Spiel zu schießen. Und danach wirst Du in die heutige Zeit zurückreisen (in die Zeit vor dem 3.Weltkrieg), wirst Dich Harry Potter nennen, und wirst hier im Forum Dein Spiel präsentieren.
Und Du wirst hier im Forum mit Deinem JÜNGEREN ICH Selbstgespräche führen.
Ich weiger mich zu glauben das Damolces anders aussieht wie auf _IHREM_ Avatar! Selbiges gilt für Michaela Schwarz.
@Sid: Nur weil DU im richtigen Leben genauso aussiehst wie auf Deinem Avatar ( ), heißt das noch lange nicht, dass ALLE hier im Forum so aussehen wir auf ihrem Avatar. Das fotorealistische Ritter-Model aus meinem Avatar-Filmchen ist übrigens mein Schwager. Das bin also auch nicht ich.
Also ich kleide mich grundsätzlich pink, trage einen gelben Baseballhelm und springe mit einem Pogostick durch die Gegend. Außerdem sehe mit 22 immer noch aus wie 10. Was soll daran jetzt seltsam sein?
@Progger: Nice start. I like the cockpit and the idea of a racing game. I`m a huge iRacing and Live for Speed fanboy. Does the game support steering wheels? Have you ever thought of buying a nice racetrack on turbosquid for you game like this one? http://www.turbosquid.com/3d-models/3d-bathurst-track/569434 That could give your project a turbo boost
Well no i want to do this game with some friends and i dont have this amount of money now^^
Now it does not support steering wheels but ill implent that later. It will be a multiplayer game for at least 8 players to fight against each other. WFG PRogger
In ein paar Jahren wirst auch Du soweit sein, dass Du fotorealistische Spiele programmieren wirst können. Dein sechster Wasteland-Versuch wird dann fotorealistisch sein. Durch den 3.Weltkrieg (ausgelöst durch den Krieg mit dem Iran) wirst Du aber dann die Lust verlieren, solche Spiele zu programmieren (weil eine verwüstete Welt dann sowieso der Alltag sein wird). Und Du wirst Dich an die gute alte (mittelalterliche) Zeit zurückbesinnen, und wirst ein fotorealistisches Mittelalterspiel programmieren.
Zu dieser Zeit wird dann auch die Zeitmaschine erfunden werden, und Du wirst ins Mittelalter zurück reisen, um dort Fotos für Dein Spiel zu schießen. Und danach wirst Du in die heutige Zeit zurückreisen (in die Zeit vor dem 3.Weltkrieg), wirst Dich Harry Potter nennen, und wirst hier im Forum Dein Spiel präsentieren.
Und Du wirst hier im Forum mit Deinem JÜNGEREN ICH Selbstgespräche führen.
Alles klar? Ist gar nicht so verwirrend.
Cool, war Wasteland Anlauf 6 dann zumindest ein Erfolg? Ich hoffe das braucht nicht wieder 8 Jahre...
-----------
Wie werde ich denn die KI im Mittelalterspiel später planen? Wird es einen Tagesablauf geben, in dem die Büttel zB morgens aufs Feld gehen, und Abends im Ratskeller sitzen? Wobei man dann realistischerweise 16 Stunden warten müßte.
Wahrscheinlich wären so ne Art "Zeitsprünge" im Tagesablauf Morgens,Tagsüber , Abends, Mitternacht sinnvoll, ohne eine exakte Zeitangabe. Die Personen wechseln dann direkt die Position.
Lese zwar interessiert mit und auf Deutsch ist es verständlicherweise für mich angenehmer, aber ich finde, wir sollten diesen Thread auf Englisch weiterführen. Es verärgert regelmäßig einige User, wenn englischsprachige Themen im Deutschen weitergeführt werden.
@Progger : Great i like the clean and simple look, reminds me Virtua Racing or some N64 Games, but some things on the track side like building or trees could be great !
@memoli: Great level composing ! You should change the grass sprite ; it looks really strange , too much low reasolution, or texture too much scaled, it's polygons are strange arranged ? Perhaps do like COD games , simple grass ?
@Harry Potter : It's "What Are You Working On" thread Pleas if you want to put big Discussion, make another thread please !
Some progress with my level for a personal Acknex project in Blender. Still a "work in progress". Street props, lighting, trees and grass still missing.
@Rendidon that's great. I find all that very grey, because you don't have put textures and lightening ? What program do you use to make buildings ? GoogleSketch ?
Yes, it`s Blender for buildings, terrain and texturing. Lighting is still missing. I hope it will look significantly better in the engine with sun, shodows and/or SSAO, fog and so on.
Quote:
Rendidon sounds cool! Cooler than Rondidon tbh =/
"Rondidon" has a special meaning to me. Of course it´s not a "cool" name or something, but it`s a personal one. I used to be a huge Harry Potter fan as a child and many other kids in school called me Ron from time to time. Later my first girlfriend called me Rondidon, I was 13 back then, and it was sweet to me and so on. So it has a meaning to me and I don`t want to change it to something else. We´re still friends.
I`m planning to use realtime SSAO in the future to see how it comes out.
Some progress (Blender 3D view). Still far from good-looking. The whole level has 48k Faces so far including vegetation (not that much right now). All buildings together are 28k Faces (about 30-40 med-poly buildings plus 50-70 lowpoly distant ones). I`m a bit afraid of performance issues, so my aim is to be chary with too much details and textures in the first place.
Added more vegetation, fixed some minor things. Getting somewhere. Next step will be street props and detailed trees. 49500 Faces.
What I`ve experienced so far: WED is indeed a barrier for optimized and detailed level design. I never wanted to believe this until I was forced to try out Blender because of the ship graphics demo in the showcase. I should have started to work in Blender years earlier. -.-
I thought the same. But the truth is: For Blender it`s just as easy. Define a material, assign it to a face, create a UV mapping, scale it and move it. Plus: You can work with more details and more exact while getting the better performance .
I thought the same. But the truth is: For Blender it`s just as easy. Define a material, assign it to a face, create a UV mapping, scale it and move it. Plus: You can work with more details and more exact while getting the better performance .
Uv-Mapping: Not really if you know how to handle it. Export: MDL-Exporter out of the Wiki from Ventilator. Shaders: I don`t have decided that yet. Maybe. LOD: The level is relativly small. Buildings and trees that are more distant from the gameplay area are lowpoly. I don`t think that I´ll need LOD for this.
Nice, but: Increase terrain brightness + contrast a little, many people think they need to have fairly low contrast textures for a realistic look but IMO it just makes the graphic look dull. I would at least double (in x and y, thus quadruple) the terrain's vertices, use for example terrain_lod and this won't be an issue.
BTW the shadows point into the wrong direction (compare last screenshot).
Very good ! From artistic point of view, you could put some 3D assets around buildings to simulate life, or some grass , rocks around them , more things. For the moment they are just building "posed" on a simple flat terrain. That's what will bring the scene to life
These pictures are just an example , we don't ask to make another Skyrim copy, we already have it , we don't need a some "copy" game without originality !
Under photoshop we can paint in some color mode that don't alter what is drawn under ! For example i could paint green, the black and white variation of the texture would remain untouched, it would just become some Green variation !
Does someone knows if that exists under Gimp ?
If not i will go cartoon, or will do haindpainted stuff !
You can do that in GIMP. Under brush options I'm pretty sure you can choose stuff like addition, multiply, colorize, and so on. Alternatively, have your grey layer as a "multiply" layer on top of your colour layer.
Nice, but: Increase terrain brightness + contrast a little, many people think they need to have fairly low contrast textures for a realistic look but IMO it just makes the graphic look dull. I would at least double (in x and y, thus quadruple) the terrain's vertices, use for example terrain_lod and this won't be an issue.
BTW the shadows point into the wrong direction (compare last screenshot).
Originally Posted By: lostclimate
lol something was bothering me, but i couldnt pick out what it was, superku got it with the shadows. should be a simple fix
The sun isn't set very powerful, that probably causes the low contrast.
Lol it seems I forgot to rotate the skycube to fix the sun angle.
About terrain resolution, maybe cd_saber can do that.
Originally Posted By: Schloder87
this game remembers me on mount & blade :-)
but looks good!
thanks!
Originally Posted By: ratchet
Very good ! From artistic point of view, you could put some 3D assets around buildings to simulate life, or some grass , rocks around them , more things. For the moment they are just building "posed" on a simple flat terrain. That's what will bring the scene to life
These pictures are just an example , we don't ask to make another Skyrim copy, we already have it , we don't need a some "copy" game without originality !
Keep it up !
Yes, the level isn't really finished, We use it for testing code and stuff, later we will add more props and finish it properly.
@Rei: tried something similar in the past, but based on several cube meshes. One cubemesh at the start, and selected cubic parts removed etc. But required one entity per Block >.<.
@Hummel Good idea, haven't thought about using shaders to make it the effect clear - thanks.
@Rackscha Using "cubes" should be quite slower, because I only create the sides of the cubes out of planes I need (plus a few extra ones because of bugs D: )
If my project (a dungeon project) will fail [and I think, like most of my projects it will] I am going to make it public.
But first I need more performance and visual improvements.
As a UI nazi I have to tell you that the use of radio buttons is exclusive for situations where you want one and exactly one option to be chosen by the user. In all other cases checkboxes are the right UI element.
These are no radio buttons ^^ That are toggle buttons... just designed like radios.
I agree with Sidney... users expect circles for radio buttons and rectangles for checkboxes (toggle buttons). You can program radio buttons with the behaviour of checkboxes, but this would be incrongruent to the user expectation, maybe they even think that this is a bug
I decided to get back to my roots and start helping total noobs in coding like me. This means very basic stuf in small mini workshops step by step.
Many things look so simple for a noob it's hard to figure out basics as jumping or a working healthbar I know i find it hard and so will all noob people There now 3 workshops and more are in the making.
Added street signs and grass. Non-brackground trees still missing. It`s sometimes a bit complicated because of the big area. I`m working preferably true to original. Many textures (street signs, facades and windows) are taken on the spot.
Next step is the other side of the river. And trees, of course.
@Rodidon : For street , search on CG Textures, or other photos on google and if needed make them tileable ! Yours have too much noise, thet looks too much agressive to the eye ! you should blurr them to try, but yes you should target photo textures, and you don't need a thousand, only some of great quality !
I finally found how to colorize a baked normal map : Just use Multiply in the selection of the tool option panel. That's good, now i can keep this method to make 3D characters and just paint highlights or shadows :
I modeled them out. Everything else would be senseless according to video memory and performance. Street texture is 256x256. The marks will be 50% transparent in realtime to look better.
@Rondidon : If you would use textures instead of polygons you will gain lot more power (less polygons to process ), and the less 3D independent objects the engine have to treat indivdualy the better he likes !
In some games like Gear of Wars , you can find 3D artist trics; like some 4 or 5 big squarre floor surface joined , that are one mesh in fact , and in game they look like they were separated : optimisation !
-------------
3D coat is incredibly complete, even for 3D painting you can use layers ,and there are lot of different types like in 2D programs so i jsut adde a multiply layer to paint color above the normal map for some color test !
The whole city consits of five meshes: Ground (including streetmarks), street props, buildings, water, vegetation. So there`re just five models to handle individually. The street marks are ~200-300 Polys for the whole level, so I don`t think there`s a big performance issue . If I would use textures for that, I needed to create various (big) image files. I don`t think this would be the better solution. Is it?
@rondidon: sounds perfectly fine to me. But a low mesh count only works well if you also only have few textures, thats why usually texture atlasses are used for texturing such scenes. Because every texture kinda means another mesh when rendering and while your graphicscard has no problem handling 500k polygons, your cpu has problems handling 2000 draw calls.
No, don't use textures! I've been fiddling around with roads on terrain in blender before, and it's a pain with textures, because aligning them to the road is quite a lot of work. polygons is the way to go, you just have to be careful that you don't get z-buffer "flickering" problems when the polygons are too close to the road.
@Slin: Thanks for sharing your brain . I use 16 different textures for the whole scene including buildings plus few textures for vegetation. Problem with atlases is that it`s a pain to get a nice UV. But for signs and other "special" things I use atlasses, of course. Special textures are as "summorized" as possible.
@Germanunkol: Texturing in Blender is still faster and more exact than texturing WED blocklevels.
If you were feeling smart, you could call it "a graphical phenomenon of the flickering kind derived from an altercation on the 'Z' dimensional distance buffer."
I always avoid it. So for me the name does not matter.
Posted By: Anonymous
Re: What are you working on? - 12/09/1108:26
I WAS working on something that could be summarized as an african God Of War, until I got into this small problem. I ran A6 on my Acer Aspire 5730z laptop and it worked just fine until I formatted my computer and reinstalled the OS. Now the 'build' button literally ignores me. I also downloaded A8 and it says (assuming my project is called 'game') Can't open game_a8.c Can't compile game_a8.c Error E3555 - Start up failure ab=ny key to abort This is the 3rd intel graphics card I've installed on my acer since I formatted it. None of them works with gamestudio. I don't know what I had before formatting that I don't have now. My direct X is 10. I started a thread on this about 2 weeks ago but have gotten no replies! http://www.opserver.de/ubb7/ubbthreads.php?ubb=showflat&Main=46503&Number=388075#Post388075
Posted By: Anonymous
Re: What are you working on? - 12/09/1109:50
Im stil working on my 3rd Person Horror Action Adventure Game (A7) <Shade-C Beta> Terrain and Water are stil wip. Most vegetation also stil missing. Player Code written by Rondidon, modified by Superku and me.
Looks good. I`d like to see the player code in action. Maybe you could make a little video? What`s that weird thing in the water? Looks like a television interference . You could work on the lighting of the scene. Moonlight is bright light and casts vague shadows.
Oh and I could help you out with some tips about vegetation. I would heavily recomment to use something like Blender to create a big vegetation mesh. Placing the objects individually in Wed is a bad method referring to performance and stability of the engine. I had a similar problem with my shooter project a while ago, too. Change the level design method as long as you`re able to. Prevent yourself from trouble and get a better workflow! I could give you links to some good tutorials and maybe few vegeation models (for free, of course). Check out Terminal26, Loopix and Dexosft. They have good grass and tree sprites and models.
Posted By: Anonymous
Re: What are you working on? - 12/09/1114:50
@Rondi: Ill think about a video once ive completed the file-package for you. About the water: Well, Shade-C (beta) water is stil wip. The Part in the middle of the water is GS watermirror shader.
About vegetation: that is why i stil wait to fill the level with grass. I wanna find a solution that has almost no effect on the performance. The Methode you suggested about using one big file for grass is one idea, but i would prefer to have a colormap where the grass models generate themself if the player is near. The Serious vegetation code is almost perfect for that, problem there is that even the invisible grassmodels are stil sucking on the performance.
Wrote a real motion blur some time ago but it did not look good (esp. edge bleeding made it look bad) and the effect was pretty expensive. Now I've got a fake motion blur that looks much better than the previous "real" one and simple ambient occlusion, both only need depth information and are self-made, haven't had a look at the usual SSAO technique(s), yet.
EDIT: Ambient occlusion comparison: http://www.superku.de/ao_compare.jpg (motion blur is still active and costs performance, but speed is 0, so no blur visible) Btw. it's of course no real ambient occlusion, it's a "depth booster".
@Chris_Oat: Vegatation is always bad for the performance. In my shooter project I created different "vegetation zones" per level. Then I wrote a little code snippet that switched the meshes invisible that were most distanced from the camera view. But that`s not very elegant and inefficient. For my new project I created one big mesh that is always visible. It has around 25k faces. Not the best method, too, but it works fast for smaller ares. Import is that you design your level in a way that the vegetation sprites don`t strike to be needy of polygons. For example use simple (one-sided) faces for the background vegatation (grass and sprite trees) and models of a higher quality for the gameplay area. Always look for the best compromise. Don`t use a model that is about 1000 faces instead of 400 faces and looks just a bit better. Don`t use a 512x512 texture if you go well with a 128x128 image. Even if it`s more work: Designing your own models instead of using pre-defined ones from Dexsoft etc. and using a great editor like Blender saves a lot FPS and nerves. Plus the result will look better if you work stylistically confident.
@Superku: Very impressive. You`re a wonderful coder.
Thank you, Rondidon! Last update on this topic: (Almost) finished this rendering setup, now with bloom. I think the subtle shadows/ fake AO add nicely to the depth perception:
(lower screenshot taken without fake motion blur, without fake AO and without bloom)
May not look as nice as other solutions and I've got a fast pc, but runs at 390+fps on 1280x720 screen resolution.
The lack of depth clipping on objects that are very far away from an edge is actually pretty cool,it looks like a crazy blend between cartoon and realistic.
@Rodidon : If you play Need For Speed etc .. i think it's textue not 3D ! Some othesr uses 3D tiles or a tool that allow to "gleu" any type of pluggable 3D tiles road made to be "glued" together ! or you could use Texture projection on top of 3D road also. But if you plan polygons and the engine handles all that, go for it !
@Redeemer: My photo is included in 3D coat by default Just kidding, it's photo from the author of 3D Coat !
Still working on the property Panel.. Now including the texture manager
It reads the Models Skins into Bitmaps, and shows them as Buttons. Up to 3 Skins/Normalmaps/Specmaps are possible to change by simple clicking the texture Button ^^ (Listbox appears to chose a new one)
@Rondidon : I don't know if i'll use the model on a project ?
I done a recolor with more brigntess colors on some parts : It was cartoon, but turned not as good as i wanted after retopo and texture baking. I prefer the original under Sculptris
Onother graphics way i see is in style of Killer 7 or Normal map with flat colors on the texture parts !
Hi, Working on a custom Path Editor: www.youtube.com/watch?v=O07tSEez7sc (sorry for the bad quality ) SOme Features: -add nodes and bezier curves(which can be changed with the help of a magnet) -move nodes over ground -insert nodes -delete nodes -save path -laod path -give nodes up to 10 skills and setting flags(for example you can set the magnet flag and the node will be transformed to a magnet) -change position and rotation properties -move an entitiy along the path(see end of video)
(The path is implemented with the help of linked lists)
I'm using the standart 3dgs GUI system. Just toggle buttons(images for hover click and off) and some background panels I'm currently working with GrünerTee (Modeler) on a game that needs a lot of pathes. I will publish more info when the first level is finished
it just stencil & simple projection shadow(for vegetation). i want to give some other details but my english not enough for this.
Quote:
Goes in the direction of the ICEnhancer for GTA 4.
Quote:
Very good work. Do you use Shade-C or your own shaders? What about performance?
Water shader is my own. Bloom from aum(Slin's shader). Also PP antialias from here. Everything else is my own, or general shaders from wiki. Performans not good. 30 fps with HD3850. But i am working on it.
@Emre : great textures ! I don't know if it is HDR ? but the white parts of texture looks too much strong for the eye , it's like having flashlights on the eyes looking ta the scene !
@Realspawn: Good composing. I like it. (Especially first song.) Is your project rts or frp?
@ratchet: All texture from cg-textures. No it's not hdr, just bloom. About the white part of the texture, you are right. I will change it a little bit. Thanks.
Can't tell to much about the game but it will come later.
I don't use any computer software for the music. I use my keyboard (Yahmaha Tyros 3 XXL workstation) and sometimes my voice then export it to any sound format i want
Thx for the compliments.
Oh and Mrguest music creation is a hobby of mine for a long time : Kind of sing a song writer if people are interested they can see all my work Games / music and so on on my youtube channel
"She's out there" reminds me 80's songs (80's pop). Good work.
@Rei: Thanks for the info. There is no problem about stencil. Because all models are lowpoly. My real problem there are much camera and shader in the demo. Depth, refrac, reflect (2), projection for shadow, proection for water, etc. I am working on enable_tree & material lod for techniques but situation look like this so far;
@Realspawn: yes, with that keyboard AND knowledge, you don't need anything else I'm a guitarist, so I need softwares to make drums, percussions, and other intrument tracks... And it was my band I quitted last year - a bit different style than yours : http://www.myspace.com/noisefusionband Currently I'm doing something similar but more complex craziness in my "home studio"...
I felt inspired so i made this little happy tune I learned myself the chords over the years but put some sheet music in front of me and i for sure can't play it
Just finished the movement code for "Downside Evil" from Chris_Oat. It was a lot of work, but I think it`s feeling pretty fine now. Superku helped out with the Z-Collision System.
Jump animation is stil missing. However the code supports it.
Nice work Rondidon! I found a bug when you press in same time forward and jump keys, then player start gliding. Also this happens sometime after jump and then player cant stop.
still working on grunts episode 2. i wanted to release it sooner but we ran into some coding issues and casting issues. so i had to re-cast the girl from the trailer, re-record all of her dialogue lines and by doing so i also recorded new ones that give you more information about your quest because apparently some people cant read
I hope the story will be better and deeper than in Episode one. As I told you back then: The game has great potential if you go in the direction of Portal with GlaDOS and so on. Epidsode one was a bit disappointing to me because it was really just a better "template shooter" with a cool graphics idea. The humour (it was meant to be funny because it`s a a template shooter) didn`t really work out that great for me . But nevertheless the idea itself combined with a neat story has potential. I`m looking forward to Episode 2.
GUI and product search for the "Einkaufsnavi" Now the whole graphics stuff is done in Gamestudio
(Don't complain about the colors, it's WIP and my design team hadn't seen it yet)
Please note the really useful treeview component (can also be used as a listview) Also the treeview has auto-scrollbars which appear when the items are vertical overflown
Broozar's not active on the forums any more, but he's quite active when it comes to CSiS. Oh yeah, I remember ulililillilililia as well... good old times
nice!, what parameters did you use? I never could get PSSM shadows look that good
Setting up great PSSM shadows isn`t that hard from my experience. You have to make sure that the scale of the level is right and the "camera.clip_far" variable is as small as possible. Scaling your level too large or too small is both bad for the performance and the result. And there might be problems when it comes to astronomical scalings like 1cm = 1 quant in huge levels. (Referring to Slin who experienced that in "Project ship"). Finding the right parameters is important.
For my level it is:
Code:
camera.clip_far = 15000;
var pssm_res = 2048; // shadow map resolution
var pssm_numsplits = 3; // number of splits: 3 for small, 4 for large levels
var pssm_splitweight = 0.9; // logarithmic / uniform ratio
var pssm_splitdist[5]; // split distances
var pssm_transparency = 0.4; // shadow transparency
float pssm_fbias = 0.00030; // shadow depth bias -> must be adjusted
pssm_run(3); //Compromise between quality and speed
//OR
pssm_run(4); //Quality
It still can be improved by setting pssm_res to 4096, but 2048 works good for me. Especially pssm_fbias and splitweight are important for good looking results.
I think i will need some shadows some day - This can come handy.
Got some free time and made a public version. Maybe somebody can give me a hint how to make tire skid marks Right now drawing pixels in bitmap - but this looks a bit wrong
More of the same. Getting somewhere with the lighting, but I`m not satisfied at all with the overall model quality (more details, street textures) and the terrain yet. Also, the water is still ultra ugly. I`m working on it.
@Rondidon : Now it looks like a real village part. For water youy should add some fresnel movement ? world reflection is not necessary ,even the shade can make it look great already :
@Esper: For the wheel, it looks very granny in the texture, like too much noise ? And there are too much baked shadows radially, instead it should be a shadow along the circle path of the Wheel perhaps ?
Perhéaps it's not baked shadow but texture variation, you should change it cau it makes loosing the circle aspect and lighetning of the wheell
@Rondi :3 Wow, This village is just beautiful! But I don't like this water- and sky shader. The sky is on the horizon to bright and the water looks like a blurry mirror. By the way: the grass is realy wunderschön!
@Espér: Bring all faces in the same direction and it looks better. a bit ....
The next enemy in the game, a dino-bird-thing! This is the first try.
Und für alle die glauben das ich dem Gamestudio forum verloren gegangen bin: Nö, bin ich nicht! Sonst hät' ich mir die post hier gespart ... ô.ò logisch ne? Ich hab nur viel um die ohren und will nicht frühzeitig das projekt enthüllen.
Esper, I don't like the logo. Sorry... I'd make it much clearer. Google "Logo" and you'll see what I mean. Logos are clear, strong, stand out and are easy to print. No gradients, only a few colors, and clear to read at any size.
Managed to work on my "A-Rpg Maker" again Today i finished the Autotile importer:
It loads a Rpg-maker XP Autotile (WinApi File Open Dialogue), resamples it completely to my format (like RmXP does intern), and saves it in the Projects Autotile Folder.
@Otter: Your model is great and cute like Mario games ; i like this cartoon design You could try Sculptriss 3D sciulpting for some of your characters , it could do good results also !
1 = the old RmXP/RmVX Autotile 2 = the new Tilemap Autotile 3 = Full transparent and half transparent Color (color picker mode) 4 = The live preview wich shows an ingame constalation of the Autotile. (all the Autotiles and the Preview are shown animated)
The Color Picker mode. It took some time to manage the pixel_to/for things.. But it´s working now. The Function of this mode duplicates each pixel to a double-size bitmap to show it 2 times bigger (easier to click). And the small rectangle shows the actual color, wich is under the picker-cursor.
For the autotileimporter: why do you repose the tiles? Doesnt make any sense to me. Your bigger "new" set, uses the same tiletypes but requires more memory.
yes.. but i don´t want to create 4 16x16 pixel windows just for 1 animated autotile.. And i will add some more Mixtures of the Autotile (2 more rows). So the RmXP Default won´t do the work ^^
@Otter : Your texturing skill is already great, it turned out nicely ! Like you said no need for sculpting. I prefer the green one, the green make more appear the flesh like texture. Just personnal taste.
really nice Birdosaururs Gamestudioisticus! but if I were you, I would change colors, e.g. use the green head with red body, or even better with red-yellow mix/transition feathers. but if eats only blue frogs, keep the blue feathers (never eat yellow frog, it's posioning)
Really nice model otter... Ive made a fast and simple colorize in Gimp... In my opinion its better to set the colors in one pattern... I take a simple Green touch with a orange beak, maybe you like it...
FPS is ~15-20 on my Geforce 8800GTS 512MB. That`s because I didn`t optimize some things yet and the level is pretty huge and detailed. My goal is to get 25FPS with max details including FXAA on my PC. The detail level will be adjustable, of course (FXAA level plus on/off , PSSM shadow quality plus on/off, HDR on/off, DOF on/off, sight distance, texture resolution).
For optimizing , do you have sprites for far away trees ? It works great and saves a lot of FPS ! For houses some LOD also like simple shapes with drawn textures instead of Full 3D windows, doors etc ... will make it also.
For the grass, wee see a lot it's a one face bilboard ! but i suppose it's not definitive perhaps ?
@Rondidon: Wow, it looks much better now. The only thing I can criticize is the darkness of the shadow of the pier (quay wall?! "Kaimauer" in german). It is much too dark - darker than the shadows of the trees. But it should be even brighter than the other shadows, because the water reflects the light of the sun.
Originally Posted By: Rondidon
FPS is ~15-20 on my Geforce 8800GTS 512MB. That`s because I didn`t optimize some things yet.
The textures of the roofs look as if you do not use mipmaps (or only 4 mipmaps?). Try to use more mipmaps (at least 10 Mipmaps), and your framerate will be much better.
Hey! I don't want to spam this thread with my stuff. Sorry But I'm so happy that you like my work. Many thanks for your suggestions. The color of the bird varies with its strength.
My next mission is: the weapon of the big Boss. Has anyone experience with warhammers?
Rondi: The water looks so much better than before! A beautiful view on the city, with the -still very bright sky-. PS: Congratulations to your 1500. post xD (eine echte leistung! Denk mal zurück an deine ersten posts und schwelge in erinnerungen)
@Rondidon: nice! use grass models instead of grass sprites. I use 24 poly models based on Loopix scattered grass model, with weighted randomized clipping at 1st and 2nd lod distances, and works fine in huge quantities. I use sprites only as details like flowers what can be swithced off easily on slow pcs.
@ratchet: I use 8 poly models for far tree lod models. how is it possible to use oriented sprites as lod stages of models?
@Otter : Your weapon without texture is already very good !
@Sivan : In fact playing Guild Wars 1 , they used some sprites for very far building and other stuff ! Even if bad oriented some times you didn't notice it a lot and approchaing , a smooth alpah transition was made to put some other real polygon LOD !
Some times because there was moutains on the side, you could be far from a building only in some part of the terrain, so they put a sprite when you was far and the orientation was good (even if approximative in some manner) !
Others techniques are real time LOD sprite generation each 10 frames or more you recalculate your very low LOD model and make the sprite of it.
Because when a model is very far it will take lot of time before you move of some degrees around it or lot of time to approcha it ! There was a subject on that in some GameDev article.
But well, you can just make some very very low poly model it will work also
@ratchet. yes I saw it in AAA games too, but for smaller objects (or I didn't realized in case of large ones because they were made well ).
normally I set as much models as possible to be static, and use _1/_2/_3/_4.mdl for lod. the last stage is usually like 2 crossing sprites (e.g. in case of trees, with the same image on all sides), but for a building a cube like thing is better with side specific texture. or in case of not important objects, I use less lods, and apply simple clipping.
alpha transition is nice, I tried, but I think it is only for low number of objects, which have the importance of nice distant look (towers, large buildings, object used as guideance). but if an object is dynamic, it is really not too costy to make a rare check... I think I will make some new tests... and it heavily depends on the actual game, how much graphic detail is important...
Alpha transition is very greeat. just check Oblivion or Skyrim !
You are in place A and have lot of details and characters around you , Place B is some meters far. Now if you are not distracted from all details in place A, if you stop playing and just watch, you look at the place B, you'll see nothing only basic structures , no characters , no details 3D assets.
Now you approach slighlty step by step and you'll see as you approch to place B objects fading with alpha transition very very smoosly, sometimes with no LOD ! Same things they were characters you didn't see that appears with alpah transition.
It works just great when you are playing, caus you are concentrated on other things and the alpha transition makes that you don't notice the appearence. Compared to old clipping engines of old days , here you won't notice anything
Just create a very small teasing promo of my upcoming project which slowly is coming in a good way to life won't tel more about it just have to wait and see
Also am Terrain (algorythmus?) musst du noch Hand anlegen. Der Editor allerdings: Sehr schön. Übersichtlich und dem Screenshot nach nutzerfreundlich. Kann man einzelne Objekte auch zu einem Mesh zusammenfassen? Wie sieht es mit automatischer Erstellung eines Gras-Meshes aus?
Der Terraineditor-Teil kommt noch. Das Terrain wird aus einer Map generiert - die Vorgabe ist, dass das Terrain ab NN auf eine Höhe x hochgeht und darauf residiert.
I just mined out my first Gamestudio trial from Nov-Dec 2010 created with A7 free, that dealt with making a 3D map of Europe, North Africa, and Middle East (Asia until India), based on satellite height data (combined with underwater topology) and ground images. it featured texture drawing onto single skin terrains (also for multi-terrain levels), what I'm just implementing into my editor project MapBuilder (terrain height manipulation and terraintex.fx based multitexture editing is nearly finished).
once it can be a base of my simple turn based strategy...
Perhaps some more crisp , detailled textures ! They look very low res like some N64 game ; if it is PC you should rise the level, perhaps some detail textures ?
yes, I will exaclty do that. as I mentioned, it was my first trial with 3dgs, I used simply a not too high resolution satellite image, 12 terrains (33x33 vertices) with attached edges, having 1024x1024 1 skin texture each, plus a detail skin from a Lite-c workshop. and I made a couple of mistakes too (like increasing ambient that resulted in hiding terrain shades)
My first COMPLETE mini-project, as a form of self-tutorial.
This is the VERY first of many based on suggestion the THIS THREAD.
It is VERY much a work-in-progress ATM, just the level and the (potentially) 'active' environmental objects to interact with. ie: ceiling fans and wall-ceiling lights can be shot out.
[FYI] Anyone got any suggestions on what sort of texturing I should do to the walls/floor? (shopping mall, sewer, etc...)
No screen-shots ATM, I still cant post pics for some reason...
Why not increase the challenge once we hit the four things ? Like moving ones ?
You could make bigger effects when we hit the ventilators, we see nothing noticeable when we hit them ! When all was down there was lightening issues on the wall behind !
Finished today the loading of hmp's, (re-)saving for terrain entities and other functions for dealing with (modified) terrain entities.
For accomplish the following screenshot, I do the following:
create terrain from bitmap (height values on red channel are encoding the height)
create a lod version, in which the mesh is reduced by 95%
load some common terrain from .hmp file
create the same again and modify the terrain by the heightfield which I got after reducing the first terrain. Since the target terrain has more cells than the (reduced) source terrain, I bicubically interpolate the height data
then I save the modified terrain to file and load the newly created file from hard disc as a brand new entity
(click the image to enlarge)
Please note that all procedurally created / modified terrains have correct normals.
And the code it is simple:
Code:
function terrainPerlin ()
{
my->scale_x = my->scale_y = my->scale_z = 0.07;
}
int main ()
{
video_mode = 11;
level_load("");
vec_set(sky_color, vector(100, 50, 25));
vec_set(camera->x, vector(0, -663, 349));
vec_set(camera->pan, vector(90, -26, 0));
// create terrain from bitmap
BMAP* bmapDonut = bmap_create("donut.bmp");
float minZ = -32, maxZ = 32; // min and max height for 0..255 grey values
float width = 140, height = 140; // x- and y-size
Hmp* hmpDonut = hmp_createfrombmap(bmapDonut, width, height, minZ, maxZ);
hmp_addskin(hmpDonut, hmp_createskin(bmapDonut, NULL, NULL, NULL));
// create entity from hmp representation
ENTITY* entDonut = hmp_entcreate(hmpDonut, vector(-300,0,0));
// create 5% lod version
float lod = 0.05; // 5% (mesh reduced by 95%)
ENTITY* entDonutReduced = hmp_entcreatelod(hmpDonut, vector(-150,0,0), lod);
// derive hmp from entity
Hmp* hmpDonutReduced = hmp_createfroment(entDonutReduced);
// load some common terrain from .hmp file
ENTITY* entPerlin = ent_create("perlin.hmp", vector(0,0,0), terrainPerlin);
// create the same again and modify terrain by
// reduced heightfield, bicubically interpolated
ENTITY* entPerlinModified = ent_create("perlin.hmp", vector(150,0,0), terrainPerlin);
hmp_entpaste(hmpDonutReduced, entPerlinModified);
// save modified terrain to file and load as new entity
hmp_entsaveto(entPerlinModified, "perlin_modified.hmp", false);
ENTITY* entPerlinModifiedLoaded = ent_create("perlin_modified.hmp", vector(300,0,0), terrainPerlin);
}
Thank you I will finish the saving of maps for my editor (regarding the features that are supported, mainly terrain and arbitrary objects, more will follow) and then I'll add terrain deformation and then I show maybe a video or the like
Yesterday, i started to do a commandline version of my GenesisPrecomipler, so integrating it into other IDEs should work fine.
Here is it running from CodeBlocks:
My Precompiler starts the LiteC compiler when no errors occured. I save the LiteC errors into the usual ackerr.txt. For some reason it only says: Error in line xy But never shows the filename where it occured o.O
It looks quite nice. I like especially the second screenshot. The demo works well, although some hick ups, but the grass is missing. Would like to hear some ambient sounds, to see some life, maybe lizards, maybe flies, an enemy, some riddles, something that gives you motivation to run around and search for shelter or amunition etc.
worked on a veeeeery old project.. my "Keeper of the Dungeons"..
i managed to create the whole map out of terrains (still buggy.. >> see <<) But with a great FPS boost (500+ FPS *__*)
And my second success.. Destroy blocks of the terrains.
Leftclick sets a marker (the hazard stripe thing). this marker kills Health of a (pseudo) block - wich is a struct object. When health is low enough, the block will be destroyed (vertex movement) and the marker erased.
Next step is: moving the vertex, so the walls have multiple faces too, to create waving or rocky shapes.
All still very WIP. obviously by the random shadows from objects that i still havent exported from blender to a7 . also the model of me has a temporary non-efficiently uv mapped bake from a different high poly with several skins. the bad uv mapping+low res is what is making the normal map on my face look all muddle-y Lighting is a mix between baked and blurred stencil.
getting frustrated by the normal maps losing depth from every diffuse texture except of course the wall which is blindingly opposite. because the wall is painted white, it loses pretty much all diffuse value in being overpowered by the normal maps
Hey Lostclimate, try to reduce the lights a bit its a bit to strong. Maybe add a "hemi" light to get more atmosphere. Maybe change the Light color to an"warm Yeallow" ?! At the last Picture: Please remove this green "dirt" (?) it looks wrong to me (Color is to strong in my opinion).
Thats great Losticmate. If you would just remake the sort of armor more metal, with hard edges you would have the prefect model What program od you use ? Sculptris ? Your anatomy sculpt is 100 times better than mine also.
@lostclimate: I hope I can explain this correct in English. The room is great in detail and looks really good.
But let us speak about this model: I think you make too MUCH detail in the face of this creature. At first sight, I could not identify it exactly. Where is the nose, where is the end,...? There are too many spines and bumps that confuse. Don't understand me wrong,.... You can still make lots of details, but defined and following a pattern. The feet are -I think- just in work, right? (Because they still look very much like clay.) And the ears are too bony, but enough of the criticism.
I like this scale texture very much. You've managed this with Sculptris? Impressive!
thanks for the comments and crits! the gun looks great so far otter. provided you continue with the same cartoony style youve been leading with recently.
as far as the monster model, yep all in sculptris. Ive learned a few tricks recently with it, so i think im kinda still refining my workflow for it. yes lots of stuff is still WIP all sets of claws are, the face isnt textured, and the bumpmap on it does look a lil rediculous, gotta smooth some stuff out.
Losticmate, you could be inspired by that for the armor in hard surface : Sculptris hard surface
Even without textures it has great patterns.
Perhaps it's personnal taste if i prefer more simple htings and less crazy details sometimes? But less details and more clean surface can be good also !
Sometimes general shapes and patterns are lot more important than billion of details.
also wanted to add about the youtube video you posted ratchet... part of the problem i have is im using a computer with a single core old amd athalon 64 and an integrated nvidia 6150 all with a whopping 1gb of ram... I cant get nearly as many polys in sculptris as that guy had, i could probably if i had zbrush throw a rough copy from sculptris in zbrush to straighten and harden the lines, but i dont have zbrush either:(
I rely heavily on painted bumpmaps for some of the smaller details in that character.
I of course still do have a lot of smoothing out to do especially in the face, and have already done a lot, still have to fix the bottom claws and finish the diffuse map. As far as the part you wanted as metal ratchet: its supposed to be bone like material that makes a hard armor cage.:)
EDIT: I swear I'll stop hogging this thread soon
coming sorta close, still have a lot of color correction/brightening to do and small normal/diffuse map details to add, but soon its off to blender
Very nice character model and pretty detailed. You`re talented
I worked on improving the vegetation mesh of my river level and added wineyards, new waving bushes, trees and stuff. That`s how it looks like in the Acknex engine so far. Of course many details are still missing, but I think I`m on the right way in terms of the colors.
And I modeled some more fish. This is a zander and it`s lighted pretty bad . For texturing, color adjustments, alpha channel creeation and stuff I use Photoshop Elements 10 and Blender for modeling. Just if you`re wondering.
@Otter: Is this Commander Keen`s lazer gun? Looks a lot like it . Very nice modelling. You`re awesome when it comes to comic style.
Creating some overlayed UI for Gamestudio. Events are sent via Windows Messages, Design is done per xml file. You can setup settings which are stored in a file, also it supports overlaying the default starter window of gamestudio.
the game looks very nice, but it runs very very slow on my system. And the installed size is 158MB.
Suggestion For an Indy game: -dont try to reach AAA game level detail and effects.
This will on the one Hand overload you with work to make a complete game, thus killing motivation , and it is not quick and easy to run on many systems.
There are too many professional titles that look better, are complete and run much faster.
-> unless you use this as showcase for getting a modelling job in a company. Then this approach is fine. I suggest to program an automatic camera ride through the scene, wich also helps staging effects easier.
Suggestion For an Indy game: -dont try to reach AAA game level detail and effects.
I disagree with this. AAA developers are people too I think the main issue is scope, rather than detail.
That is, detail can be sacrificed for the sake of scope, especially for Indies, but I would not want someone to think "low detail" is a good general rule for Indies.
@losticmate : I understand with a lower CPu you can'o as high. But Sculptris have the hidden feature to speed sculpting on some parts only
From what i see you are alreday high in details. Even with the details you have i could make more regular curves.
I think the main problem on your model is lot more that you don't paint regular cruves, even on muscles, this is distortion on surface curve.
Some example : [URL=http://imageshack.us/photo/my-images/820/sculptriscurves.jpg/]
If you want more control on your surface sculpt i recommend a lot that you use "Clay" feature. This way you keep big control on surface cvolume, like making volume by adding layers.
It takes a heightfield, a probability map, an average distance, a minimal distance and generates a distribution (1st pass).
If relax is = true, all generated positions will be iterated in descending order of field probability (so that items with high field probability have more influence over lower probability items in terms of position stability) and too near items are pushed away, if the minimum distance is violated (2nd pass). If the previously generated random item probabilty then violates the (predefined) field probability, the item will be deleted; otherwise it gets updated (height and field probability).
Then, in the 3rd pass, pairs are searched for, that still violate the minimum distance, and are deleted as well.
After finishing this, for each item a callback will be called and the heightfield, the XYZ coordinates (in heightfield space) and the field probability at that location are passed. This gives me the opportunity to decide by myself what to do for each generated item:
In the function above I generate a cube on the terrain, whereas the scale of the cube represents the field probability at that very location.
It looks like this; the picture in the upper left corner shows the probability (black = 0%, white = 100%); parameters were average distance = 24 and minimum distance = 12:
(click to enlarge)
Posted By: Anonymous
Re: What are you working on? - 01/20/1212:27
very cool HeelX, are you gonna share this one with us once its finished? will it be also work on A7?
very cool HeelX, are you gonna share this one with us once its finished? will it be also work on A7?
Thanks! - Maybe, but that decision has to be made in the future and although I own the rights for the code (it is part of a client's project), I don't know what happens afterwards.
@Pappenheimer : Yes the reduce Brush will help a lot specially on surface with no details : surface details can be painted on normal map after the baking. The point is that he needs to make lot more regular surfaces, less disturbed, it looks very like some Clay trying on curves.
I made some torso in the same style :
Perhaps it's better to have basi surface good and regular before painting details on the normal map.
Perhaps it's beta ,the contrast and luminosity between palm trees and grass is too much (grass is too much blokc color dark and palm tree is too much lightened).
You can take this picture as example of good textures color and luminosity :
@RealSpawn : They are not from me , but by people having made these iphone games.
@Rondidon: I think your are right ! They was just to demonstrate something ! next time i'll juts post links if i have examples to show instead of pictures
Minecraft2WMB-Converter: I know i'm doing very much stuff with minecraft, but i simply like the games style and also the really great dynamic map generation. So i thought i could some stuff with GS on Minecraft generated maps: Only Problem: Total compile time of this map was about 3 hours
@Germanuncle: I get about 20 FPS on 1920x1050 with 118k Tris. The map is currently not compiled as mesh and not optimized from my side.
@ratchet: I don't want to make games anymore I will make templates, snippets, plugins, tools and such stuff. If this project gets good enough for me, i'll publish it.
Only Problem: Total compile time of this map was about 3 hours
Did you remove invisible cubes? You could write a simple function to remove any cube, that cant be visible from the outside. Just itterate over all cubes, and mark those invisible, wich have a solid cube on all 6 sides. This cuts down the number ob cubes by 80% on average. (you probably want to do this before exporting the data to wmb)
Also check that the cubes are exactly on non floatingpoint positions, this should help the buildprocess a lot.
I've done this already. But I've postet the screen before i've done this. Next step is to set all faces to invisible which are next to another block. Hope i can get some speed boost. Another thing i thought about was adding only visible faces to the map, so the compiler only needs to optimize the mesh i have and calculate lights for those. I think that would be the fastest solution.
Minecraft2WMB-Converter: I know i'm doing very much stuff with minecraft, but i simply like the games style and also the really great dynamic map generation. So i thought i could some stuff with GS on Minecraft generated maps: Only Problem: Total compile time of this map was about 3 hours
Lol I am already compiling a map for my game for 4 days! It will compile for another two before it is done. And in the end, you check the level and wonder why (the secret are the detail blocks) it took so long...
@MasterQ32: Very, very nice - I like your project.
Originally Posted By: Damocles_
Did you remove invisible cubes?
If I were you I would try to join cube surfaces that have the same texture and have the same surface normal. E.g. if you 3x3 cubes on the same z-level with same texture, you could make one cube.
Rigging and animating is such a pain in the ass. Does anyone CHOOSE to be strictly an animator? To me its just tedious. Anyways, this is the first time i took the time to create automatic controls for the rig and took time to really paint weights in nice for one of my own models. (actually i dont create controls usually for client models either, i just do it with the bare rig). I really got this one decked out with pole targets, stretch joints hooked to ik chains... makes it really easy to pose. Im anxious to see him run but all i have is small level testbeds so the run animation looks wierd. I dont feel like making a whole new level just for the model.
[/random rant]
anyways...
BTW, this scene runs at 29 fps on a computer that struggles to play doom 3 at its lowest settings. I was trying to capture as much of the proper shading and sillohettes as I could while using very few polys, this scene is about 7k, obviously most are in the player. I'm also hoping to keep level sizes fairly small to save polygons.
@Losticmate : Do you really need facial animations ? I avoid that , i have some RPG in mind , i already know that i'll just put big pictures of 2D drawn faces for dialogs, with the good expression instead of going 3D ont that part.
Yes that's lot of work, perhas you can find good tutorials on internet ? Some 3D engine have some morph feature, but no one use them !
@HeelX: Have you ever thought of applying to a job on Conitec once your tool is finished? I mean, seriously, you`re the perfect man for the job. You`re very tough at Gamestudio programming, disciplined, not that expensive because of your age and Conitec can need some extra developers. I personally don`t think that JCL will contact you per PM and ask you for a job. But maybe with a written application you`ve got a chance.
Thanks for your kind words, really appreciated. I can't foresee the future, but I will start thinking about what to do once I am half through finishing my degree around summer --- and game development is always my beloved option #1, if I can live from that
@HeelX, great job, that procedural sea-wall thingie is a really neat idea.
I also had fun with terrain editing, brush based vertex manipulation and setting hull and normals is okay, but I'm too stupid for the SDK (and absolutely have no time to dig into that deeply). I simply use the .dll what Widi used in his terrain deformer, but it works perfectly only in A7 compilation, in A8 the first chunk remains unchanged somehow (and unfortunately Widi couldn't help me)... fortunately in my editor saving/loading a vertex modifier array has its role beside direct storage in hmp.
moreover, as a relaxation, I'm experiencing with hierarchical pathfinding taking into account different unit sizes and movement capabilities (atm building up the necessary database):
Looks good. I think the colors indicate the distance-parameter for the objectsize in the pathfinding.
Since you have quite a large grid, try to make "LOD" stages for the pathfinding. Where you make one or more less detailed pathfinding-grid-overlayers, with a lower resolution. (for example: 1 field=4 lower grid fields, make it passable only when ALL of the lower grids fields are passable)
So for a distant search, the pathfinder would first look for the rough open area-connections, and then go one level deper in finding the exact way. (keeping inside the found path-fields of the "over"grid)
yes, I do something like that, to achieve group movement, briefly:
all is done within my MapBuilder editor (without manual settings required): 10x10 tiles form a default area (called as cluster), which are automatically subdivided when they contain some homogeneous sub-areas, like an impassable high slope abyss, or water, or badly walkable hill or swamp. Each area move cost is also calculated, and real passability checked too, and stored. It will be used for the high level pathfinding, thus in case of 500x500 tiles, the area array consists of approx. 50x50 clusters = 2500 elements. it can be searched fast enough by A*, taking into account move cost, and later some additional AI influence too.
The tile passability is based on calculated maximum unit size (shown in the picture by different coloured squares) depending on 4 different unit capabilities at the moment, by measuring terrain hight and slope data, plus obstacle data updated periodically, plus moving unit data updated frequently. Each units within a group will do its own, low level pathfinding in small tile arrays, using a faster (but more stupid) solution than A*.
Now I'm ready with these data structures, plus the A* array systems, plus I have the 2 pathfinder algorithm to be used (should be altered a bit of course). What is next is group order and unit behaviour system to be implemented... I realized that making pseudo codes is useful and fun, but requires a lot of chocolate
Playing arround with cluster-algorithms; I want to cluster a generated set of points into equal clusters to procedurally bake (later) trees as batches into big models instead of placing one entity for each tree (for good reasons). I am not yet done, though:
@HeeIX: looks very interesting. Can you explain it a bit further, for example where do you need it explicitly, or what are you going to do with it at the end?
Having thousands of entities for trees in a level consumes much processing power. Since in the game for which I am writing the editor for, has much vegetation on 5x5 km maps, I suggested that we could take the x/y points for each tree on the map and bake, e.g. several dozen trees together as one single model. If the trees share textures and use e.g. texture atlasses for the leaf-parts, a huge performance gain can be reached. I'll keep you updated, because that is what I am going to do after I finished the clustering stage. I implemented k-means clustering for vectors (xyz), finished a balancing postprocess, for minimal mean distance to center -or- approximate equal size of all clusters (size = number of datapoints per cluster) and will now design a hierarchical clustering that tries to 1.) create as less clusters as possible and 2.) don't overcome a limit of datapoints per cluster (this is important, because if my tree with highest polycount has e.g. 1000 faces and since I can only have models with 2^16 = 65536 faces, in this case the limit would be 2^16 / 1000 ~= 65 trees per model == 65 datapoints per cluster. A lower bound will also be introduced to avoid, that if a clusters molders back into a 1-datapoint cluster (when k, the number of generated clusters gets to high). In that case, I use the k-1 clustering and apply the same algorithm to those clusters, that exceed my upper bound and do everything until the upper- and lower-bound condition holds ultimately.
@Joozey: I just draw a greyscale bitmap with gradients and soft-brushes and used a Floyd-Steinberg algorithm to dither it; with an eraser I deleted some parts. I wrote a function that creates an array of VECTOR's for each white pixel in a bitmap, with each VECTOR representing the coordinates.
During tests - no. Later I will create a probability map from topographic maps and use the seeding function I showed earlier to generate the datapoints. This way, you can throw in an arbitrary map and matching forests are generated.
hmm... 5x5 km with extensive vegetation really needs tricks. probably this game won't be a flame thrower simulation I mean if vegetation is static, it is a really fine solution!
@HeelX: Do you use the mdl7 sdk to bake the models?
Yes, I will
Originally Posted By: sivan
hmm... 5x5 km with extensive vegetation really needs tricks. probably this game won't be a flame thrower simulation I mean if vegetation is static, it is a really fine solution!
It's not a rainforest game either maybe I am also totally wrong about the tradeoff, but I have a strong feeling about this.
I am new on game programming, i've started game studio a month ago and i made soft point light shadow shader, however i couldn't put image here.edit: i did it now thanks heelx
i'm just working on my little game and got some empty brain, no idea, maby somebody can help me, if he has a idea... I need two icons in a kind of "bubble" or something like that over the head of NPCs, where the player can buy things or can talk with them (should be just in general "buy" and "talk").
(click to enlarge). On the left you see my pictures for this in gimp, but it doesn't fit in a medival world. I dont like my actual version and maybe somebody has got a better idea or some food for thoughts?
@JoGa: put the message onto paper scrolls as the first typical solution... (can be animated nicely to open and close to get the attention of the user, even if they are small and positioned next to characters)
Yes, it looks really neat! Would you mind to make a small demo for us to have a closer look, especially when it comes to performance and image quality?
@Rackscha: i didn't do anything for game programming before, but i have a friend who can help me about gamestudio. He made me to start and he is good teacher.
@Heelx: Actually, i am trying to improve this with max 3 pointlight shadow but i am still learning transferring datas from engine to shader. I try to figure out how vecLightPos works etc..
@ damocles, sivan, Rondidon: Thanks for reply and the ideas! I made a model of a bag and some coins, which are rotating over the head. I added 2 radiationsprites, which are rotating in opposed direction. They are for standing out the bag and coins from the background. Hope, it looks better now - the paper scrolls are great ideas, too - I will add them.
The coinbag is really nice now. Not shure how the ring looks animated, its not referring to an offer as intuitively as the speachbubble. Though the player will quickly learn that a hovering icon means an offer / quest.
Just make the ring less dominant. (more transparent for example)
If its animated, it will visually stick out even with high transparency.
wow, thanks ratchet for your effort! Your idea is great, but I think I use the single model + sprite without a border because it can be seen good enought (I hope :D) Bag and coins for trading. Book for details of the story. Paperscrolls are used for abilities, the player can learn. I use the border for Items on the ground and the a little scroll of paper for the mouse-Icon, that chances if the player move the mouse over the NPCs. First time, the player should be in contact with this guys, he will get a message, that he should talk with them so he should learn quickly for what this symbols stand for. For those who want to see the icons: http://www.youtube.com/watch?v=V7RPwof8-Es The Icons are for testing not on the right place, so don't be bemazed . I hope it's good now (I dont belive how much time is "wasted" for creating so little thinks) and thanks again for all your effort!
Finished my clustering algorithm, works now like a charm with lower- and upper-bounds; spatial distribution is now taken more into account than before, see yourself:
click me
and because I don't want to bore you with algorithmic yaddayadda, here a shot of Michael, the main character of one of my personal projects - Razor for my Valentine, as I have worked on him in my spare minutes over the last few months:
The hair has to be improved, though - was my first trial with Cinema's mograph spline deformer.
His name is Michael and he`s black, but he looks like JustSid(`s old avatar) with that haicut. I smell a conspiracy. Seriously: I like the model a lot. It looks interesting and has some unique style. Like a depressive mixture between emo and innocent cuteness, but with an evil touch. Keep it up! Good work.
how do you determine cluster middles? searching for local density maximums somehow?
I do particular nothing, that is an inherent quality of the general k-means algorithm, since the mean distance to the cluster center is minimized. What I do is that I just say "no more than x" datapoints per cluster and the algorithm creates an approximate solution for it. I create per dataset several cluster instances and do some analysis on the cluster statistics (especially the variance on the size) to decide which cluster-instances are good and which are not, which has a relative balancing effect on the overall quality. If datapoints result in clusters, that are below the lower bound, I mask all datapoints in clusters which are exceeding both bounds and apply the same algorithm on the remaining points and merge the results later on. If I cant even make 2 fitting clusters on the datapoints, I do some sort of brute force solution: sorting the points in terms of distance to the smaller cluster and assigning in ascending distance order until the point swap is even.
So, at the end, I don't know beforehand, how much clusters are "there"; I just have the boundary constraint of the cluster size and I am searching in a recursive fashion and changing focus on datapoints.
You didn`t really made this in Med, did you? O_O It looks pretty detailed and frightning. I`m looking forward to your project Keep it up! Great work as always .
@HeelX - ahh okay, just the order of pictures was tricky, and I thought the cluster quantity is determined first by item quantity... (which would be also not too complicated, but pre-defining their middle positions by item distribution would be a trouble for me)
I must say, I really like this thread, even if I don't post often here. HeelX, thanks for so often providing a bit of background information about the code you use, it's really interesting to read.
~~
Ingame news shown. Players need to know whats up. Additionally, small skits reveal more about the characters, and allow us to tell stories about them. We want the players to grow fond of the characters, but it doesn't fit Dungeon Deities' gameplay, so it's placed in here -- by chosing the TV in the throne room, you can access these stories.
Finally got my daytime and weather system to work feat. - rain - thunderstorm - nice weather - sunrise - dynamic wind strength - seasons (summer and autumn so far. It works by changing the RGB values of the vegetation mesh at runtime) - JCL`s PSSM shadows . I could finally tweak them to work with my project, although they`re still very slow and halve the framerate.
And I worked on improving my level. I photographed some animals at the real scenery and modeled them in Blender.
That project is nicely coming along! Very nice one tip: if you would bake ambient occlusion into a 2nd uv map, you can achieve much more realism for almost nothing
great screenshots! The plaster (der Putz?) of the houses looks a bit too clean/to less diversity in the colour or too less shadows; I don't knwo what it is exactly, but the orange house looks a bit strange to me. Maybe too less details in the orange plaster?
Thanks a lot for your suggestions. I`ll make the rainy weather setup a bit darker. Thanks to my smart weather system that`s only a matter of some sliders . @HeelX: That`s a great idea! Thanks.
Edit: Okay, I tried to find a better "rainy" setup now. Of course the level and the lighting is still not perfect. In the game it looks better because of the moving sky, the waving grass, the rain and so on
Very nice. Unique graphics style. Reminds me on the games on our first family computer (a lightning-fast 286 at 16 MHz and MS-DOS 1.0 plus green/black monitor. ). Keep it up!
Thanks updated the pic to a newer version. I did it draw_ style for easy and quick prototyping. Creating models or sprites takes a lot of time, and I wanted to focus on gameplay rather. I do add sounds this time (usually I don't) to improve emersion. That definitely works. Beeps for important messages, slow-paced ambient techno/trance on the background.
Started working on mdl instancing; bones already work:
Code is pretty straight forward; my ackmdl.dll is hiding the ugly stuff:
Code:
Mdl* mdl = mdl_create ();
mdl_addbone(mdl, mdl_bonecreate(NULL, NULL, NULL));
mdl_addbone(mdl, mdl_bonecreate(NULL, vector(50,0,0), mdl_getbone(mdl, 0)));
mdl_addbone(mdl, mdl_bonecreate(NULL, vector(25,25,25), mdl_getbone(mdl, 1)));
MdlCollection* collection = mdl_collectioncreate();
int i, radius = 50, segments = 150;
for (i = 0; i < segments; i++)
{
VECTOR pos, angle;
var a = i * (360 / segments);
vec_set(&angle, vector(a, sin(a) * 25, 0));
vec_set(&pos, vector(radius, 0, 0));
vec_rotate(&pos, &angle);
MdlInstance* instance = mdl_instancecreate(mdl, &pos, &angle, NULL, 1);
mdl_collectionaddinstance(collection, instance);
}
Mdl* mdlCollection = mdl_collectioncreatemdl(collection);
mdl_save(mdlCollection, "mdlCollection2.mdl");
@Joozey: Your game looks so awesome. Guess what? This lead me to the wikipedia page about airports and airport infrastructures; have a look, it's darn interesting to get to know the real stuff, that drives the real world - you might enjoy it, while getting a flood of ideas for your own game. At least, I had for a theoretical, rather realistic tycoon'eqsque game I am never ever gonna make --- best wishes!!!
Looking cool. What's the goal? What's ImageFactory gonna do?
Edit: "Combine" looks pretty useless to me. Why don't you just allow multiple outputs for one value of a node? i.e. Link the value directly to the R, G and B inputs of the bmap, link the "Image" output of the Defocus node does in this picture: http://www.freewebs.com/pwnage303/screens/bloom_DOF.jpg Or am I missing something?
Variables can only be connected to one filter, everything else would require more lines of code and since this will be an entry for the 200 lines contest... Anyway, the goal is to procedurally generate textures by stitching filters together.
(P.S. - second price is also an A8 or 500€ (or Dollar?))
200 lines of code seem to be tiny for such a project. Will you be using external information for what the nodes do? cause otherwise I can imagine it'll get squeezed in there...
What we want to enter is the editor itself, everyone can use their own shader/functions. The information about them are saved in an external file, and we just read them out, pass values and process them.
At the moment we are creating a "big" version, which contains way more than the final contest-entry, especially fancy GUI animations and stuff. [Though we have not decided what will be "in" at the end]
We are at 34 344 byte pure source code - ungolfed [long definitions, no defines, with unneeded fancy stuff]
[+ Thanks to George's rules "lines" are actual lines in the code editor, not ";" ^^]
Edit2: I never want to work with the standard gamestudio ui again ^^ Little update, that really disturbed me: left/right alignment:
Just workin on a new end time game. The Game is set after a nuclear meltdown. Here is a foretaste of the game:
Hope you like it!
regards, TheShooter
Posted By: Anonymous
Re: What are you working on? - 02/07/1211:53
he always coming back, TheShooter:)
well to your screenshot....well....lighting looks ok, textures on the wall could pass aswell but why does the ceiling have the same texture, and why does the tiles tripled their size at the end of the room? Why does the "table" fly? Not much else to see, almost feels like a deja-vu
Yes, I'm always coming back. The Table is mounted on the wall, I just have to add some screws. Ive got an other texture on the end of the room:
I know, it does not look perfect, but I will not give up!
regards, TheShooter
PS: And I'm getting better everytime, eh?
Posted By: Anonymous
Re: What are you working on? - 02/07/1212:08
if the table is mounted on the wall then the shadow in totaly off. the little i can see looks ok so far, but tell me (us) why did all your other attempts failed? Was it because of the level design or the lack of programming skills?
@chris_oat: The other projects have failed because I can not really program. This one will probably fail too, but on every project, I get a little more far. Yes, I have some problems with the shadows, because I have brighter lights in another room, and I cant assume an object only to receive light from one source.
if you make a low res leveldesign, why not use low res textures then. I think this would fit better together, and saves also lots of work on the textures.
@chris_oat: I'm trying to make a detail normal map shader, with color map, displacement, diffuse map, and an occulusion effect. So not realy a complex shader.
Also, I'm thinking of, to create a new 3D lighting pipeline, but thats complex...
Posted By: Anonymous
Re: What are you working on? - 02/07/1212:29
Originally Posted By: TheShooter
a detail normal map shader, with color map, displacement, diffuse map, and an occulusion effect
Sounds complex enough to me:) Does any commercial game have that kind of shader? if yes may you direct me to some screenshots of it?!
For example, "Metro 2033" got that kind of Shaders. In addition, its got dx11, and tesselation as well.
Posted By: Anonymous
Re: What are you working on? - 02/07/1212:47
and you think you will be able to build a game like that with all the required good looking art (models and Textures, Normalmaps, Hightmaps and Detailmaps)? Dont you think a Normalmap Shader is enough for you? Even you are able to create such a shader, will you be able to programm the rest of the game (Player, AI etc) ?
In my Opinion you missunderstood somethings about shaders... The Shaders doesnt make your Level looks much more impressive without adding much more details...
So , your Game starts after an Nuclear Meltdown...okay cool...Abandoned City style... But your "ROOM" doesnt fit this style, it looks to much clean for this. Search for Google Pictures for Tschernobyl and have a look inside the Houses, you see much more destroyed furniture / dirt / cracks a.s.o...
@The Shooter: Hi, I always find some strange minor things in level designs, this time too: that type of lamp is not a direct/indirect type, so no light on ceiling should be visible (as I see it's a 2x58W IP65 industrial fluorescent lamp). Sorry, occupational disease And I agree with that more detail really does matter.
Made a little example for u... Nothing special, just 20 minutes work... Added a broken shelf with some boxes, borken wall, hole one the ground with some plates and some cables. Try to do this too and add dirt & crack decals / broken furniture / brick stones on the broken wall...
@Rackscha All the "functions" we have need to be stored somewhere, and a 5 KB source is a bad place for them. Because of that we have decided that all "functions" are just loaded by us and everyone can create their own. [That way we save size] And well, as the user needs to write them, we need to interpret them.
Thanks, but this is the only part with "real" comments and only because I had a great headache with all this stuff, because we need to bind C functions to the script (otherwise everything would have been a bit more simple).
separable hexagonal blur like described in this DICE paper: Looks pretty neat, although I would prefer a disc shaped filter but as it seams there is no efficient way to make them separable.
Very interesting , could be a total new UI system for A8 custom tools : Visual Scripting , basic shaders editing (only variables tweaking and adjustment) , 2D animations composing , cutscene editor etc ...
Will you have some sliders with min , max properties values ?
Don't think so, as you can input anything via those "vars" at the moment
We are discussing if we make a whole system out of it, if there is a big interest in it. Of course we would then make it way bigger, with a real UI and stuff.
No. We have an army of UI testers here and studies show that the border is perfect, everything else results in either exploding brains and/or crying people.
Implemented some of the shaders from the wiki to see if everything works and will be easy-to-use for the user later. [Emboss, Negative, Black/White, Sobel Edge]
I also added a color-to-alpha and Sid a special "right" version of a monochrome shader:
Also improved the quality of the connection lines, still thinking about blurring them.
Hey JoGa, your game is developing really nicely! Here are some (maybe) valueable remarks from the video; they not only cover the dialog system (I hope this doesn't offend you) and I just wrote them down to give you some hints how to improve here and there, too:
- The main character seems to float? - Is it right that the dark character (the merchant) with whom you are speaking with, seems to be 1,5m taller than the main character? - I really like the inventory, you were inspired by Neverwinter Nights, aren't you? This is so great - Are the character skills the one from you or the one from the merchant? Because I don't know who of the both has which money (there is only one money value written to the screen). - The whole GUI is done really nice. Though, I had problems to read the text... I don't know, have you tried a different font? - Is your dialogsystem already able to cover conditional and branched narratives?
Hey HeelX, thanks for your (of cause) valueable remarks, thats great! - he does not float, but the shadow is bit bad, i think i will add a decal under him, to avoid the "floating-feeling". - I so delighted about the working script/panels I didn't recognize, that I scaled the merchant in WED AND in the script, thanks! - The skills are not included yet. I have to write a script, which create the purchasable skills in the inventory of the merchant equal to the player-level (as more the player "grows" in his level, as more skills will be purchasable). Sorry, didn't mention that ( 1 skill insert ) - Okay, that's sad, but maybe I attunded to the font so I will have a look for an other one Is it just the dialog or all text (the values in the inventory on the left for example or the headlines)?. - The dialog is really, really simple because I want to tell the "real" story as audio, so the player have not to read that much and can play. (for example he have to walk a distance and as he walkes, he reads a letter as audio comment...)
I never played Neverwinter Nights except a little demo (just Dungeon Siege and D1+2). I hope some pleople forgive me, because I know it's a great game But thanks for the hint, I forgot this game, maybe I will play it now - for "inspiration" how they solved some problems in such a game.
I thank you for your comment and the time you have taken!
Finished vegetation generation; though, we waived for now the mesh baking via clustering, because it turned out, that my algorithm is too unstable for low cluster sizes - too bad. Plus, I finished the city editor (took me only 2 days, eek!), which you can use to place vast city environments in a fractional amount of time than you would need with WED. Plus, vegetation and buildings adapt to terrain deformations.
I don't know why, but I have the feeling you're going to soon be releasing game after game using this engine. And these games are going to be ones people talk about
Cool scene Rodidon ! I see lot of polygons even on the far horizon, perhaps you could use lot more low poly trees or billboards fr distant trees ? Perhaps i'm wrong ?
um, those shots are from in blender, i doubt theyd have anything in low poly versions ... thats for when he brings it to GS i assume
great shots, the only thing is and am sure youre aware is that your textures are rather clean but thats expected, maybe if you could bake AO into them, it'd greatly improve your scenes
Improved the detail grade and added some grass, a forest cabin and some more stuff.
I`ll try to add some baked SSAO in the future. At the moment I don`t really have time for experiments, but I`ll look into it as soon as I have some open space.
Managed to build a relatively fast (as fast a the hexagonal filter of same size) separable disc shaped filter. I put a small demo together showing the hex and disc filter in action:
Actually, Bokeh is just the result of blurring image parts with high brightness contrast using the right kind of filter kernel (f.i. disc or hex, gaussian wont work->too soft). HDR DoF is the usual application. Building larger kernels isn´t such a prob with the hex filter since it has linear complexity. Adding additional passes with the right sampling offsets can give you quite a large radius. But I´m not sure in how far this works well for the disk kernel. A common technique is to render sprites to achieve the bokeh effect where it is necessary. This way you can have quite large, arbitrarily formed results.
I wondered, because corresponding to your original image (the 2px particles), there is no blurry mish-mash like with regular gauss- or box-blurs. I'm just asking, because in the long run I'm looking for better and faster alternative to blur my SSAO images, the better if the kernel is dynamically sizeable
I think I have a run at the moment; instead of playing Skyrim I coded a contour tracing and region segmentation algorithm in Lite-C. Even though I didn't fought against dragons and took some Skoma, it was quiet fun, too:
The changing color in the contour image indicates the order of contour points, stored in a linked list. It can handle degenerated cases, too (1px lines).
If you weight all samples equally for your box-blur, you will achieve the bokeh effect too...but quads are usually not perceived as aesthetic as hexagons or discs for this case.
What applications do you have in mind for your shown algorithm?
Finally exported a first version of my new level to Gamestudio. Still a WIP. Some sprite trees in the background have to be replaced, normal mapping on the stones is missing and the water needs some color adjustments and a better coastmap.
@Rondidon : you are becoming a great level designer and creator HDR + PPSM looks great ! The trees shadows are done from where ? it is shadow on trees, texture projector or some other shader ?
The trees are split into two different meshes. The leaves are animated and use an alpha test material and the trunks are the standard gouraud-shaded "mat_model" material with a custom ambient lighting formula. There`s nothing special about it . For the future I`m planning to use some normalmapping.
Quote:
you are becoming a great level designer and creator laugh
What applications do you have in mind for your shown algorithm?
I am going to extraxt pixels from a map that belong to harbour areas and auto-generate meshes for it. After getting to know these pixels, the next logical step to take is to get to know how many regions the binary image has and what contours each region has. Then, I'll filter the contours and place walls at the water-sides of the harbour areas and generate meshes for the floor parts.
Quote:
Finally exported a first version of my new level to Gamestudio.
This looks really cool, is this a portfolio project or work for a client?
Looks really good, esp. the first screenshot! The only thing that I dislike is the shoreline, I would simply reduce the transparency of the water where it meets the sand (that is where it is white right now).
This is made in Acknex, so it's Windows only until jcl decides to do the right thing. The game is part browser-game (you create the dungeons in a browser, buy the tiles, ...), so maybe for that, we'll create a special mobile site. But I don't think this client (where you do the dungeon-venturing) will be made into something avaible on smartphones.
Thanks Rei! I do like the theme, though it's probably going to get replaced with something tailor-made for the game. Can you elaborate what you mean with the "one-step movement"? It's actually intended that the blocks can always only be moved one complete tile - turned out to be less confusing.
It looked like you had to run backward and forward again to push the block again, if the movement of the block (accelerate / move / decelerate ) is one part [eg, even if the user still pushes always accel/decel] I think it would be better.
(Looks like to much work for a simple task otherwise, but probably I am wrong, I haven't played it (: )
Ah, I see now. No, that was just my erratic playing behaviour - you can just keep the button pressed and it keeps moving (as long as there's space).
A shame you haven't played it - we had the public alpha announced here -- but not too many people from here played it. To those who did: Thanks again, you rock! To everyone else, make sure to join the fun next time!
25-35 is great for some adventure game or presentation , simulation. Not for a game with too much additionnal stuff : complete UI, characters on the scene, special effects, Some basic AI etc ...
For shadows, you should use some UT3 clever technique : texture projection, used from lot of years , same result : Your trees won't move from their place And biggger performance (why not tunning in otpions like static or real time for trees ?)
You should use billboars that orient to camear direction for far distant trees, instead fo some static 2 crossed planes : they are too much visible, a billboard would give lot more best result for even less processing : only 1 plane !
The language can be changed with a config file. But we create the tool in german.
The file formats are compressed dumbs (like the marshal dumb rxdata files from rpg-maker xp), wich can only be accessed with a DLL we write.
This is just the 1st tab of the Database from the ARC (A-RPG Creator) and not just a simple NPC Creator. The ARC will be distributed as a standalone RPG Creation Toolset (like the Rpg makers). The Curves are just a visual hint of the values per level behind each column. If you click on one of that curves, the Curve-Creator opens, and you can change the parameters for all levels or create a random curve (or a curve based on a start->end algorythm). Clicking the samll tri-color buttons at the bottom of the charakter and face field, the Character-Creator and Face-Generator opens. Means you can use predefined/precreated Charaktersets or create a new one out of Bodyparts.
The parameters and options in the Database are hardcoded, but you can extend them by vidoque (script wrapper language) ^^
At the moment, we change positions and sizes of the GUI Elements, to create a more neat arrangement.
arrr, Espér, das sieht ja wahnsinnig aus! Darf ich fragen, mit was du die Editoren programmierst? Mmit was habt ihr die Icons gemacht? Die sehen hammer aus! Das Frostschild rechts oben erinnert mich etwas an das, des Eismagiers aus WoW bzw das links unten aus Guildwars. :-D habt ihr die Icons größer entworfen und am schluss dann kleinskaliert? Bin auch grad dabei mir gedanken über Skillicons in meinem Spiel zu machen, deshalb die Frage
Seith auf jeden Fall super aus - gibts vom workflow mit eirem talentbaum-editor dann auch mal ein Video?
Der Editor wird... TADAAAAA.. mit der Acknex Engine programmiert ^^ Da ich zu testzwecken keine andren icons habe.. hab ich mir GW und WoW icons geklaut (sry and die Entwickler ^^)
Videos folgen ^^ Aber erst die ganze Database vollens coden
1. i dont think im ignorand, im just buzy and wrote with @ , but ur right, next time i pm if i dont ahve the time to translate. google translator: @ Esper: Oops, I ask myself with Acknex something too complicated to program against, so hats off laugh But self aware that I know the icons on the left below was from me yet known TdP runs from ele-bonders (burning ropes): -> I'm really looking forward to a video!
@ Fogman: Wow, great, this is a great tip! viieeelen Thanks! I had the problem that the icons were washed out right, but with your tip, you can counteract the beautiful, thank you!
It's not aimed at all Germans and especially not tired ones, just the arrogant ones who refuse to keep the thread in the language it was started in time and time again.
be positive: this forum is a great chance to learn German (I put a Google translate to English button to my browser soon after I joined this forum Not the best but works.)
yes ratchet, i agree, the blocky thing is just placeholder and was made from blocks in WED, i hope to get one of my artists to remodel it, which would fix all the blocky areas and the coloring isnt final, true, thanks for the crits
It should be more samurai sword, but i want originality : Borred about all stereo typed games I like original studios sometimes: Suda Team , Clover Studios etc ... and japanese games using totally original monsters : Persona 4, some Final Fantasy titles etc ...
Perhaps i'll turn the texture to some bad evil warrior in a dark armor (it's simple blocking color for now !)
I choosed simple paint, i'll keep a minimalist style and minimal texture. Like Amstrad and old games, the shapes suggested thing, your imagination did the rest I bet more on fun smooth gameplay, special effects (Zelda Wind Walker), and features than graphics.
Aber schon interessant, dass du bei deinem ersten Blender-Modell mit den sekundären Geschlechtsmerkmalen anfängst, dann zu den Primären kommst, dann die Beine machst und das Gesicht fehlt währenddessen immer noch . Und jetzt hoffe ich, dass die Amerikaner das nicht durch den Übersetzer hauen...
Back to English. Sorry for the German excursion.
@Ratchet: Very nice. I especially like the green alien model a lot. Nice basic style.
I just updated my editor, so there is now another 3DGS application with the capability of wmb compilation. Time to deal with the dirty documentation...
Is there any news about the new WED in development?
common package is the common staff for both the A7 and A8 versions, containing system and graphics files, nearly everything, except the exe file and the exe specific dll-s.
so the only difference between A7 and A8 compilations is the exe and the system dll-s, this is why they can be downloaded separately.
I got crash only when I put MapBuilder into a very long named folder and wanted to make a screenshot by pressing space. And another crash situation was during entity placement, when I selected another entity to be placed by a hotkey, and placed it too fast: e.g. pressing 1, left click, pressing 2, left click within half a second. I always forget to resolve it.
And please be careful with wmb compilation, it calls the wwmp2wmb.exe, the editor gets minimized, so you can check compilation result, but the map compiler window must be closed afterwards, because the output files are copied only after this action to their proper folder. Probably later I modify it to be the compilation customizable.
I updated the page, I hope it is better now. Unfortunately the file size is maximized to 15MB by the free host, this is why there are 5 parts. Maybe it should be stored somewhere else.
@Sivan : I'll try the editor next week (busy for now), this is the whole package and terrain editor that is integrated that really interest me for some future casual game using terrain perhaps.
--------------
Alien with final texture in same style as first character ! (Next is animations !)
There`re many really nice projects in the making at the moment. Fireflies looks really good. A little bit like Superku. Nice to see that people are still holding the Gamestudio flag in the sky . We`ve got some talented people here in the community. That`s a good thing. Time for the new WED and some progress in Gamestudio development. Let`s rock A8! Time to kick ass and chew bubble gum. (Or something like that)
Thanks guys yeah, but actually it is called "FireFlight" and will not feature villains. This version of the game is a biofeedback game (for a client) and the user is already 'challenged' very much by relaxing his/her heart. The fireball waves are the most difficult part of the game and sort of a restriction, because the lavapit stage is infinite (in opposite to the green valley- and dark cave-stages before). I reserve the right to make a non-biofeedback game version of the game and if I do so in the future, I will certainly put villains in it like crows, bats and the like.
Good suggestion , specially for outdoor levels ! I can't imagine how must i would loose making some complex level with BSP compared to use some modeler with all advanced extrude , loop cut, loop select, ring select etc ...
Thanks I have some little holidays, so the game will be on pause. I don't have deciced on characters for player ... but they remain in the same style !
I got an infinite loop of coding-debugging-learning-coding-bugfixing-going crazy-bugfixing-giving up-learning-coding-success-new ideas-coding-failing-learning-coding-bugfixing, so now take a rest and do some low poly modelling
Skyrim alike inventory system with full customization options (color, alpha values, fonts, sizes, ...), categories, item preview and mouse/keyboard control
Next things todo: - Item Event System: -- Equip -- Use -- Drop - Information Window (to get a short description of the item and see two addition informations)
EDIT: Forgot to say: All sizes are screen size adapted!
The background is only a screenshot of skyrim, but looks much better with this in background
When you have fun with MDL stuff: Maybe you could write the requested Blender 2.61 MDL exporter. That would be very useful for many people here I think.
You could take a look at Ventilator`s Python script. I think the best way would be to convert it to the new Blender plugin functionality, but I`m not into this at all and only know the advantages of Ventilator`s exporter. It`s impotant that objects can be exported / converted individually and not only the whole Blender file with all lamps, model objects and stuff. Cool thing that you`re trying to help us poor Blender modelers !
Well not shure, I would start with simple objects, chaning parts and looking wich part of the binary changed. I think the structure or approach to pack data will not differ so much from MDL5.
look into my blender exporter. it's python source that writes mdl7. should be easy to figure out.
the .blend format is a complicated memory dump of blender scenes. but there are libraries for reading it (i think somewhere on the bullet physics website).
writing an exporter in python probably is easier and more user friendly though.
(when i tried to export bone animations for the first time i couldn't figure out how to calculate the matrices correctly so i first wrote a mdl viewer in python that was able to play back bone animations. it was fun and it helped a lot in understanding what's going on.)
I just updated my homepage with a new version of MapBuilder (now terrain deformation and saving is really fine), wrote download/install instructions, and also added some new screenshots: http://mapbuilder.zxq.net/screenshots.html Due to Google Analytics the winner is Germany with 35.84% of webpage traffic, followed by USA, Iran, Turkey. Unfortunately nobody from 2 continents: Africa and Australia .
now my high level pathfinder works lightning fast (so optimization will be possible even if that makes it slower a bit). time to get trouble with the low level one
those are showing maximum unit size that can access that tile. orange= no access, blue= 1x1, green= 3x3 tiles. other= 5x5 (this is the max size at the moment because of speed reasons) they will have role only in the low level pathfinding, when units have to find their way very fast, depending on their size (and their other capabilities).
How do you structure the "large distance" grid? Is it simply regarding not 100% passable areas as non passable? Or do you calculate a passability value for each "subgrid". (you could make a short-distance pathfidning-testrun from each of the 4 sides "side-middle"points to each other 3 middle points to check if there is a connection) Or use a floodfill to see if each side is beeing touched...
each tile has its own type and cost value. first clusters are evaluated based on percentage values of their tiles. and then the cluster type and move cost is changed, or set as further processing needed.
there are cases when the cluster area is subdivided: e.g. if impassable tiles are over a level within a cluster, a flood fill checks their connectivity. if they produce a coherent area, they are cut out of the cluster, and form a new one.
walkability of neighbouring clusters is also checked, by checking all the border tiles and their out of cluster neighbours, thus determining neighbouring clusters (ordered as walkable and non-walkable)
I was a bit chaotic, because I wrote this part before summer, but this is the main process. it works at the moment, but not totally perfect, so some further tweaking will be necessary. especially when I will add buildings, which will also form separate new clusters. now only obstacles are handled.
@Dice : Could you post a video gameplay ? That's great , respect for finishing a game project
Now my taste and bad talking : The level looks very sad and grey, dark ? Why not making it alive, more green clear grass, more better backgrounds like distant islands, moutains etc ...
That looks massive rondidon. Last posts show some reat quality of everyone! A please to watch . This thread was a good idea!
Here's my graphics combo breaker ^^.
Not much change visible with the eye, but somewhere I decided to port it over to C++. Embedding 3dgs, using Eclipse. Works great so far, but a lot of work. The engine has a loose coupling through interface. Theoretically I could switch engines without the need to touch the entire game logic. I plan to make the game playable on mobile platforms as well using Mosync. Drawing lines and text is the only graphic requirement... in theory.
Long long time ago rrrecap!
Originally Posted By: HeelX
@Joozey: Your game looks so awesome. Guess what? This lead me to the wikipedia page about airports and airport infrastructures; have a look, it's darn interesting to get to know the real stuff, that drives the real world - you might enjoy it, while getting a flood of ideas for your own game. At least, I had for a theoretical, rather realistic tycoon'eqsque game I am never ever gonna make smile --- best wishes!!!
I'm really happy to hear my concept triggers people to look into stuff . I got most of my "airport" knowledge from the game Airport Tycoon (2), or Airport Inc. That game needs some work especially on interaction to make playing enjoyable, so I plan to make my game solid on that part at least. I wonder what this tycoonesque game is you talk about . Thanks for the wishes!
The project is challenging, has potency in my eyes, may target a wide range of platforms and does not require much artistic skills. I might stick to this for some more time .
Joozey, this project does sound interesting! However, I find the visual style to be almost... too busy. I may be alone on this, but the high amount of lines makes it look somewhat confusing. Could you try increasing the grid size, perhaps? Instead of a grid, you may also try to only "suggest" a grid, by using an alpha channel (or even just different contrasts) that is very strong where two lines connect, and less so inbetween.
I dunno. I mean, I like the kinda elite-like throwback? And perhaps it's not so much a problem in motion. But I think it might be worth a try!
Though, who knows how much of that is placeholder art. In that case, sorry, and please disregard the comment.
Your way of working sounds pretty cool though - being able to switch engines on the fly, so to speak. Neat work! ~~
Can I upload a super-boring shot that no one will comment on?
I certainly can!
Think of the children. Report offensive content.
(FUN FACT: To have this option avaible at all times is a requirement or otherwise this game would, according to those FSM guys (german link, be warned) be only for ages 16 and up! Laws.)
Joozey, this project does sound interesting! However, I find the visual style to be almost... too busy. I may be alone on this, but the high amount of lines makes it look somewhat confusing. Could you try increasing the grid size, perhaps? Instead of a grid, you may also try to only "suggest" a grid, by using an alpha channel (or even just different contrasts) that is very strong where two lines connect, and less so inbetween.
I dunno. I mean, I like the kinda elite-like throwback? And perhaps it's not so much a problem in motion. But I think it might be worth a try! laugh
Though, who knows how much of that is placeholder art. In that case, sorry, and please disregard the comment.
Your way of working sounds pretty cool though - being able to switch engines on the fly, so to speak. Neat work! laugh
Agreed, especially in the back lines tend to get very busy (if that's what you mean). I could add transparency on lines, but I am making draw calls for each entire row and column, not separate drawcalls per gridcell. Otherwise the fps drops to 1 . So transparency would have an effect over the entire width of the field. That picture is zoomed out as far as possible. If you get closer I think it's less busy, but then you have less overview as well (naturally). Can't change the width of draw calls, sadly. What I can also do is try other color schemes, and see if I can lessen it. The grid could blend in with the background more, perhaps that fixes a lot already.
@Dungeon Deities: Neat intro screen! The style you present is great, and I guess also relatively easy to create, even for programmers.
@MasterQ32: Can you give some background? What does it do? Does it really merge entities and save them as one single entity to be loaded when the level is played back? Because that's an awesome feature. What does it do with multiple skins, how does it handle them?
atm i'm merging the models in runtime, i do not save them to a file merging contains vertices, faces, 1 skin per model and no animation but maybe i can add multiple skin support, but i don't know how...
There is no objetive at all. I always liked to play with randomly generated things as terrains, roads, mazes... I think that random generation it is not properly exploited into games market. It could be a great inclusion for infinite playing to those loved games
If someone want to take a closer look, here is a compilation: download
@rayp if I'll be ready with the simplest version of low level pathfinding, it definitely will be immediately integrated with unit movements, because it is essential to see bugs. when it works more or less, I'll make a video.
Rewrote my dynamic lighting system (not the mostly model based 2D lighting from my side-scroller, but the one I've posted in this thread a while back) that now supports multiple lights and IMO looks much nicer. There is still no additional rendering involved despite the camera view and no postprocessing, it's an object shader (PS 3.0) applied to WED level geometry. It continues to use line segments (simply defined by two points) to perform the shadow calculations, so the shadow for a cylindrical column can be realized with one line (that is updated per frame), a rectangular pillar (see screenshot) with two line segments (thanks to Hummel's Twilight Box for the idea of using two lines diagonally).
The ground has a simple dynamic texture shadow system that illuminates the tiles in respect to a heightmap and the light(s) and draws basic shadows (don't know if you can call that bumpmapping or something like that, I don't have experience with that stuff). I can upload a demo/ video if you want.
Not too long at all Thanks! I really like it. The floor self-shadowing effect is really cool as well! It's a really convincing effect. Well done! You know you have to make an uber-pretty first person game now, right?
@rayp it is integrated to my MapBuilder editor, and no one knows when it will really work smoothly, and for what purposes can be used. the aim is to handle group formations too, but maybe only a basic group handling will be achieved with limited unit quantity. I'll decide later about making it open source or not, depending on the quality of the result.
FINALLY, I got my CEGUI plugin doing what it is supposed to do. Working. After fighting through a swamp of dependencies and input incompatiblities of the acknex engine (btw mous_valid is not updated in the pluging sdk) I managed to got the xml layout and configuration thing running. Was quiet some work to figure out how to bind events to the elements.
Here a screenshot. It is not much, because it is work in progress. And there are a lot of bugs and unsave code I have to remove before I can put some nice demos together.
@jenGs: That is really nice! Is it a standard dll compiled against the Gamestudio plugin SDK? --- I am also not familiar with CEGUI, is the rendering resolution independent? Is the skinning bitmap- or vector based?
When it's ready and bugfree (mostly) I will share it. For now I will create a video tutorial how to implement CEGUI into your project yourself. This way I hope I will get some people to help me with some problems I have with the implementation. (UNICODE INPUT)
Quote:
s it a standard dll compiled against the Gamestudio plugin SDK?
Yes. It initialise's the system and loads the design and injects the acknex input.
Quote:
is the rendering resolution independent?
it can be scaled automaticly. But it is not vector based, so you have loss of quality if you are using small skins. It's skin based.
The system has a render library for a lot of platforms (opengl Directx 9 10 11 ...) And special systems for specific engines. To init the system you hand the d3d9 device to CEGUI, than you have to put the render procedure to the acknex render_layer event.
The advantage of CEGUI is, that it is complete and has a nice gui-editor. And I like the xml feature a lot. That create's the possibility to create real model view controller systems.
Edit: Disatvantage - it has a LOT of third party dependencies. They are all free for commercial use. But you always have to check if this change's, when you load the newest version.
W O W! i wish i could draw like you this is no joke. And i like the characters especially the zombie scientist. A great design ! I cant wait to see this in 3d.
Ratchet (or somebody else) iam looking for some low poly cartoon style models, like that one above this post (done by ratchet). Perfect would be with some simple animations, like walk, attack, idle. Someone around here want to share his/her (old) models ? Or a tip who sells such modelpacks. Looking for a "cute big head style".
@Rayp : Just send me some sketch of what you want and approx number of polys (2000, 5000, 10 000 ?) ? what sort of texture any game screenshot that would show it ?
Some wip of the re texture on my thief character : (I know the Uv are not optimized and done not well)
I wonder why medieval thiefs always wear hooded jackets... if the hood covers your head, your auditory and visual perception is very limited. That is not so helpful, I guess.
Heelx: I'm not sure thiefs that would need to survey around them would be dressed on the head like that. In the western cowboys days some just covered below the eyes with something , and in the ninja days they had only eyes visible. This sort of head cover would be moe for some easy dicret operation And its non realistic ennemies for some casual Arcade RPG !
@Damocles : wow, great style and design , i like a lot your cartoon style, i think i'll go this way. ( you can find games using less pronounced lines and they work also with plain colors : http://www.youtube.com/watch?v=mjR0FO7j1KA ) @JibbSmart : my low poly style is very very simple, even you and other people could do the same models very easily
Yes I pronounced the lines a bit stronger than initially wanted. You can also leave out the shadwos I drew around some lines, and draw the contours with a less contrasting color.
The good thing about outlines is that they define your Model clearly by showing the elements its "made of". Making the Characters well defined when they are not big on the screen. (good contrast)
And making the contours them out of relatively straight lines fits very well with the low poly approach.
Best of all: the workload per model/Texture is quite low, and managable without having strong artistic 3D modelling experience. So you can get your assets done in a reasonable time.
after high level group pathfinding, group order driven low level unit pathfinding is working (stupid version) with some error handling. but the nasty things are just coming (unit & group behaviour and movement loops to handle both movement and new path searches)
Finished the texture for thief class ultra low poly (I could put black lines on boots, some head parts and other parts, but the UV wasn't done well on that parts so black line of one pixel turned too much stretched.)
cool, looks more defined now. Maybe you also Highlight the eyes too. You could experient using a darker blue line there, so the black line is not too strong)
also try to keep the resolution of the skin relatively even over the model (bigger polygons using more skinarea) - less stretching. This looks better with the very low pixel-res style, making the resolution even over the model. (->mix of pixelart and 3D modelling)
Yes i tried balck outline on the eyes, but the result is bad ! I'll test some darker blue. , yes the eyes don't popup enought and that what's sould popup a little in characters !
For the resolution i'm stuck with 128*128, i have done the UV with some quick unwrap, no real work on UV to optimise them ! So it's somewhat to late now. But i'll do it for next models indeed
@heelX, the guys who just stole my friends' bikes wore hoods. It's not just medieval thieves, it's pretty realistic today. The reason is that you'll be less likely to be recognized, I'd guess. You're not planning on being seen, but in case you are, it's no good if anyone sees your face or hair.
@losticmate : it depends on your cartoon style in fact
It's not like people targetting realistic look where lot lot lot of things must be respected to make it look life like ( mass effect characters etc ...) Yes low poly is really easy and fast, and fun caus you don't care about details , proprotions, you can express your imagination.
Another way is making high resolution cartoon sculpting (look at my next zombie posts) , than retopology and bake normal map for texture guide ! Sculptris is the king for that
@heelX, the guys who just stole my friends' bikes wore hoods. It's not just medieval thieves, it's pretty realistic today.
Yes, if you are stealing around noon. But at night, when you are sneaking into houses and stuff, .. whatever
@aztec: Nice, it is pretty dark, though, I can't see the ground. Maybe you can add some lightning from the moon direction to the surface, so that you have a thin line of moon shimmer on the ground? Like this:
For the shadows, they are CPU/GPU intensive. those actual simple blobs are really ok. What could be done would be make them less transparent and more noticeable like in some Iphone games.
@Joga: Great sculpt (yes sculptris rules) ! Your retopology seems very hight in terms of polygons for the low poly version ? will it be a giant fly or a big boss ? caus such level of details should be for camera very near and perhaps it won't be lot of times where the player will see it closer ? At some distance from camera, lot of polygons won't bring definition on your model, you can optimize. Anyway great model
thanks for your answers, MrGuest, rayp, ratchet! And thanks for the tip, ratchet, that's a great remark. I planned the model for a normal enemy, but also for a boss. So I think, I use this actual model for the boss and reduce it further for the normal enemys, later. Skin is not finished, but here the state of work
edit: for the wings, I have copied the polygons and flipped the copy for the two-side-effekt. The wings are just brushes I used a brush in black and white for the structure and after this I selected the structure(selection from alpha-channel), increased the selection (1-2pixel) and filled it with two different color-Gradients (toolbox->blend in gimp). For the question "have you assembled them to the model in MED ?" you have to excuse me, I didn't understand this question :-( Normally, after all, I export the models as fbx and import them in MED. It worked with my last model, so I hope, it is working here again
A market and a great statue of who might be the saviour/a hero of the city or a grand ancestor . And the little main character dreams of becoming himself a statue too
Octree for polygonal collision detection in KarBOOM.
The transparent terrain is what was imported into the octree. What you see is cyan polygons which were contained in the magenta nodes, which are all the non-empty nodes that intersect the yellow AABB.
Two days ago, after playing the A5 mission demo, I had the idea of creating a game with old school graphics using only textures from standard.wad. This is what I I've got now:
does anybody know this 2GEN game editor? it was used for making Supercan, and the upcoming second part Supercan2: http://www.youtube.com/watch?v=131rjsgtkdc my Turkish is not perfect (=0)
Sivan : just amazing That's really how A9 should be Powerfull workflow for people to create the basics of a 3D level ! Could not Conitec contact them and find some partner agreement ? Conitec could get this amazing tool for A8/A9 and all customers perhaps ?
absolutely agree, seems to be a real-time all in one level editor. unfortunately their site is Turkish only... http://www.2genoyunekibi.com/ I put a post also at 3rd party tools, because I was asked via an email to do it. It came maybe from one of the authors, so I asked a few questions, and will share it when I get any answer.
I think some of us perhaps a good bunch , would be ready to pay for such prooven and complete level editor with such amzing tools
I have 3DGS commercial and would be ready to pay if it was sold by Conitec separatly for some 60$ to 100$ max ! I agree it won't be as cool as some other 3D engines having all in one , but would be very great, while it could be integrated in some future 3DGS version like A9 ?
Sivan : just amazing That's really how A9 should be Powerfull workflow for people to create the basics of a 3D level ! Could not Conitec contact them and find some partner agreement ? Conitec could get this amazing tool for A8/A9 and all customers perhaps ?
I would not like it being a core part of 3dgs. It would be cool as a framework like shade-c. But what I like about gamestudio is that I can create Editors like that myself with gamestudio. And if you think about it. Its not too hard to create an Editor like that. It's just work. And making such an Editor perfectly fitting for your is often essential too and normally you do not know enough about such a framework to easily modify the parts you want to change...
But what I like about gamestudio is that I can create Editors like that myself with gamestudio. And if you think about it. Its not too hard to create an Editor like that. It's just work.
I agree with 3DGS flexibilty, indeed it's a big strenght !
The problem , is that you won't find such 5 or 20 editors made with 3DGS as complete and ready to work !!
And some of us are not programmers at all, or won't spend months trying to make one.
There is one in the makings, but i really prefer this one a lot and all it's features. Visually it contains all tweakings i could search for shaders, paint, fullscreen effects and objects placements.
We had lot fo threads like : This great terrain editor can be easily done with 3DGS etc ... ok , but there haven't been made any time. Sometimes, it has been something begun by a programemr than he abandonned, too much time consuming, bugs etc ... borring sometimes... or simply no more motivation !
And we wait for a terrain/vegetation editor for more than 2 years perhaps or more in the forecast page! (not ot be harsh, but if we would put the date of the post of a new feature on the forecast section you would be surprised about the time you wait or have wait for some of them ) ! But i'm ok with that, no problem i don't target terrain based game with A8 and some people perefectly do things manually with MED or any other tool for terrain without a real time terrain editor !
Ok it could be done , but some of us , would just be happy right now with a working tool
The video just shows the editor is here with all features demonstrated , some of us i think would be ready to pay to have it indeed. That's the difference, the matter is not talking about it could be done, we don't care .... the editor is ready and working and some of us i think would pay for it !
Sorry it's WIP thread, we should discuss about taht on the tool section !
I don't want to have to code an editor in order to be able to work with Acknex properly. Period.
I'd say you don't have to, but this'd be a fun discussion to have elsewhere. Let's not this derail this thread. I like coming here and seeing screenshots.
The SüperCan editor is great, but specialized to this gamestyle and the underlying architecture. Its a typical workflow tool for a specific project. Its also not usable without the underlying SüperCan codebase.
Generalizing it to suit more gamestyles can blow it quickly out of proportion. Because then people also want to have their "actions" and "skills" and import options, adjusting all kinds of parameters, and lots of other stuff.
Wont work. Either people need to have more restrictions to keep the workflow easy, or have the skill (and invest the time) to adjust the code to their needs. This editor do everything but easy idea will not work. (Unless lots of $$ is invested wich is only profitable in a big engine)
In many cases an editor is not even needed and things should be shifted to code or ini-files. -> if you plan to make an actual game.
There are also often requests by artists who dont want to code. In this case it would make sense to create a specialized "scene stage" Editor, wich is primarily used for setting up a nice scene (Objects, Foilage, Shaders, Terrainediting, Camera-movement) and enabling the user to wander around in it. Witout all the little restrictions that a propper gamescene requires.
(-> The editor should have a fixed set of effects included)
Agreed. I currently have an editor I designed for my own project. But the code is so sloppy and its so specific to my own game that releasing it would just be useless. I do however plan on releasing the editor as a map editor similar to Halo's "forge" mode.
- Edit terrains, save edited terrains as *hmp - Painting directly to the blendmasks - creating shadowmask with / without static objectshadows automaticly - creating shoremasks automaticly - creating colormask for terrain lod automaticly - Load entities with filedialog (not shows up with fraps?) - a seed function (for plants and others) is integrated - give the ents a action from a dynamic list (pulldown menu)
About a weeks worth of work so far, still working out a few kinks.
Its a multiplayer FPS (and yes, I finally learned how multiplayer works), called Vexer(name subject to change). I am going for a happy medium somewhere between Halo and Unreal Tournament, but with a bit of my own personal flavor. I only have 2 weapons done for now, but let me know what you guys think.
Haha yea, I just leave it like that for testing. But I was wondering, what is a good setting? I usually use 80, but are there any other "recommended" settings?
@exile Yes the most recommed fov is 75. I hate games like CSS where the fov is 90 o_0 BTW: The game looks good, textures are good styled and have same style like your weapons. The walking bob is cool. I like the explosions, from where you have them?
Finally the conversion starts to work. Had to built my own draw_line3d function with the 2d variant that comes with mosync. I keep it isometric, probably kill the fps if I'd start calculating the projection.
About a weeks worth of work so far, still working out a few kinks.
Its a multiplayer FPS (and yes, I finally learned how multiplayer works), called Vexer(name subject to change). I am going for a happy medium somewhere between Halo and Unreal Tournament, but with a bit of my own personal flavor. I only have 2 weapons done for now, but let me know what you guys think.
left click: select group right click: set target and group order then find high level path then set unit order to move then get low level path periodical updates for dynamic obstacle avoidiance
FOV is a matter of opinion, i hate it too low or too high, 90 makes me feel as if i'm a spectator of the world rather than in it.
Personally, I think FOV affects immersion also, smaller FOVs like 75/80 make you feel closer to the world but reduces the overall view of a scene, its a tricky balance, your FPS seems ok as it is though
@3run : in some cases, because the movement and other update loops are not well scheduled yet. but it will be eliminated as much as possible, keeping collision quality and cpu load in balance (the higher unit quantity I want, the lower quality can be used).
@DarkInferno : I really like the style of the player hand and the level mood It seems strange fullscreen effects on screen or it is effects on player view like water drops ?
Can you give us some informtion on the game type/style ?
@Chaos Coder: The blood's sprites themselves are mine, but the effect is made in PEB (which is an amazing tool), then refined in engine.
@3Run: The blur is just the standard gamestudio blur effect which is being affected by the mouse and other game variables Check out this page as its where I got the base code (thanks alibaba).
Just tried the multiplayer and while it works, its very buggy. I will keep refining the engine to make this work. I ultimately plan to release this game along with an actual effective 3dgs multiplayer template when its all done and working
Looking great Darkinferno! Definitely a shoe in for the competition/contest! Oh BTW, could you clear your inbox? I have been trying to get ahold of you hahahaha.
That's great DarkInferno Lot lot of good ideas ans stuff.
Now my taste and preferences : -Bigger hand ? (two hands also ?) -More animated hands when running/walking ? -It's a demo but why not improve : animated and visible powers : i don't mean Skyrim like , but more Visual attractive and visible, more on screen. -Hands moving with fists like when you run , you don't shoot at same time ? Or can only use powers when walking or strafing ? - Normal map shader ? it will bring a lot in visual and seems really good for that type of game not so demanding on polygons ?
I like the concept of defend game , the powers and the world.
everything you mentioned is planned, we ran out of time, plus we had limitations since the HD3000 isnt the most powerful video card ever, this isnt a DEMO, its a proof of the concepts behind WrathOfAngels
nice idea, darkinferno! I played it and it was fun! Two thinks I meantioned: Is the camera position at start of the level on a far point of the map? At start, the game needs some time and the Life of all towers are shown - so I thought maybe the camera could look in a direction, which does not include the whole level? The second: Maybe you could create a kind of "radar", which shows the position of the remaining enemies? Your idea is nice, I am excited, how your game will grow.
Just finished an other model nearly, a kind of mausoleum, the middle of the graveyard next to the start-village and the entrance to a dungeon:
@Joga, nice Model so far. But you've wasted too much polygons in my opinion. Ive made a very basic and simple example to show you how to reduce some polygons... Dont get the details like your one but if i add the "details" to this Model, the polycount shouldn't be over 700 - 800 (Just the half of your count).
heres my "example" (fast and ugly / 368 polys so far)
@Esper... Finde den ansatz deines Artworks ganz gut, dennoch würde ich eher größere Flächen die so zerfressen sind bevorzugen als viele kleine, wie du es zur Zeit hast. Das sieht für mich zu "pixelig" aus. Schau mal wie es aussieht, wenn du wenige Große mit vereinzelnt kleinen Flächen so zerfressen machst , anstatt so viele kleine.
@MPQ - Thanks, absolutely, but first I complete my editor to be suitable for fast level creation suitable for RTS games, mainly for battles (I started its development because of the special requirements what WED and other game engine level editors do not offer, or what they offer is too complicated for me to understand, because I'm just a hobbyist). Not an easy task, but hard but fun. These tests are required for the further development as a proof of my ideas. If this basically 2D pathfinding works well, I can improve it to include climbing ladders to walls, using multi-floor buildings and stairs etc. I have the plan and some working codes, and now it seems to be really achievable The pathfinder requires little modifications only. The hard one is the group movement and behaviour. Maybe I should read more about flocking, but there are more simple solutions too (e.g. follow the leader(s)). Anyway, the pathfinder is suitable not only for groups, one person groups are also okay, a few hundred should be handled without problems even on my old pc... I will test that too.
if things are going well just like until now, and I can achieve some improvements, if someone is interested, I can send a small package for trying/testing it. (MapBuilder is needed, just the exe should be overwritten, and by some tricky hot keys everything is accessible. And I should finish my really low poly animated character with LODs for better performance, to replace the current infantry character got from a game.)
You are going in some special game type, the characters seem already low poly in the screenshots ?
AAA RTS games using lot fo characters having programmers that have made specialized and compiled C++ routines for it. And using instancing sometimes also.
But yes with special tricks you can achieve good things with A8.
Really impressive is the Dead Rising Engine (org. made for Sportgames). It is some sort of special instancing. Unbelieveable what the programmers did there.
programmers that have made specialized and compiled C++ routines for it]
Well you make that sound so mystical. They also just boil with water. Its mainly just lots of manhours put into it an sticking to well chosen designpatterns. (The brute force way)
I'm definitely unable to make an AAA RTS But I try to achieve an acceptable unit quantity beside enjoyable graphics quality, by simple methods, like very low poly models with close lod distances, restricted camera, and balanced function calls per frame. Unfortunately only 3dgs Pro offers instancing, but only for sprites and particles. And I'm not planning to buy it, just upgrade to Commercial. Terrain LOD would help a lot, but I was not satisfied with it when tested with the Trial version...
@ rvL_eXile: Thanks for the hint. My model has 960 polygons. Well, after your post I recognized some, I could spare, you are right, thanks! I made some models, but couldn't include them in my game (my game is on my PC in the town, I live to study, now in holiday, I live at my homevillage, so I had a look in blender on my laptop) result is not THAT good, but I am confortable with it, because it's my first try with blender rendering :-D
edit: added some more stuff, torches will come next days:
I will try to bake shadowmaps in blender and use them in 3DGS, never did this, don't know if it's working, but I want to test it next days.
Nach den Überfall auf Andrea Rossi und seiner E-Cat, werend eines Experiment, sind mehre Kathalysatoren geklaut worden die es möglich machten eine Kalte Fusion zu erzeugen.
1. Suche nach Hinweisen um den Dieben auf die Spur zu kommen 2. Schau dir das Überwachungsvideo an 3. Spreche mit Prof/Dr Ellen
...
models from: -> Free Resources and Model CD for A5
So far you see at the end of every round who has the most points, who killed most and who died most. I nearly shows all data of the round, i.e. How long the round took. I greatly improved the server log, as you can see it now shows every command i do. Thats everything at the moment. I'll now add some buttons to the menu.
@Zeusk : I like it a lot , very original and great colors Will it be on user contributions ? i know a project that could use it if you don't plan a game using it or selling it ?
This will probably end in nothing, as I'm lacking of ideas ofr the levels, unless someone wants to help maybe with some sketches or designs about how the maps could be
painkiller, maybe, you could add some more colored lights: red and green for entrances, and bluish for the screens, more yellow for the wall lights... at the moment its veery grey. A tip for more level ideas: broken walls which lead in very different areas, dead end rails, storage rooms with 'towers' of old machines, boards that connect them ... ...stairs, even a few steps make a difference, windows, bars where you can look in further rooms and floors or yards with some sunlight...
@Painkiller : Please change this very very old door texture (I know it's a fun project and using old WAD) !
I'm playing Mass Effect 3 : Level design is so much far from the Mass Effect 1 and 2 , that's just marvellous, can't imagine what they'll do for Mass Effect 4 and some new Hardware power ? Each level was amazing on details and effects !
after one week of insomnia because of my newest baby, I had some time to improve my pathfinder: - smoother animation due to modified unit loop (okay, it is not sensible on these screenshots, sorry) - basic follow-the-leader behaviour that resulted in better FPS, and much better terrain bottle-neck passage behaviour (the upper groups are just slightly scattered after passing the narrow area in the middle - it was a pain for them before). I think this follow-the leaders feature will be the basis of formation control, so after scattered state the group can be reformed again hopefully
Congrats on your fatherhood! Does following the leader mean only one entity searches for a full path and the others only do close-range pathfinding? If so, what happens when there's small places that they need to get through, does the leader take into account that there's more entities than just him trying to get through? Looking forward to seeing the changes in motion!
Speaking of follow-the-leader: Have you tried what happens when one entity is moving around, pathfinding, and another group of entities is set to go to him? Sort of like a moving goal. This would often be encountered in an RTS situation, when players tell their units to attack a moving unit...
+1 . Experimented with some basic transparency (based on a transparency bitmap) some time ago. I`ll try to find a shader-based solution for this when my modelling work is finished and there`s some time left.
@Rondidon: Sieht super aus. Das einzige, was mir negativ aufgefallen ist, das sind die fehlenden Spiegelungen vom Schilf und von den Felsen im Wasser.
Noch ein paar Verbesserungsvorschläge: 1.) Die Steinmauern würde ich etwas höher machen. Die Proportionen wirken etwas unnatürlich, vor allem im Vergleich zu dem sehr hohen Turm. Niemand würde einen so großen Turm bauen, und so große Tore bauen, und dann bei der Höhe der Mauer so sparsam sein. Die Mauern sind so niedrig, dass man eigentlich ohne Leiter rüberklettern könnte. Ich würde die Mauern fast so hoch machen wie die Mauer bei den Toren.
2.) Der Innenhof rechts neben dem Turm (auf dem 2.Bild) wirkt etwas leer. Du könntest in der Mitte vom Platz entweder eine Grasfläche machen (zum Beispiel in kreisrunder Form), oder einen Brunnen bzw. ein Wasserbecken (für einen Springbrunnen).
3.) Und ich habe noch nie eine Parkbank/Sitzbank gesehen, die so aufgestellt ist, dass die RÜCKSEITE direkt an einen Weg oder an eine Straße angrenzt. Normalerweise zeigt die Vorderseite zum Weg hin. Weil niemand würde sich auf so eine Bank setzen, wenn hinter dem eigenen Rücken Menschen vorbeigehen, und man nicht sehen kann, das hinter dem Rücken so passiert. Da würde sich jeder unwohl fühlen. Also entweder stellst Du sie auf der anderen Seite vom Weg auf. Oder Du platzierst hinter der Bank eine Hecke, eine Mauer oder einfach nur eine Rasenfläche, die zumindest 1 bis 2 Meter breit sein sollte.
Thanks for your suggestions. I really apperciate them. The water is still to be improved. The missing reflections is some scripting mistake and not a big thing. I`ll fix this.
@Germanunkol: The group has an imaginary leader in the middle of the group for high level pathfinding (areas), for low level pf (tile) the first line of units are leaders, others are moving relatively to them, but taking into account tile accessibility too. Thus I can later set group orientation and different formations by assigning leaders and desired leader positions. This is the plan atm they have constant orientation. I'm just imrpoving group collisions: in some cases (when 2 moving groups are crossing each other) 2 units could enter into the same tile, so a 3 stage avoidance is appliad at the moment (1-turn away and slide on each other, 2-new path towards another tile, 3-set unit to passable temporarily if nothing helps). But later it should be improved with a unit priority based pushing system... (e.g. because of siege units, vehicles, or cavalry) But in reality crossing of group of soldiers, or changing formation also resulted in broken formations, thus sometimes resulted in loosing the battle, so I think it should not be perfect and easy task in a game too. Of course, it can be annoying if the solution is too stupid. Targeting is group based, updated in cycles, so following a moving target hopefully will not be a serious problem. I also hate when it is stupid, so that will require some enemy path estimation (e.g. by stealing a few steps of its path).
What is so special on Harrys water ? Its just a photo
No, it is not just a photo. It is a combination of semi-transparent texture layers and a shader. And it was not so easy to 'paint' the underwater (ground) texture, because I had to remove all reflections and the visible surface of the water (waves, bubbles) from the photo.
@Fogman: Wow, das ist ja super. Sogar mit einem meiner Fischmodelle und dem Rio Ebro Waldlevel auf dem Cover. . Kanns kaum erwarten das Spiel im Mediamarkt zu sehen.
Was mir negativ aufgefallen ist: Die Screenshots auf Simuwelt, Weltbild und anderen Shops sind schon anderthalb Monate alt und kommen aus der Pre-Alpha .. inzwischen sieht das Spiel besser aus (vergleiche Screens vom Gardasee-Level auf Seite 129 hier im Thema). Vor allem das Terrain, die Schatten und die Beleuchtung bzw. Farbabstimmung, aber auch das Wasser. Kann man da noch was ändern und am Ende dann finale Screenshots an den Publisher übergeben, wenn ich dann mit meiner Arbeit fertig bin? Nicht dass das jetzt eine Priorität wäre, aber das gäbe halt ein stimmigeres Bild im Hinblick auf den Ersteindruck.
Quote:
The character looks a bit dead/bored to be honest
I second that. I didn`t make the character model, but I`ll see what I can do to make it look more realistic if Fogman allows me to
That cant be from a computer. Its missing the computery stuff. Every computer picture has some computery stuff™ in it. I really cant find any here.
Here is your 'computery stuff':
Wie man sehen kann, besteht die Mauer im Hintergrund noch aus flachen Sprites. Die Betonung liegt aber auf "noch". Wenn der Level mal fertig ist, wird man wirklich keinen "computery stuff" mehr sehen können.
Thank you guys for kind words. @Superku, yeah I know that mate, but it looks odd for those who don't know about Doppler effect @chris_oat, blood was made with particles. There is only one smoke sprite, behind the smoke particles.
looks really nice, rvL_eXile! What about some junk and concrete-chunk at the base? And will there some windows be in the top or should it be a water tower? I will stay tuned, great job!
my navmesh plugin, now working on path management for multiple agents for my demo ,i had 10 guards at one point each generating paths but without avoidance and still got around 512 fps at avarage. nothing big yet but i have been asking around on mikko's recast discussion group as to where find the new source and will be upgrading my recast source from 1.4 to the newest which has avoidance for agents ,crowd control and other goodies ...
3Run i always learn something new when you make something even if i reasearch it . Sivan the basis of your code has to be good as you show rapid progresion. Yes its been up here a while in 3rd party tools but under developement stil
Hab leider viel verpasst in letzter zeit. Aber sofern ich lust und zeit hab, zeig ich euch, an was ich gerade so arbeite.
@rvL_eXilealias "das Ding Vater": Du bist wirklich ein begnadeter Gebäude, Landschafts und Struktur designer. Muss man echt mal sagen. Ich versuche mich gerade an etwas, dass weniger meinem und eher deinem stil entspricht. Daher wäre es echt cool wenn du mir ein paar ratschläge gibst. ^^ Allerdings will ich nicht verraten was es wird. Stattdessen zeig ich dir einfach ein pic davon: Who knows what that will be?
Danke Otter, sowas hört man gern... Also was mir gerade spontan zu deinem Bild in den Sinn kam, war so ne Art mittlerer Osten Basar Gebäude oder sowas. Der Bogen vorne auf der Ecke und oben drauf erinnert mich irgendwie daran.
Der Kiel und das Ruder passen nicht. So ein Schiff könnte man niemals steuern. Das Ruder muss sozusagen der verlängerte Kiel sein, an dem das Wasser entlangströmt, und dann am Ende umgelenkt wird. Dein flaches Heck lässt es gar nicht zu, dass das Wasser am gesamten Ruder vorbeiströmen kann. Das Ruder wäre daher wirkungslos.
Lösung: Der Rumpf muss hinten abgerundet sein. Der Kiel wird zum Heck hin immer höher, und hat schließlich die gleiche Höhe wie das Ruderblatt.
Das Ruderblatt selbst sollte so aussehen: Ruder Die Ruderpinne (der lange Hebel) wird durch eine meist quadratische Öffnung am Heck ins innere des Schiffes geführt. Das Steuerrad vom Schiff sollte dann direkt über dem Ende der Ruderpinne liegen. Weil am Ende dieser Ruderpinne waren Seilzüge befestigt, die durch das Steuerrad bewegt wurden (die Seile wurden an der Achse des Steuerrads aufgewickelt).
Find ich voll Lässig das du dich da auskennst Harry =D Danke auch für die pic's, die sind viel besser als meine. Hier die glätung des "Boothinterns" (wie vollprofis wie ich ihn nennen):
This is supposed to become a community contribution, esp. aimed at the beginners as it is very easy to use: Simply set up WED paths and #include "pathfinding.c" and you are ready to go, there is no manual path calculation or entity setup necessary, everything is handled automatically. There are a few functions, but you can keep it simple and only use
If you want to move ent to vtarget, you can call this function in a while loop and it will calculate the position where ent has to move to to reach vtarget. Example:
Code:
action enemy()
{
while(1)
{
ent_path_get_target(my,player.x,temp);
turn and move to temp!
wait(1);
}
}
The code tries to keep the calculations to a minimum (it remembers old positions, start- and target-nodes and checks if a new path is necessary), but if you know the start- and target-node, you can use other faster functions. I plan to release this when jcl fixes a path_next bug.
not to beat a dead and probably decomposing horse, but i still dont get why you only use med. you are obviously very talented so why not use something that will push your talents more effectively?
not to beat a dead and probably decomposing horse, but i still dont get why you only use med. you are obviously very talented so why not use something that will push your talents more effectively?
You could heavily increase the quality level of your stuff by using a professional modelling tool like Blender. But I like the detail grade of the mesh and especially the bow. Great job.
1st: Object Button list instead of 3D objects (scrolling with mouse wheel) 2nd: 3D (self-rotating) Preview when clicking a Button 3rd: Changing Daytime on the right 4th: Zoombuttons for the Preview + Resetbutton for the Preview 5th: Drag&Drop to shift preview-cam positions X and Z
Thanks to Slin who invested some hours in shadow optimizations it runs at 30-40 FPS at full detail grade with shadows, HDR, DOF and stuff in 720p resolution on my Geforce 8800 with 512MB video ram (~4 years old PC). 60-70FPS without shadows. It runs perfectly fine on newer PCs. So it`s still in the green area and playable.
looks really nice rondi... But what do you thinking about to put some dirt under your objects? (painted on the terrain) Not much just a small line because right now it looks to much "right now its placed on the ground"... i'm missing the "the building stands there for a while" feeling Hope you understand ...
The lake Garda also has some places where the mountains aren`t that high (and time-consuming ).
Nevertheless I`ll try to do my best and rework some parts of the level to fit better to the original. I´ll especially concentrate on the mountains in the background. I can`t change much on the vegetation and other details and make "that big and great level" because we don`t have a big budget. To make really great levels like yours I`d have to make my research photography work on the spot - and I`m afraid that wasn`t possible.
Quote:
Hope you understand ...
Yes, I understand. That`s a great idea. If there`ll be some time left in the end, I will get into that sort of fine-tuning.
@Rondidon: BravoBravissimo! Ich wundere mich immer wieder was man aus dem ollen Gamestudio so alles rausholen kann.
Why i'm working with MED? Just because I'm too lazy to use these high complex supertools. This is mainly due to the long "learning curve"(eng.???) without visible success.
Click to reveal..
But with a bit luck, is this ship my ticket to a "kollage"(eng.???) for Multimedia art. And they will force me to work with other 3d programs. "Use it! Or endless pain will befall you!"
@Rondidon : That rocks Would you put a download link ? I would like to test the level on my good PC and on my EePC also ! To imporve FPS you can find lof ot tricks : Shadows are the thing that takes the most power on games if not very optimised ! For trees : texture projection would be lot lot more fast ! And i don't know if you have distance managed for the grass visibility like in games ?
Have you sprites for distant trees ? big trick used by Far Cry and used today also !
Great, dirty look and very professional polishing. Maybe you should work on a bigger demo level instead of single models. That`s also great for learning how to model efficiently without influencing the framerate too much.
Hey guys, I thought I'll throw in a couple of pictures, too. I'm working on two things right now. at the Moment I'm focusing on the woodscene. Hope you like it.
City crossroads
Just trying to make some low poly but yet detailed buildings. You may call it "for learning purposes only"
The woods
This is what I'm focusing on right now. Started a new project with a friend and we thought we'll start to make a nice map before coding this time just to see how this forest may come to live one day. It's missing some trees though. That's what I'm up to next.
Added some details... This buidlings would be 4 floors height... (the second floor is just copied so this one would be change to get an more various look)
looks good, great job! I like the details and the great texturing!!
The ground of the balcony looks a bit overpowered - it looks like a whole fundament of a house. Did you model from a original picture? For my static understanding there is a floor of ~50cm concrete a bit too thick for the balcony, usually. The wall, on which the balcony is hanging on, is les thicker than the balcony itself, i think; I would be interested, if it's from a picture or made from imagination. Nevertheless, great model and great model-skill!
Danke, jetzt müsste es sich nur noch so spielen .. sind gerade noch sehr viel am Bugfixen und daran das ganze Ding etwas runder zu machen. Wenn man halt noch einen Monat hätte und manches von Anfang an etwas anders gemacht hätte.. Aber etwas Zeit ist noch.
nice da kriegt man direkt lust auch mal wieder was zu modellieren
Quote:
Danke, jetzt müsste es sich nur noch so spielen .. sind gerade noch sehr viel am Bugfixen und daran das ganze Ding etwas runder zu machen. Wenn man halt noch einen Monat hätte und manches von Anfang an etwas anders gemacht hätte.. Aber etwas Zeit ist noch.
improved follow-the-leader system with real formation support, better performance etc. far from perfect, some more path smoothing processes are needed...
Awesome building! But you mixed up chinese and korean characters. Just if you don't know. (The big sign has chinese, the two smaller ones have korean characters.)
1500 - 2000... A bit more Have added an Wireframe Shot in my last post... EDIT: Just a hint. The Building, without those "Metal Plate Roofs" / Pipes has 3600 Tris
+ Rotating of Objects + Selection Frames (see below) + Small Tooltip with Model Infos
Small question.. can someone tell me how to rotate the min_/max_xyz values, so my selection box rotates with the Model? thx to Rei and Puppeteer works now
really nice Rondi! But maybe adjust the contrast/brightness of your rocks in my opinion there looking to white... Doesnt fit to the rest of your level...
Im realy astonished to see what you still are capable to create with 3DGS... Seems like 3dgs is the perfect platform for Simulations. Damn if I hadnt forgotten my scripting skills I'd start new project
3DGS is reliable and stable. It works great for quick prototyping and smaller games. The only big problem is the Windows-only thing. And the editors / the workflow could be better. WED isn`t really suited for ambitious level design tasks without workarounds.
Anstatt euch mit fortschritten beim Schiffsbau zu langweilen, möchte ich euch heute einen teil seiner Crew vorstellen! Oder besser gesagt wie ich mir deren design vorstelle.
Warnung an alle mit ausgeprägter Herpetophobie . Es handelt sich um total abgedrehteDINOSAURIER PIRATEN
Click to reveal..
Click to reveal..
Na? Wer fühlt sich an die guten alten Zeiten erinnert, als noch Krokodil Piraten die Weltmeere unsicher machten
Sind sie etwa zu nah an ihren "Originalen"?
Was sagt ihr zum "Jolly Roger"? Immer das gleiche und einfallslos oder gut so wie er ist? Habt ihr Ideen und/oder vorschläge? Her damit!
Ich geb mir mühe beim zeichnen, also gebt ihr euch mühe beim kritisieren! Das ist ein Befehl vom Kaptain! OTTER
Translation by JustSid:
Quote:
Ahoi ye old gamers!
I don't want to bore you with the progress of my Ship, so instead I want to present you a part of its crew. Or lets better say the their design.
Warning to everyone with an Herpetophobie, it is about totally crazyDINOSAUR PIRATES.
Ähm, hab ich jetzt den Thread gecrasht, seit ihr sprachlos oder habt ihr gar keine meinung dazu? Seit ihr vielleicht alle Roboter? @John Connor: hat der Tag des jüngsten Gerichts schon begonnen?
Translation by JustSid:
Quote:
err, did I just crash the old thread, or are you all silent in awe? Or has nobody an opinion about it? @John Connor: has the day of the last judgement already begun?
@Otter: Wir sind nur alle sprachlos vor Begeisterung.
Aber mal eine Frage: Was ist eigentlich aus Deinem ANDEREN Projekt geworden?!? Das mit dem kleinen schwächlichen Sid und dem starken Ritter Sir Harry? Habe ich da irgendetwas verpasst?!? Oder gehören die Piraten zu dem selben Projekt dazu?
Translation by JustSid:
Quote:
@Otter: We are all just silent in awe!
Small question though: What happened with your OTHER projects? The one with the weak Sid and the strong Knight Sir Harry? Did I miss anything? Or are the pirates part of the same project?
apparently I achieved my minimum aim in the development of my RTS pathfinder, namely group formation works fine, target group direction can be set by mouse just like in Cossacks, and stupid units now know that they are stupid time to go back to editor improvements...
Otter, nicht traurig sein. Die Community ist von Dir so wundervolle Modelle gewohnt, dass die (gezeichneten!) Figuren da nicht ganz mitkommen können. Sie sind okay - aber allein dadurch wie Du modellierst, werden sie wahrscheinlich noch deutlich an Charakter gewinnen.
Dein Waffenideenwettbewerb hat allerdings alle neugierig auf ein sehr witzig klingendes Projekt gemacht, dem man verständlicherweise etwas nachtrauert, wo Du Dich anscheinend auf ein anderes Projekt konzentrierst.
Translation by JustSid:
Quote:
Don't be sad Otter The community is just used to your wonderful models, that the drawn figures can't keep up with them. They are okay - but only by your way of modeling them, are they going to gain a lot of depth
However, your weapon ideas contest made us curious about a project that sounded really funny, and its regrettable that you are apparently concentrating on another project.
Aber mal eine Frage: Was ist eigentlich aus Deinem ANDEREN Projekt geworden?!? Das mit dem kleinen schwächlichen Sid und dem starken Ritter Sir Harry? Habe ich da irgendetwas verpasst?!? Oder gehören die Piraten zu dem selben Projekt dazu?
Welches andere Projekt? Piraten, Ritter, Drachen, Roboter, Zombies, die Community.... Sid hat viele Feinde!
Translation by JustSid:
Quote:
Quote:
Small question though: What happened with your OTHER projects? The one with the weak Sid and the strong Knight Sir Harry? Did I miss anything? Or are the pirates part of the same project?
What other project? Pirates, Knights, Dragons, Robots, Zombies and the Community.... Sid has many enemies!
ohhh and by the way: good work zeusk! You make the eyelids separately, right? My eyes looks like this, after 2l. black coffee ............. and a arrow in my knee.
Cool! Richtig gut sieht das aus. Besonders die texturen gefallen mir (vorallem beim Soldat)! Zum glück haben wir nicht den gleichen stil, sonst würd' ich neidisch werden
Machst du nebenbei auch mods für skyrim?
Translation by JustSid:
Quote:
Cool! Looks really good. I really like the textures (especially the one from the soldier)! Luckily we don't need the same style [for our games], otherwise I would be really jealous.
Ja, die texturen für den soldaten sind 3000x3000, also schon schön detailiert. nein, aber ich hab mich von skyrim zu ein paar modellen inspirieren lassen, wie z.b. der helm.
Hier noch eine Veranschaulichung zum Detailgrad.
Translation by JustSid: [quote] Yes, the texture of the soldier has 3000x3000px, so its really nicely detailed. And nope, I just use Skyrim as an inspiration for some models, eg. the helmet.
I took the time to translate the last few german posts from Otter and the responses, sorry if its not always the same meaning as the original authors intended (they are free to edit or completely remove the translations of course). Hope it helps the english users a bit, and please, like Superku said, keep it in english from now on. kthxbye
Great model, great texture work. A hint: 3000x3000 pixels is not a good size for any 3D engine. You´ll need something based at the power of two (e.g. 2048x2048 pixels)
awesome stuff Schloder loving your texture work on the weapons epecially. And about the punk rock chicks eye lids. the one area i can never seem to get right is the top eyelid! XD it never looks as natural as it should and the shape is never quite right but i'll post new pics as soon as i tackle this obstacle
And about the punk rock chicks eye lids. the one area i can never seem to get right is the top eyelid! XD it never looks as natural as it should and the shape is never quite right but i'll post new pics as soon as i tackle this obstacle
The top eye lid only needs to be lower, it should overlap the iris, while the lower eyelid touchs the iris.
@Otter : Tht'a great ! And you could make it lot more cartoon with bigger canons also and bigger skeletton head on the ftont of the ship also But perhaps it's not the style you want to match ?
God.. this will be funny work to bring that into A8 ._.
It is rather simple: Make a duplicate of it, triangulate it and export it as FBX, then import into MED. Maybe you have to touch the skins in the skin manager, but in fact it is rather straight forward.
nice galleon or caravel or whatever ship. some ropes and rope-ladders could really improve the look if animated together with the sails. and those cannon-portholes should be over the middle stave if it is not a submarine
Hey Otter, really like your Canon Model But the Canon itself is to clean imo... Add some scratches and maybe some smut to get a more "this canon shot's so many Ships"
Yeah I got a few basic images, and assemble a random amount of them into a "platform" with random angles, which generate at random intervals while you fall. If you pass a platform, I convert the position of the player and the rotation of each image on that platform to the pixel position on each image, and check the alpha values to determine whether you collide or not.
Nice. I like it. But the flag is blowing in the wrong direction.
An interesting phenomenon. At 500 quants height, the wind in MED blows in another direction!
Originally Posted By: Xarthor
That lower row of canons looks a little bit too low. I mean you don't want to have any water in your canons aye?
That's right! I don't want that someone takes a bath on board. Ugh! A bath.
Originally Posted By: ratchet
@Otter : Tht'a great ! And you could make it lot more cartoon with bigger canons also and bigger skeletton head on the ftont of the ship also But perhaps it's not the style you want to match ?
I know! Unfortunately, that let the whole ship look smaller.
Quote:
OMG!! What a freaking-pumpkin head-dirty-plump-looking-whisky drunken-plumber. It's Malleo! Obey him, or die! How does it look like if you make the princess in this stil. =D
@Joozey: Nice! What is the story behind this? ^^
The cannon got an upgrade. Now with more dirt, more structure and less red! I hope this convinces the critics.
This is an editor for cutscenes, or little skits that present ingame news. The game itself checks for new such skits on the server, then, when the player chooses to watch one of those, downloads the definition files for any such skits, checks if all files are there (and downloads them if they aren't), and finally plays them.
I just created a new 3 min. long video of 3 moving groups with 40 units in each in an example level containing vegetation and buildings (functioning as different sized obstacles): http://youtu.be/U68sFfo0oNk Everybody is moving fine without further map processing required, so I'm happy now Graphics is nothing special because I use 3dgs extra on an old pc... and better usage of lod will improve performance.
@3run: no, they are avoiding obstacles, but only those which they should avoid. I set all stones passable i.e. handled exactly as terrain, they are impassable only if has a very high slope. but of course they could be set as obstacles similarly to buildings trees and bushes (as I did in my earlier tests). since there's no any c_tracing used for movement it is not desirable to use such a small models as passable because sometimes the unit legs are getting hidden (but the same happens in any Total War game when units are moving on high slope terrain). @wjbender: thanx, I'll start fighting system soon, and also have ideas for walkable buildings having 1 or 2 floors for seige battles.
Getting past pathfinding is a hard task but when you finaly do ,a whole lot of restriction falls away and its smooth sailing from there on forward into game developement which is what were really after ,so i congratualate you for sticking it out especialy since you had to do it all on your own , its such projects that keep us inspired
sivan, that's pretty cool! Looks like your secret dream project is an RTS? Or is it a secret, never-before-seen game that just happens to need that kind of AI and pathfinding?
~ ~ ~ ~ ~ ~ ~ ~ ~
Thanks for the feedback!
There's a lot of putting things together right now. Can't say that this is particulary motivating work, but comments like those make it easier!
It's just that I have a stupid way of working. As soon as the part isn't "interesting" anymore, because I have solved the main issue, I just... stop. I have to force myself to finish stuff. Instead, it piles up and all those little things have to be added in the end all at once. Take that editor I posted before, for instance. Now that this is working, I don't add convenience features anymore. For instance, I want to add an option to change the order of nodes. It's super-easy to do, with no challenge. But I just... don't do it. People, tell me I'm not the only one with that problem!
Now: Encrypting and decrypting music so that it's not trivial anymore to get the tracks (and all that without pro!). I'm not pretending that it's super-secure (it really isn't), but the goal has always been to only make it more difficult than recording the "Stereo Mix" ingame - which I believe this does, for most users. The thing is working already, and I can live with the slowdown the encryption causes, but my tool of adding new music to the library could probably be a bit nicer to use. Oh well, can't have everything!
It's just that I have a stupid way of working. As soon as the part isn't "interesting" anymore, because I have solved the main issue, I just... stop.
That's absolutely my way, sometimes, I wasn't even able to finish a work for university, means, write it down, once I understood the given subject! What a mess!
A good way to get motivated is when you have some people who depend on you finishing the thing. For instance I extended my GameEditor for a friend who uses it for a simple question-answer-game that is presented at sort of games convention. http://www.stadtbibliothek-bielefeld.de/5007.htm
@Error014: Looking awesome. Even before the next public version, it could be cool to have a VioletTV countdown on YouTube and IndieDB for the release, as well as some important announcements. Then once players get their hands on the game they'd be quite pleased to see all those announcements and countdown episodes available from within the game from the get-go
@Error014: yes, some kind of RTS is the aim, the system what I found out one year ago seems to be working it can help the AI decision making too, but that part will be another hard thing...
This is just a lighting and composition-test. I'm not satisfied because some things look a little off, but well I wanted to show you anyways. Maybe you got some ideas on improving this scene?
thank you schloder, i'm lovin it to get compared with commercial project But hopefully you don't mean the first Shenmue cuz their texture are so blurry
but to be serious, i love the style of shenmue, because it is like it is. so blurry textures belongs to this too i think. this has a special graphical charme of its time!
Got inspired to do some level design for once by chaos coder and RVL_exiles recent posts . Heres a rough concept of the level. 3D screenshots coming soon.
I like the way the objects pop up when creating. Oh! And the Editor itself with his loading bar and this many crazy numbers looks "saugut"
@exile:
@zeusk: Nice sketch! I'm really curious how it will look like in 3d. But I believe it will take some more time.
"I'll open a bottle of rum for you landlubbers." My Galion looks SOooo fantastic, that I wanted to write it in the "Mich freut es gerade total" Thread. But it hides from me. Anyway, pics coming soon.
Boss fight is awesome. I love the music and the way everything collapses. Well done! What I'm not too fond of is the boss. I didn't understand him at first... It's just the "eyes", right? Maybe that's just me, but I find he looks a little unfinished...
Posted By: Anonymous
Re: What are you working on? - 05/07/1207:48
indeed, very good boss fight. It reminds me of the run-away scenes from uncharted where everything collapse in the background
Though I'm not quite sure what exactly it is, it looks good to me. Will it also have interior? How did you came up with the style (red and wood kinda reminds me to the old japanese style)? Part of something bigger?
Good work! It would look "deutscher" if you split it into different sections with different colors (and don't forget to forget the ruler!). Currently it reminds me of an Asian townhouse. It fits well with your Chinese Restaurant.
The boss is the whole "shadow-wall-with-eyes"-kinda deal. Maybe I'll animate the eyes or something to make it seem a bit more "alive". The player has an ability that allows him to summon shadow hands from the walls - I suppose we could use that art asset as well for this guy. But in honesty, we're a bit pressed for time, so I can't make any promises.
ratchet, thanks for the suggestion. And while I like the style of 3D dot game heroes, it is very different to ours - which I also like. And it is a bit late for such a big style change! We had "hard coded" animations before, but we've replaced that animation system since. It looks good for most other characters, I feel (for instance, here's some of them dancing, but I do agree that the player's animation aren't too great yet. This perspective is a lot trickier to get right with this style, but I'm sure it's possible to do better.
HeelX, yeah, balancing is still a big use here. Falling down is relatively easy at first (I've played this battle lots of times by now, so I tend to make fewer mistakes as the "standard player"), so I hesitate to add more punishment to that. But right now, it is still too easy a battle, so some tuning will have to be done. If you fall down, not only do you lose time (which can be a problem in the latter stages of the fight, where the boss is pretty fast), but the laser will also be broken (the smoking one in the video), so you lose out and cannot attack this round. But that's hard to convey in a good way. I may resort to a simple sign telling the player about that in a room before this battle starts.
rayp, looking great! Only thing I dislike is the trail at http://www.youtube.com/watch?feature=player_detailpage&v=khY6RgyVYZI#t=12s 0:13, when the rocket speeds up. It consists of multiple individual particles, which doesn't look good. Maybe try either the streak or beam flag. Or, if you have fps to spare, use more particles and place them at random places between the entitie's current pos and its pos from the last frame.
WHAT THE HACK? I can see strange things in the corner of my eye. o.O Working on this dammit boat hack decoration today since 16 hours! And this is already the 3rd try in the last 4 days. Sometimes I think I'm going crazy!!!
Sorry that I don't show you my ship today. I never thought that this small thing took so long to build.
WHAT THE HACK? I can see strange things in the corner of my eye. o.O
This can either be http://de.wikipedia.org/wiki/Mouches_volantes or http://de.wikipedia.org/wiki/Blinder_Fleck_(Auge) where the latter case is only a sign that you are probably overworked. (btw. the wood looks a little flat compared to the nice decoration).
looks really nice Otter ! Can't wait to the the whole Ship ! 16 hours... Well, the details are so much time consuming... sometimes incredible how much time such a small object used to get finished
There are people taking care about their boast and cars , always shining like some ferrari cars No scratch , no rust , i'm sometimes borred about all that next gen games with destroyed environments , destroyed stuff etc ...
Hmm vllt hätte ich mal ein bisschen früher posten sollen... Meine Arbeiten am "Camping Manager 2012" sind jetzt fast abgeschlossen. Endlich... nach so vielen Monaten Arbeit! Hier zwei Videos zum Spiel:
hmm maybe I should had write a little bit earlier.... My works on the game "camping Manager 2012" are going to be done. Finally .... after so many months of work. Here two videos of the game:
Okay, I don't know whats so hard to understand about "This is an english thread", but obviously it must be something about the word "english".
I would really like to see this thread staying in english and as I don't want to translate all your comments (again!), lets make this really simple: You translate your posts by yourself, and in the future you just write them right away in english, OR, I will move them to the archive and write you a nice PM about derailing threads...
If you can't speak english, you can either create a new thread in german or you just don't post at all.
Thanks Benni! Also, nice work you did there, I really like the night scene, although you might want to tweak the rain a bit? It looks a bit weird when the drops face the camera all the time.
This is "slightly" inspired by the tron movies... And it's all about bits and bytes:
Seriously like this, I always thought it'd be awesome if you didn't limit yourself to just "space", hope you call that level training room or something though because grids in space does seem a bit... um, unlikely? Tough based on what I've seen so far from you, you have a knack for making "space" look "interesting", so, that's good;
I'm just hoping I might see that atmospheric level some day
Good luck man, I know you've been working hard, may Conter.Strike.In.Space become an awesome title!
Thanks for the comments, guys. @darkinferno, we're planning on having a level-design contest. Top 2 (or so) levels will make it into the game. If you have the time, why not enter with your idea?
@3run: the framerate is never lower than 512fps on my good enough machine, but there are only 15k polygons on the screen. It is an extension of the shadow mapping example from 3dgs samples folder. It is basically same code but with one more depth camera that modifies the mesh though vertex shader to fill the shadow volume.
last week I went crazy while optimizing my pathfinder (group cohesion, bottleneck behaviour etc.). performance vs. quality is an eternal fight... but now everything is acceptable, you could test it soon.
this helped me to get motivated to turn back to my editor, and corrected the terrain texture painting features, I stopped developing in Jan. some new features are also added, works smoothly (except "physically" removing a skin). I'm using the terraintex ffe, but rgb blend-map export is also planned, especially when I'm upgrading to commercial soon...
I have to add single skin texture painting too, and moving skins up/down (deletion and insertion works).
Sorry, but seeing this Unity game I only think: It is SO stupid to make an FPS in these days... You cannot compete with BF3 or CoD and you do not have a big player fanbase behind your game
Still working on the hidden ancient shrine level concept and transforming it into 3D heres an example of the steps i take when converting a high poly model to a game ready.
@Rei_Ayanami : It was to give ideas to 3Run And 3Run game seems ro be perhaps somewhat different ? If he wants to make some FPS why not, he didn't says he will compete with AAA titles !
@Zeusk : If your scene is little, why not using decimate , bu the best option is really some fast retopolgy !
@ratchet: thank you for some ideas, but my aim isn't to make something Battlefield 3 or COD type games. I only want to make simple, horror project, with some simple old school graphics. So, I don't aim to bring such stuff as destructible environment and as I'm not a level designer that level was used only for testing purpose. I'm looking for a designer to take care of that part of the project.
@Rei_Ayanami: I agree with you, but as I said in this post above, I don't aim at compiting with such game monsters as Battlefield 3 and COD series, that will be ridicules
The wall seems ok to use it and make some sort of rooms or closed fields/corridors :
How to apply a normal map ?? I failed ? I know i hhave to include "Fx" file than create some action defining : My.mat ="" ... I'll retry ... Well ... i'm very borred to not be able to choose that simple shader and apply it directly by mouse
------------------
We are in 2012 ... old A6 workflow should be very far away and not here after 10 years ?? Why do we keep A6 Workflow ? This is the very annoying and outdated workflow that makes me uses Unity for this project !
------------ Applied the shader and textures really quickly,terrain deformation and paint very smooth and big frame rate : instant result
You can add lights, shaders continue working and are lightened by them : Just what a 3D artists wants, nothing to do only add by click and adjust the lights settings !
Why A8 is so complicated in these areas and makes me loose so much time ? Well i stop, this is not a debate , and i won't start again saying what i would like in A8 ! I like A8 for casual little things, and i know it's programmer oriented.
still trying to import my Titelmenü Room into MED.. still getting ***** FBx Importer Error ._. i start thinking of recreating the whole fu****g room -.- (3 weeks of work for trash.. shit -.-)
Cause the Uv Mapping is getting lost with that way... fbx is the only format that worked from C4D to MED for me... I don't know why the MED gives an FBX Importer Error ._. A8 needs a direct FBX support, instead of running everything through MED and mdl..
Wow! Impressive, Espér. Only the columns on the texture look very flat. And it seems, that the chest is made of plastic. What exactly is this anyway? Looks interesting!
Ratchet: Your last stone has a photo texture right? He looks so different than anything before.
small question: How do I make an alpha texture display correctly? Small example:
Click to reveal..
without alpha
with alpha
I need a transparent texture for the ropes of my ship.
@Espér: Cool screenshots ! The chest should have less specular in the shader , caus it' have wood parts, just normal map would be better perhaps ?
@Otter : It's a shader from 3D Coat using two bitmaps (that's must be photo sources i think) ! You think photosources looks better than cartoon so ? Should i go this way instaed of hand painted ?
Ratchet: i think it's easier to make nice looking textures with photos than ones with good hand painting. And this even though I never use photos for my textures.
if you tell me how to make that i've a problem with the ground fog.. I use partivles with a random atart point and vel values... it works so far, but looks like circle clouds that appear and fade out xD need to work on my partivle skills...
eh... it depends. What do you want to do? Since I know that Esper uses Cinema 4D I thought it would be fruitful if he creates a lightmap with radiosity in Cinema 4D and adds the dynamic lightning on top (or the like).
Lightmaps are pre-calculated textures that store light and shadow info (not to be confused with shadow maps!). It is a bit hard to combine that with dynamic lights though and there are many techniques - I would generally advice you to buy some of the ShaderX books or Shader for Game Programmers and Artists and/or Real-Time Rendering (3rd edition), since they cover most of the used techniques out there. Realtime Radiosity is not possible unless you have a raytracing engine or the like, but there are ways to fake that, e.g. light probes or dummy lights. Then, you have to decide if you do besides or instead of classic forward rendering sort of a deferred shading / lightning, which opens up new possibilities as well.
Hm... technically speaking, there are some mistakes in my post, but it is also quiet hard to find an appropriate (short) answer to your question
working on a new Ground Fog idea for my Titleroom.. A Model wich has a fog texture and simple shufts my.u in his action. perhaps a second entit below that shifts the v parameter.. i think with LIGHT and TRANSLUCENT, the effect is better than particles
When you finish dungeons, you can give hints to other people in your team. You only ever hear one hint from those group of heroes who made it, but those who die? Their spirits haunt those dungeons, and instead of ruining everyone's day, they tell the player their hints. It's pretty neat! The location of the ghosts itself is, of course, also a big help, as it helps with identifying traps.
Poor ghost dudes.
[EDIT: I am aware its a bit hard to read in the shot. The whole thing is animated ingame, so just by that it's easier to read. But I definitely want to make sure that text never disrupts gameplay, so it appears with no trigger, and only transparent (and thus cannot hide anything important)]
because the parallax grass is computed with the camera render target and it is totally screen pixel dependant, so if you change the camera tilt the grass waxes and wanes.
still busy with my navigational mesh(recast) plugin im thinking of holding off any updates and demos i have been usualy uploading ,until i have a more complete version...
the tiledmesh is implimented ,needs further work for the dynamic stuff..
navmesh gets build on a seperate thread ,going to see how and where i could add more thread support .
i also am going to try and create a caller for script functions with threads(but thats a big if)
Besser kann ichs net sagen wie Alibaba. Was passiert wenn Du andere namhafte Software benutzt ? Reisst Du dann ein Loch in die 3D Matrix das alles verschlingt ?
ganz tolle Arbeit und voller Respekt. Bin echt gespannt, was du alles noch aus dem Hut zauberst. Viel Atmopshäre und Persönlichkeit. Was man verbessern könnte, wäre eventuell noch das Wasser. Sieht nicht wirklich nach Ozean aus. Aber der Rest ist super.
Was man verbessern könnte, wäre eventuell noch das Wasser. Sieht nicht wirklich nach Ozean aus.
*pssst* nicht so laut! Das ist ein älterer Shader vom lieben Herrn Superku. Sonst steckt er dich in eine seiner Milchkannen und setzt dich am ende von stage 12 aus!
Übrigens Rondi ... du sprichst die "verbotene sprache"! Bedenke welch unaussprechlich schreckliche Dinge dir jetzt widerfahren werden! ...... ...... ... mist
Thanks, thanks, thanks guys.
Harry: Galeotter? .... hmm nice suggestion. But "last day" fits better to the Slave history of the pirates.
@Wjbender : sound cool, sometimes it is better not to publish minor updates, because it can decrease the interest of testing development versions... it will be a big new feature of 3DGS.
Thats true ,the fact of the matter is im just wasting time with small updates ,i was trying to eliminate bugs by doing small demos of each feature i add by exposing it to others,since chances are better a bug wil be showed among more than only me trying it out.. But i can do more work by not wasting time on demos and updates . And just to be clear here to everyone else of how generous i am with developing or atleast trying to develope this ,i myself cant even use the plugin i dont own a gamestudio version i own a trial version and basicly the free version .im dont really care about how much over-all intrest there wil be in it when im done what matters to me is the experience ,learning and the fun involved in doing it ,nothing compares to taking up some challenge and nailing it ,thats whats so cool about coding ,the challenge .finding a way to beat it ,kinda mcguiver stuff lol
I was leaning towards the american mcguiver macgyver i dont know the spelling but its the dude that solves all situations he gets into like give the guy a army heading his way to kill him and just throw him a bag of bolts he will then start talking (in thinking mode) and then build ten tanks from that bolts and capture farts in a bag and airdrop it onto the enemy with atomicly explosive results couse his just that kinda resourcefull dude
Currently only 4 sides (front, left, back and right), but I'm planning to add all of them (so that is going to be 8 sides like, front, middle-front-left, left, middle-left-back, back, middle-back-right, right and middle-right-front).
This is certainly a hell of work, to sort all the images for the animation correctly. But cool looking sprites! In the second picture; the Flathead has a little TOO flat head.
Now i'am working on the (I love it to repeat the thread title over and over again) first pirat-guy.
I'm also try to search a name for this reptile race.
I'm also try to search a name for this reptile race.
Nice work, as always! Here are some ideas: - "Ghari" -- from "Gharial", a thin, long crocodilian native to India. - "Magar" or "Makara" -- meaning "water monster" and "crocodile", respectively -- the origin of the name for the "Mugger crocodile" (also native to India).
Sorry, but seeing this Unity game I only think: It is SO stupid to make an FPS in these days... You cannot compete with BF3 or CoD and you do not have a big player fanbase behind your game
Why does it have to compete with CoD/BF3? There are alot of FPS's out there with their own stable fanbase. I think its a great genre, its just that most people making an fps make it SO damn cliche repetitive and boring, almost everytime I see a new Fps being made by an indie, the only thing "unique" about it is the choice of weapons and theme.....same boring crap, be it with unity or gamestudio.
A mix of both. First, parts like arms were added by script then replaced by a gore-model. But this was slow. Then i tryed vmask with the head + replacement model. But this time i tryed "damage" models for the zombies. Textures and animations shared, bone animated. So right now u can say its preanimated. Bad: Not much wound combinations possible, much work Good: Fast, good looking
Making more progress on IntenseX custom animation and IntenseDialogues is now complete. Lot´s of thanks to felix (masterq32) for his execute alternative Enjoy: http://www.youtube.com/watch?v=lpzzYgGLUjM
Rayp, heres an idea. Add a texture thats just a completely invisible black alpha image. Make the head and limbs their own group, then just change their skin to the invisible alpha image during gameplay and create a limb gib (should have a bone at the joint) and set its initial position and angle to the bone its being created on.
EDIT: Oh, and the assets in my game arent mine. Theyre all just placeholder (except the hud and sound which I made). I am not an artist, and dont claim to be. I am a programmer. Now if an artist wants to work together on this project thats fine. But until I can find an artist, I'll keep working on the programming aspect itself to make sure the game actually works as intended.
Thank you for that tip. Thought about it too cause its nearly the same as using vmask. But when using alpha-transparency i had troubles with shade-c (if i remember right). But i think for now i use the gore-models. And i have the feeling its faster without the extra models and there while loops.
currently i am work on my textures, for a better quality i create this in blender. My target, i design a modular scifi textures for an fast and flexible level design.
Trying to find a good looking design for a futuristic space shuttle used for my PAX-ET-BELLUM project. This shuttle will be used at the beginning of the game to bring scientists (time travelers) from the Earth to an orbital space station.
@HarryPotter: just check for inspiration the new space shuttle of Pentagon called X37B returned to Earth a few days ago (it is unmanned and really ugly), or the earlier plans by NASA called X33 (it is really good for a base sketch). I would not place windows onto the front edge of your space shuttle, where heat could blow it when entering atmosphere. or make it really different from existing plans.
I would not place windows onto the front edge of your space shuttle, where heat could blow it when entering atmosphere. or make it really different from existing plans.
With these big wings I guess it would enter the atmosphere at a much slower rate, as it has a greater drag. How do you want your spacecraft work? It could even use a special elevator system like in Freelancer or make it a space SHIP, like in Cownboy Bebop where the space ships can also land in water.
We don"t see a lot of things in the video , and the camera is too far and don't ove to show the normal and specular effects, a little useless ? What make me sad , is that other engines propose complete lightening and shaders integration, and some propose Deferred shading alreay Not only Unity 3D , some low cost engines like LeadWerks engine propose also that deferred shading.
So 3DGS makes me a bit sad in that areas. That's like lightmaps import on models that is a hell on workflow compared to some other engines.
i stuck in some problems with path functions... Created paths in WED.. but can´t use path_spline to walk on them 8cause they are not closed).. Now i give it up, to try to move the camera from node to node, for today ._.
No ratchet, you are right and these are things that should be sad which have thus far been ignored (except from some exceptional members of the community ). At the risk of going off topic, light-c really is a powerful engine, so I feel its about time we got some tools to really push the engine artistically. This has been the elephant in the room for quite some time. But in the interest of the topic...
I think that the spaceship looks awesome! Only, I think its a bit too rounded. Maybe if you gave it some angles to give it a somewhat "stealthy" look to give it a bit of personality. Right now it looks kind of toy-ish (But its still great work. Better than what I can do at least )
@ratchet: The only thing "sad" about txesmi's video and demonstration is your reaction. I mean... really, you call that "useless"? You talk about deferred shading (again and again) like it's the solution without probably really knowing what it is about, deferred shading is a lot less user-friendly, all shaders have to be adapted to work with the specific implementation and it does not improve the visual quality automatically. I don't say that GStudio's lighting system is highly advanced or very user-friendly, but such effects like the one from txesmi are way more useful for an actual game than all-in-one solutions. Not only you but a lot of other users always demand a better lighting system for A8, but they never say in concrete terms what can/ should be improved. If you have a specific thing in mind I suggest you make a researched feature request, this will help much more than continue complaining.
Love the design, wouldn't change it because it makes your game not look cliche, has its own voice, however I don't like the wings on the fourth
@Superku: Well, I wouldn't call it useless but it does seem like something that pretty much won't be used for anything. I don't care for deferred shading though;
I demanded a better lighting system too and am sure its currently so bad that they shouldn't even need a list:
We make levels with blocks right? okay, why cant block have shaders without screwing up shadows then? Why is the lighting a model with the default lighting still soo ugly?
The problem is, yes, with a LITTLE work, we can avoid that but guess what, we should care about the guy thats going to download the trial... import a model, place a few textured blocks, add a light then say: "bleh, NEXT engine" Because of how ugly it'll look, he's not going to "spend a little time" to get it, not when it works immidiately in another engine.
And I think this community has given up on feature requests ages ago, unless its some little minor function that can be added in hours;
@Darkinferno: What is useless, omnidirectional lights? WHY?
@txeesmi: ah looks so nice and smooth. Perfect for proper dynamic indoor lightning(which is not yet possible until you write your own stuff) Are you going to contribute this? Currently imagining remaking my Bombermanclone with omnidirectional lights in Explosions :3
@Superku : Yes, you are right , and im' not specialist shaders at all. I just hope 3DGS to improve lot more in workflow ... it's just a wish
@txesmi : How it is complicatde to implement in 3DGS ? Anyway that's some great work ! But i think only some real scene with very visible specular objecst and normal mapped ones will really show how great it can look. But it's not your matter if it's visible or not , if it can render at ggod speed handling some directionnal light and shader easily it will bring 3DGS to some new level for some people
Yes, it is gamestudio, i don't know the exact data size, but it's about 256x256x64 currently. The size of all voxels is not fixed, you can expand in all directions (-x,-y, -z gives some strange behaviours, but should work though) also you should be able to render nearly "unlimited" voxels, i will do some tests later (have to fixup some system parts)
I'm also interested in how you did this. I did some tests with draw_point3d(pos,color,alpha,size) some time ago, but the squares rotated with the camera and wasnt really fast enough anyway... I also tried with oriented sprites like you did in "CubeX" but I could never get enough speed, at 128x128x16 no matter what I tried it always started to get extremely laggy
Very nice! Perhaps a way to load functions from another .fx file and express them as a node would be good? This way people can make and share their own nodes so that those who are petrified of shader code can just drag in more complex things like the worldViewProj transform (even though it's just a multiply, a lot of people are scared of matrices) and building the tangent-space matrix.
Thank you. U can blow out the fire by activating a pump system first then the water-spenklers. Maybe i make a vid these days. Fire-Blow-Out effect took me some days. Its one of the map quests to get behind the fires.
Edit: By the way your doom levels look really good, ive seen some shots already. KEEP THIS UP PLEASE.
@Rei_Ayanami: (Are you having a MATERIAL with RELOAD flag, or do you recreate the shader each time that the user changes something?) At the moment, the shader previews on the nodes are recreated every time you change something in the previous or the node itself. Atm, The final shader is recreated every time a button is pressed (for example the c-key). Then a new effect is created and assigned by effect_load. So no reload flag...
JibbSmart: ty, yes this would make sense(it is also implemented in the source shader editor). It will not be that difficult to implement such a node. However I need to create a lot of other nodes too (dirty job -.-), like texture samplers, swizzle and append nodes, etc. etc.
Dude your helmet totally doesn't fit your shirt. You look like a tool.
[Shows graphic generation! The player's sprite is updated depending on what equipment he uses. Works by overlaying three bitmaps in different ways to achieve the most flexible results]
The game's in no fit state to be played right now (debug stuff all over the place!), but we're getting there - I guess the finished game can be seen in the distance, and not too far away. We'll let you guys know as soon as this thing is playable!
I've been working on a lot of new stuff centered around modeling characters for my new 3d assets store thats going to be up and online soon. The first model pack will be a standard medevil fantasy pack. Heres some preview images of one of the characters:
Goblin Foot Soldier -
The pack will include a Demon, Knight, Mage, lesser Dragon, and this Goblin
I have been working on creating an Infinite Terrain Generation system. The terrain is generated randomly and the maximum height can easily be set as well as vegetation and ground textures. The ground textures are blended to make the terrain flow. It is easy to restrict vegetation to specific ground textures and easy to restrict specific ground textures to specific terrain heights.
It still needs some work and I have some plans to make the program run better. as of now terrain is randomly generated and places that the player visits can not be visited again. I believe I have a method in mind that would save the terrain that the player has been to so that if they return it will look the same.
nice goblin, although he looks like his best days have already passed and got retired =D. But it's good to have not always the same grimm and evil looking ones.
@msmith: cool idea, you could use a pseudo-random algorithm and a seed to generate the terrain.
Nice work msmith! What's your idea on saving the terrain? I believe if you do random_seed(my.x+my.x) whenever you create a new entity at (x,y), then you will get unique numbers for each tile, that are the same every time you run past that tile. You might have to take the absolute value of the thing though, in which case you might get a point-symmetry around the origin. To avoid this, you can try: random_seed(absv(my.x+my.y + sign(my.y)*my.x+ sign(my.x)*my.y)) or something like that. random calls after that will always produce the same numbers, but only for the same tiles.
This means you don't have to save anything. Or, if you do want to save something, at least you don't have to save everything
Your infinite terrain generation is not bad, but it looks very flat and the same everywhere ? You should make big variations on terrain height perhaps, and why not some texture variation using multitexture, put some whater plane, this way on very low height there would be water. Trees variation also, rocks placement ... all that to find a non "looking the same everywhere" terrain
A nice little beauty shot of the Cr-17 Arion, assault rifle from my game "Fallen Steel" (Made entirely in med, keep in mind I am not a modeller/artist)
@Germanunkol - Great Idea I had not thought of that. My Idea was to basically save the tile specs in a String in a double nested arraylist of strings. each string would represent a section of the map and would be initialized when the player traveled to that section. The tiles would first try to generate based on the string. If the tile had not been created before then a random tile would be selected and saved in the string then loaded into the level. This would allow the player to revisit areas previously discovered.
The infinite terrain generation slowly came together as I was experimenting with some new things. first I had created the tiles to generate as the player moved in any given direction and remove the tile behind the player. This worked well so I applied a grass texture and decided it was to plain so I added a dirt texture and randomized the textures creation. The problem then was that there was a clear line between the grass and dirt tiles.
To fix this I took a detour into BMAP manipulation. I wrote some functions that would blend the edges of a texture into the other by a specified amount. Once this worked well and I could blend textures I implemented it into my terrain generation. A quick problem I ran into was that changing the skins to these modified bmap textures would change all of the skins of that model type. after messing around with ent_copyskin and so on I decided to ditch the models and use ent_createterrain. After switching over to that I had sucessfully implimented my blend texture functions.
It worked but now the terrain seemed to flat so I changed the height of the tiles and wrote some new functions that would sew the tiles together. Sure this worked but it was blocky and the terrain was rough so I wrote another function that would smooth the terrain tiles out. This worked great so I added some vegetation to give it a better feel.
At this point there was a major problem. every tile created had its texture blended to the one next to it and its vertex's altered. each new tile had to run hundreds of thousands of calculations and each tile section contained ten tiles. I had been making the program run through millions of calculations every time it created a new tile section. This would make the program glitch a bit when this happened. To fix this I created a library of bmaps that would store the blended textures that had been created based on the tiles and its neighbors. When a new tile is created it checks the library for the correct bmap. If it exists then it simply copies it. If the bmap does not exist then it is created and saved in the library. This way the calculations would only need to be done once.
That is how it is currently done.
A problem that I am currently faceing is the nexus is filling up. This takes a while but is still not desired. ent_remove removes the tile from the map but it is still stored in the nexus. I have been thinking about this problem a bit and I believe I have come up with a solution but I have not tried it yet. My idea is to recycle the tiles instead of removeing them. when a tile becomes out of range that tile's texture, location, and vertexes will be reset to its new location and from there the tile manipulation functions can do their job.
This is all experimental right now and suggestions are appreciated. Ill keep trying things and improving things until I have a solid foundation for a new game.
@ratchet - Thoes are great Ideas. I will play around with them and see what works. My idea is to have the general look of the terrain change as the player travels. sort of like entering different biomes. perhaps desert or snow or islands and mountains all with their specific settings for the terrain generation. I will experiment with that soon hopefully.
I really am a fan of the Gameboy-Color style, I must say (which they were inspired from as well, they say). I feel it's a bit underappreciated - everyone always goes for NES or whatever. But I have more and better memories from Gameboy games than from NES I really like the mix of modern-ish lighting and the less detailed tilesets. In fact, that's something we have in Dungeon Deities as well! However, I'll be the first to admit that in our case, it's far less good-looking (there's reasons for that, of course). I'm actually doing tests with coloring walls as well (before, only the floor would show effects from lighting, and entities moving there)
@Error014 : Yes the game is made under Blender, i testd it, it's fun to play and runs well, the developper could have choosen a real 3D engine ? Your style is amazingly great : don't change anything Have you another gameplay video ? When will we have some little demo ?
Thanks guys for the comments! The next chance to play this will be a public beta session that we hope will start in summer still, so not too long into the future!
For textures, we often use simple Paint (16x16 pixel tiles are surprisingly easy to make there). I've also tried the free edition of Graphics Gale, which is pretty nice
Thanks guys for the comments! The next chance to play this will be a public beta session that we hope will start in summer still, so not too long into the future!
For textures, we often use simple Paint (16x16 pixel tiles are surprisingly easy to make there). I've also tried the free edition of Graphics Gale, which is pretty nice
... Uhm, I'm not posting too much on this, right? You guys would tell me if I post too much about this?
as always your game looks frickin amazing subscribed to your YT channel just now Spread the movie you posted around to some colleagues here at work, when the Demo comes out, I think we'll all have a go at it!
I just released the new version of my MapBuilder Editor. Have fun with it, of course for free
You can check also the pathfinder demo, whose development screenshots and videos I posted here earlier. It's inside the editor, you can control 120 units in 4 groups. Please note it is not a finished system at all, but works on any terrain, if no long walls or too large entities are placed. I would be interested in how fast or slow it runs on your pc...
I am still working on my CEGUI plugin. It is allmost ready for release. I need help with two points:
- I need to create the documentation. Where can I get this ugly gamestudio documentation stylesheets? Is there any good (free) tool out there which helps creating a documentation?
- I need to create a demo, to show off the CEGUI features. Do you have an idea what kind of demo/project would be good for a gui demonstration?
I hope you'll insert automatic panel adjustment for different screen resolutions. So if I place panel lets say at "pos_x = 10" and "pos_y = 400", whatever I change resolution runtime, panel should stay at the proper position, updating it's X and Y automatically.
CEGUI has two positioning methods. The unified positioning and absolute. with the unified system you can align the window to the left/right/top/bottom/center side und use an relative offset to define the distance to the border. you can also use exact pixel offsets, or a mix between relative and absolute positioning sizing is possible too. So yes, you can position your widgeds resolution independent - or, if you don't like this - not.
The premade skins are ugly, so I created a Win7 lookalike skin.
The pieces are finally falling together. Lots of time and work are finally paying off and I couldnt be more proud of my project than I am now
So the game I have been working on finally has a set name, its now called "Fallen Steel". The game is a multiplayer futuristic shooter (multiplayer ONLY, at least for now), which will have an "energetic" feel to it. In terms of gameplay, the best way I can describe it is like a mixture of Unreal Tournament and Halo. The plot of the game is, in 2310 the entire planet of Mars has been turned into a dumping ground and and prison where the worlds criminals go. However by 2560, the entire planet of mars is being overrun by garbage and the prisons are far too full with thousands behind their walls. So Orion, a weapons manufacturer, offers to buy the prisoners to be used for weapons testing which is approved by the GSF (Global Security Federation). At first, they prisoners were only being used for weapons testing, but eventually employees and their associates began initiating betting circles and underground competitions. Orion noticed this and saw a major profit to be made which lead to the creation of the controversial bloodsport known as "Natural Selection", which was televised on a weekly basis on Orions personal broadcasting network. All seemed well until the GSF banned all bloodsports for being inhumane and torturous in 2568. But that didnt stop Orion, after the ban Orion had begun working on a drone soldier project which would revolutionize the combat industry. These drone soldiers were "piloted" by real humans with what Orion called "Combat reality" suits. These suits allowed the pilot to feel as though they were the drone itself, responding to every muscle and movement command as though it were their own body. By 2570, Orion came back with the new bloodsport known as "Fallen Steel", which used real athletes piloting the drone soldiers in a globally broadcasted event. By 2584, Fallen Steel had become the most popular sport in the world with champions from every colony and country across the globe.
Once I flesh out more of the story and the primary backstory characters and lore, I will make a Showcase thread for this project. For the time being here are some videos which show off a couple features within the game (Keep in mind that a lot of assets are still just placeholder, as I am not an artist).
"The explosion of the rocket looks a bit small, the missing texture are not so cool." - Yea, I agree, the explosion looks a bit small. I am working on a "billowing texture" effect which will make the explosions and smoke a lot more dynamic in the looks department. As for the textuers, well, im not an artist. I have been looking for someone who can unwrap the models I have made and texture them but as of now no luck. So I just gotta work with what I have for the time being
"A death animation would be cool (even if the screen only turns red, while you sink into the lava." - Agreed, as of now I am working out ragdolls for the game. One of the animations I had in mind (for lava deaths) would be a slow sink with the players hand reaching up terminator style. I'm just trying to get ragdolls+networking to work properly. However I was going to wait until I had a proper base player model to work with (which I dont).
"Two "bolters"?" - Yes and no. The bolter thats there as the world pickup model is the actual bolter I designed. I just used another model to test animations and such (just to get everything ready for when I get an animated version of the actual bolter).
"After you pick up a weapon, they should be removed for a few seconds, including their "light" - Actually, they do. I just have all the weapons set to a 3 second respawn time for testing purposes, but the weapon absolutely goes away 100% of the time. Also, the "light" is coming from the weapon spawner, not the weapon itself, unless youre talking about the outer glow of the model in which case goes away with the model.
"And I wouldn't give my user so many possible things to customize in the UI - he might think you were unsure what to do there." - Well heres the thing, I actually have 4 different UI designs which will be in the final version of the game. But one of the biggest complaints I hear are about UI design. Sure the artist takes gameplay into consideration and tries to create the best UI possible for the game, but the fact of the matter is different people like different things. I believe in giving choices, not taking them away. If a player wants his hud to look blue, to have the health display a bit bigger of the ammo display a bit smaller, or to have his font in a different style, why shouldnt they be allowed to make that choice? Think of it like a car. I wanted the player to have the ability to tune their experience just for them. Of course there will be some things they can't do, but when it comes to asthetic design I say "fuck it" so to speak. I'll put it like this, look at all the different texture packs for minecraft This game is being designed for PC and one of the things that makes PC gaming unique is modding. All I am doing is building this feature directly into the game
Oh and for the record; Killing Floor, Natural Selection, Interstellar Marines, Hawken, Nexuiz, are just a couple successful indie FPS titles
Awesome. You should work on the 3D-graphics and 3D-modelling, but the menu and the overall "feel" is great from what I see in the video. Very useful and good looking 2D art design.
Still working on concept art & animation. This time: Concept animation for the opening and closing of a puzzle, regarding rings and rods (click image):
By the way, I used NHC VideoPad for this and it is a really cool, free, video editing tool. Recommended.
A Zombie/Survival Game. Since I'm a Level designer, I currently have no enemies. the first two are the warehouse before the power is turned on. the next three are with the power on, and the last is an outside view of the whole level.
Hm .. well .. you should work on improving your skills. Add multiple layers, textures, details, architecture. It looks like some Wolfenstein map without textures right now. Doom 1 had better graphics .. Nevertheless it`s a nice start. Thanks for sharing with us.
I can't add shaders (can't afford the paid versions), or anything "pretty" for the fact that my computer is not meant for gaming, and barely runs this. This level is still in block out mode, so nothing looks good. Once I get the level flowing the way it needs to be, I'll spend some time making little assets such as railings and stuff.
Also, When I turn the lights on, my FPS drops to 20 approx. (and there are only 12 lights)
Here's a sample from my current project. Since I don't own the paid version I'm using the free version, too. There are no shaders involved as well here.
But you're able to achieve something like this, and even better graphics without slowing down the engine. As you can see I've got over 500 fps. (top right corner). And I wouldn't call myself a level designer to be honest. Just did some models, some textures and copied some effects I saw in a bunch of games I played. I work all by myself. Programming, models, levels, textures, sound, story and character design.
This runs on a 2 year old notebook by the way.
And don't get me wrong. I don't think I'm the super hero with the amazing work. There are a lot people in this forum with awesome skills. I consider myself not even as intermediate. I just wanted to give you an idea of whats possible with the free version
@Rondidon I've worked on many levels in the past and they all came to a screeching halt because I got to detailed to quick. BTW, IMHO this level looks nothing like wolfenstein or DOOM. Trust me, me and 3run just worked on a doom level, and this is MUCH better than doom and wolfenstein.
@Roxas I'm woking on a 4 year old notebook. a Dell Inspiron 1545. BTW, how much of that screen shot is WED level blocks, and how much is assets (models)?
One of the doom engine feats was walls with other angles than 90°. yes in ws3d are only 90° walls...oh and "multi-floor" came with doom too. ^^ edit: sry multifloor came with RiseoftheTriad. A mod ws3d engine.
I know xbox can do great doom maps by the way, what i saw from the screens. Maybe you posted your warehouse a bit to early xbox.
@Roxas Yeah looks really nice, fits as Rondidon said already.
Thanks @rayp. I'm beginning to sense that I posted it a little early. I've been watching this thread for a while so I felt obligated to post, what I'm working on. Oh well, I guess we live and we learn.
@xbox Well to be honest ^^ Nothing is WED Blocks. Just a Skycube and a flat block plane underneath so that the lighting affects the models properly. I often try to reduce the polycount. I delete faces you're not able to see anyways and so on. Textures are all handmade in photoshop.
@rayp and rondidon Thank you ^^, but its not finished yet. I think I'll add some more little details. the next 2 maps are also designed.. on paper that is but I like to draw my maps before i get to work.
Thanks for the hints I´ll try to do that, my goal is, that it also runs on old pc´s. EDIT: the other half of the word is missing. "purification" is written there
Just a first test. Not final, especially not the colors. I`m aiming for more variety and more realism to get more interesting levels. At the same time the polycount is 20% lower compared to previous levels. On the technical side I`m using Blender 2.63a now and export my levels with Blender FBX export instead of Ventilator`s old Gamestudio exporter to get them into WED. While I had problems with UVs in older versions of Blender, the newest version seems to work flawlessly.
@Rondidon looking good, however I cna not imagine what you can do to improve realism in the level as it seems from the screenshots that you've managed to achieve quite good atmosphere into the level.
@Rondidon I really like the looks, the graphical quality of the assets is great! But I think the lightning still doesn't look right. Its much to flat, the tree shadows in the distance are simply not visible on the gras. Maybe you can find a solution to that, but i guess it'll mean using shaders which could be slow with the high amout of polygons you already have.
Don't get me wrong, it's a great level, but I think there is still room for improvement!
@Alibaba: I like your level and your work a lot. Is there a timeframe when we can play Mother`s Den 2? Is there a story? What are you aiming for in terms of game design?
@Pad, Keks and Walori: Thanks a lot for your comments. I`m becalmed to read that you see things similar to me. I`m indeed planning to write a new lighting and environment system on base of a new Shade-C version that is not released yet. It will feature dynamic daytimes, dynamic weather changes, a better sky system (now it`s based on the buggy Gamestudio sky stuff) and seasons. Hopefully this will bring about better graphics and more performance.
On the technical side I`m using Blender 2.63a now and export my levels with Blender FBX export instead of Ventilator`s old Gamestudio exporter to get them into WED. While @Rondidon : it is from ages that i use Blender.49 to FBX without problems, even for animated characters, yes Blender rocks
Anyway, great level
Will you put some birds , characters or wind to make the level more alive ?
Maybe I did something wrong. I don`t know. Hm. But I`m very happy that it works now.
Yes we`re planning to create a more living atmosphere. Some birds will fly in the sky and maybe there will be animated ducks or rabbits as ground animals. Wind is already integrated and the final game will feature dynamic weather and day/night change plus seasons. There`s a lot of work to be done.
Game making is HEAVY time consumming ... I need some solution to make lot of different characters : NPC, ennemies on a same base shape (RPG Maker like style, little big planet etc ...) AS FAST AS POSSIBLE , i mean just changing texture, adding custom items and clothes , so i decided to make a generic base character for any game style using characters i would want to make.
I begun some basic texture color, my goal is to make fast customizable characters somewhat like MII characters of Nintendo or Avatars of Microsoft , even if i won't reach same quality
3D coat for 3D painting is really the must (paint lines and width, paint splines , paint with brushes like "make clear", "make darker" etc ... )
And Blender just rocks !
Does anyone know if there is some rim shader for 3DGS ? like this one ? mario galaxy perhaps 3DGS toon shader will do the job well on on some diffuse detailled texture ?
I found that with the actual physX plugin it is dificult/impossible to attach multiple joints, so i took it differently and return an internal joint id instead of the joint pointer. It is then possible to remove/add/configure multiple joints (a function can be implemented to return the joint pointer if really needed). Actually, I implemented all joints found on the physX SDK (Fixed (PH_FIXED//added), Spherical(PH_BALL), Revolute(PH_HINGE), Prismatic(PH_SLIDE, PH_PRISMA), Distance(PH_ROPE) and D6(PH_6DJOINT)).
I tried to name functions like those of the actual plugin (added 3 after the X), so, except vehicles (I'm still thinking about a simple way to implement it for GS) it is more or less compatible with the actual plugin.
PS : I'll implement joint configurations before working on vehicles.
It took a lot of time to create all the shaders but now I've got a post processing shader pipeline with real HDRR. It includes adaption, bloom, camera blur, chromatic RGB shifting, static gaussian blur, color grading and other stuff
The key concept in the game are orbs - magical artifacts. When you finish dungeons, you (almost every time) find these. The more dungeons you have created (and that others can traverse), the more orbs you can have. And to make sure that your abilities get changed around often, your orbs can get stolen, and they also break after a while.
Orbs come in two flavors, the orange "active Orbs" and the purple "passive Orbs", one of which you can have equipped at any one time. The former are special abilities, such as a firewall (pictured above), but really there's all kinds of stuff there. The latter are permanently active and slightly skew the rules in your favor. The one in the screenshot above lets your active orb last longer (takes longer for them to break). And to show you when orbs are effective, they spawn orange/purple particle effects around you (which you can also see in the shot above).
For me, it's mostly adding lots of content like this to the game thats left before we can start into the beta. Can't wait for that!
@txesmi: Neat! How many of those hexagonal grids are there? What kind of variety is there? In the video, there didn't seem to be TOO many differences, which I understand is due to the early work-in-progress nature of it. But what kind of changes in the tiles can we expect in the full game?
(Also, are the enemies' names randomly generated? :D)
--
@ratchet:
This game was started and worked on for quite a while in A6. I've since ported it to A8 and lite-c (quite an undertaking, but it was worth it); but it means that some things are done differently to how I'd do them knowing what I do know and if I were to start that project from scratch in Lite-c. To answer your question, the dungeons are pieced together via simple "tile"-models (there's one for walls, and one for the floor tiles, as well as a few additional models), whose skins are changed according to what tile it is.
The dungeons are created by the user in their browser. Here's an old screenshot of that. Once that's done, you piece the rooms together to form Dungeons in the Dungeon Editor (click for shot). These Dungeons are then selected by the server and sent to the Client so you can play them! There's all kind of stuff to place - Tiles, sure, but also enemies, decorational items (such as trees or grass), Treasure chests, Buttons and laserbeams for puzzles, Rotating platforms... We're still adding those kind of things right now. If you have any ideas, freel free to share (and understand that you grant us the license to use them without paying you ever ;))
Right now, the only shader-like effect we're using is for reflective ice tiles (an effect that can, of course, be switched off). Everything else is done without any shaders. The lighting visible in the shots is, for instance, all done by hand using the bmap_ and pixel_-instructions for rather small "lightmaps" that are then filtered. The generation of that map isn't noticeable as it is, so I'm not too worried about speed (though I may still add some improvements in speed to it)
The game supports joysticks and gamepad, and the standard button layout is for a 360-pad. But it only supports them via the joy_-instructions, so no support for the big button in the middle (though I can't think of any game that supported that one, actually!)
Once more, thanks a lot for the feedback and encouragment! Means a lot to us - the final stretch of game creation can be pretty tough!
How many of those hexagonal grids are there? What kind of variety is there? In the video, there didn't seem to be TOO many differences, which I understand is due to the early work-in-progress nature of it. But what kind of changes in the tiles can we expect in the full game?
There is just one hexagonal grid. It is used to constuct the path. Each light is near the center of each cell. In the example the grid has 16x16 cells, but it can be up to 128x128 cells with tons of performance problems because the grass models. I tried merging them to less entities but with no luck until now.
The actual state is something like a glue for special events triggered along the path. The path has space in both sides (now occupied by trees) that can be filled with unique models that will give some diferences to the sceneries. I agree, right now it looks very flat. It has just seven different models indeed!
-------------------------
I added an atialiasing and dof blur effect to the render pipeline
But I'm still having problems with my distance blur [...] effect because the DepthMap doesn't work correctly for sky-entitys.
What exactly doesn't work? I am not sure what is written into the depthmap (is it 0, maybe?), but since camera.clip_near should be > 0, I would overwrite all pixel depths below camera.clip_near and beyond camera.clip_far with the value of camera.clip_far.
Then what is your problem? Don't you wish that the sky is blurred? If not, you can solve it rather technically by using something like clip_far - 1 as threshold and write 0 instead into the depthmap, or you create a buffer, in which you write 0 and 1, which parts of the scene should be combined with the blurred image (where the sky would be black then) - you take this mask as lerp factor between unblurred and blurred image in the composite pass.
Why to hell do you want to use Shade-C water when you are able to write your own shaders?!
Well... good question . I've got almost no experience with vertex-shaders so I have no idea how to add 3D waves, calculate the reflections but I'll definitely try when I've got more time.
Originally Posted By: HeelX
Then what is your problem? Don't you wish that the sky is blurred?
The stars in the night are a sky-entity and it doesn't look very good if they get blurred... but I'll find a way to fix that
Originally Posted By: txesmi
When a skybox is rendered in a depth view, it writes the camera clip far value.
I am currently trying to make some (hopefully) eerie urban levels. To achieve the floor reflection I use a slightly modified watershader from the library.
Indeed atmospheric. Just the ceiling texture looks wrong. You wouldn't epxect dirt "spikes" goint to the center of the floor on it. These spikes build up when liquid runs down the wall. You'd expect spots and stains on the ceiling.
and the ceiling texture could be replaced by a cracked concrete material e.g with some decals of piece of collapsed plasters. and similar dirt could be added to the floor.
and wall textures could be also more realistic, if instead of the grey colour, you would choose another mat colour, to show that it was painted a couple of times before, but the layers are worn out, for example by keeping those white traces on that ugly greenish paint used in hospitals formerly, like in the basement of the Technical University of Budapest 10 years ago .
and yes, the maintenance crew is definitely concentrating only to floor clearing
First, thanks for your advice! I added HeelX's SSAO. Also the ceiling is not affected by the light anymore and has a different texture now. The floor is not that shiny anymore.
Looks great. How did you avoid the ceiling ligtning? Iam asking because it looks like its just disabled for the ceiling. While the cone of the lamps is very small, the most upper corner of each wall is still fully lighted
Thanks. Yes I just set the ceiling to flat with a low ambient value. As I said I will change the lamp models. I think I will use some which are more "open". So I might have to reactivate the ceiling's lightning.
Very good atmosphere, but the floor is still too shiney. Like just cleaned and still wet and slipperly. I guess some relicts of your reused watershader.
I don't think it's too shiny. It depends on the material used for the tiles, and how well polished it is. I would blur the reflection. And if you add some dirtiness to it (which would be good), have the dirty parts less shiny (or not shiny at all).
mostly a door is bound by an doorframe. In your case it seems that the doors hinge is inside the other room. Thats leading to the conclusion that the walls have a thickness from not more than 5 centimeters.
OK ...no problem... the walls could be some sort of simple insidebarriers without any strong static structure.
BUT! .... the waterdamages at the walls shows that water ran through the ceiling from upstairs. So the upstairground doesnt lay as a single plane over the whole object. This 2 walls have to hold the corridorground from upper floor...both should have a thickness of 30 cm.
In other case if I open one of these doors the creatures of hell comes out and past me without attacking me. Because they only want out of this building very quick....they dont want to die through an imploding building
@sivan That is a good idea. But I think I will use those bulb lamps in the apartements of this hallway. For the hallway I might use something like this lamp.
@lemming & JibbSmart I think the reflection is fine now. It is also slightly distorted, so I think blurring it would not change that much (but I guess I will look into it). It actually looks better in realtime than in the screenshots. But I will add some decals for the ground as I did for the walls.
@gri Now I get your point and it makes perfect sense I will add doorframes and make bigger walls so the creatures of hell will survive.
Pretty cool effect. But there is a small cam stuttering when teleporting between the rooms (or however you made it possible to move THROUGH the portal)
PS: can you roughly explain how it works? just interested.
can you roughly explain how it works? just interested.
*magic*
(I hope my explanation isn't that bad)
Yes the player gets teleported, but its not 'just' the teleporting!
I'm using something similar to the system in portal. I've got 2 Views. One of them is placed on another position in the Level. It is exactly the same as the room you start in, but it has a tunnel in it, which leads to a second room. The second View has also got the same orientation as the first.
At this point I make use of my HDRR shader pipeline. Without it I couldn't use colors brighter than 1.0 or darker than 0.0 Theres a portal entity (portal.mdl). I set the color (with a shader) to RGB -1.0, 0.0, 0.0
Its pretty safe because it can't happen accidently that an entity has got exactly this color.
In one of the shaders I scan if the pixel color is -1.0, 0.0, 0.0 if so, then the pixel gets the color of the second view. if not, then it stays the same.
With this way, the 'other' room is only visible through this 'portal'. You can see the effect pretty good, when scaling down the portal.mdl
without this It would be impossible.
for the teleport i just use 2 empty models. If you're near enough you get teleported.
I also had to work a lot on the teleporting to make it 'seamless' It was caused by my player script which smooths the movement and also adds a force to the player, if the ground below is moving.
that's it.
But even with knowing the 'trick'... it feels strange to look at things like this
EDIT:
Quote:
But there is a small cam stuttering when teleporting between the rooms
I don't see something like that I only see a black seam at the right end of the portal model. But thats a pixel-position problem with the shader.
Ah, so you do viewportal blending aswell. Nice. Yesterday, after trying your demo i searched the web for Portal/view blending. Someone did it with a depthmap, but your color approach seems much "simpler".
You mentioned a HDR pipeline. What/How does it help in this case?
Idea: For your portal, i had an idea yesterday. For moving objects through. I thought, by deforming the collision shape of an object with stands on a floor portal for example, it should move thorugh the Portal(visually, a clone does the same on the other side).
HELP, iam thinking in portals
PS: For the stuttering: stay in front of the box, look directly into it(but stand in front and not in it). Now strafe left/right. rotate the cam a bit if desired. And when passing the corners(from your mainview perspective), it seems to stutter, while you're in the same room.
EDIT: So the stuttering doesnt appear when walking through the protal, but in the startroom.
Wow! Great effect! You can do some real nice levels with this. I once played a UT2004 map and wondered upon the really huge size it seemed to have, but it was relatively small and just got portals like these at the gates and always lead back into itself.
@Kartoffel: It doesn´t really work for me. It seems to run really slow and moving the mouse just very little makes me look somwhere completely different, same for the movement.
am i stupid...where do u see the framerate ? Which shaders ? I only see the portal ? I have more then 2-4 that is for sure. Edit2: Forget what i said about the camera, it was my mouse.
There is a cube model with the attached cutout material. It clears the given space, so the background becomes visible(by still respecting objects in front of it).
The cool thing is: The hole, while invisible, is still color encoded. A postprocessingshader can easylly detect it:
by checking for the colorencoding, its possible to differenciate the holes during the postprocessing. Above hole just filled with red.
EDIT: i had the idea for this cutoutshader after kartoffel explained is paradox demo.
Cool. I would like to see such an cut out shader on an alpha channel of an sprite. Would be great for making damaged walls or whatever. But i dont know if this would be possible.
@Gri all objects till the sky are cutted out. Otherwhise extra rendering with extra views might be required. Because the entity with the cutoutshader still write to the ZBuffer. Disabling it would result in everything overlaying the entity-hole.
EDIT: The technique behind it is pretty simple(maybe not optimal). The cutting entity just draws its colors with alpha = 0 to a 32bit rendertarget while writing to the ZBuffer. The final composed image is manually dran to the screen(draw_quad) to respect the alpha channel
@3run Sorry, I didn't ignore your message but I'm still busy on the Vehicle integration . Vehicle is quite different from prior versions so I still need to integrate some new APIs while keeping it easy for GS users.
I'll dig a little deeper on the destruction, after solving the problem on Vehicle integration (Soon).
Yesterday i added a portalrenderer ontop of my cutoutshader.
The blue point on the portal is for debugging and marks the closest point on the portal to the camera.
Why do i need the closest distance? We all know, if a portalcam moves backwards(as we do), it might pass a wall behind it and the obstacle is rendered. But we dont want it. SO i adjust the clip_near value of the portalcam at runtime to allow a freelook at my target.
see? The block is placed directly behind my portalcam, but the block is cutout by the clip_near value when moving backwards.
Sadly, it still has a flaw when the distance is too big and we rotate the cam. To much will be clipped away, because of the frustrums pyramidshape
Havent fixed the viewclippingproblem yet, but tried object transition. Using an action which creates a clone from its parent and moves relative to the portal, it went fine.
I runned demo with portal on my new PC (on which I've finished alot of games, such as battlefield 3 on high settings) and I got about 30 fps and jerking movement, but portal itself is awesome. Long time ago, I had an idea of how to do the same, but without shaders, maybe I'll give it a try. Anyway, is that portal shader so 'havy'?
3Run: No, the portalshader shouldnt be sooo performance eating.(maybe you are using to many portaliterations, the distance how far you can see through an endless portal tunnel) Can you tell me your rig o.O?
EDIT: keep in mind that a demo is most of the time created with older/non performant code.
I just runned the demo, and walked around and it was jerking (fraps shows about 30 fps when I run demo with it, but without it I still get same jerking movement and low fps). What you mean by rig?
Hi everybody. My last game is 'BallsMaster', Please check it out here < http://www.mediafire.com/download.php?0tcindfdngdklj4 >. It's open source. It's in romanian language( because my english is not too good ) but it's easy to understand and it is funny enaugh.Please try it. Please feel free to reply to me at my email < senior_nemesis@yahoo.com > Cheers!
please add more contrast to the heart and numbers. it's too difficult to read
Adding a black 'glow' effect (kinda shadow) should work pretty good, too:
@Rondidon
looks veery nice!
Edit: There's just one thing I don't like that much: lighting. The PSSM are nice but the diffuse-lighting looks a bit too... ( I don't know how to describe it ._. ) 'all the same' (or something like this)
I'm just refining my pathfinder (creating own maps for large sized units, better collision and avoidance), and making some crazy tests e.g. ordering 300 units in 10 groups to move to the same position: before:after: what do you think, if I have 85ms function processing time on my 5 year old core2 duo 1,6GHz laptop with 1MB RAM, what could it be approx on an up-to-date pc? (I have this when 600 units are moving in 20 groups using a common tile accessibility map)
maybe soon I'll make a separate test project, currently I'm doing it inside my editor because the 2 system are mutually related... hopefully I'll get a new i5 notebook within 2 weeks from my employer with a not bad video card...
Working on a generic Level Editor for Gamestudio A8 It can't compile to wmb, but you can script youre whole game in the editor with Lua or C# (i think i'm going to implement both of them, but i dunno)
(Click image for larger size)
EDIT: Forgot to mention, the PropertyEditor on the right is really WIP, it currently just shows the ENTITY as is and not how you could use it
Working on a generic Level Editor for Gamestudio A8 Since A6 days, it is one editor : WED If you need specific, tan yes you can write your specific editor. It is another try to replace WED ? Or a custom tool for personnal project ?
I want to simplify the usage of Gamestudio, it's not only a Level Editor but a (atm in)complete scripting environment. Two new screenshots (Click to enlarge):
@MasterQ32: nice! will you make an English version too? for wmb compilation you can check the codes of Emre, I use it in my editor with some modifications and works fine.
i know the wmb compilation code, i've tried to convert minecraft worlds to wmb but they were too large and yes, i will, but first i'm simply using the words which i type (could be english or german, depending what comes to mind first )
I also had problems with wmb compilation when too many entities are on the level, and I have normally too many entities so I use my own file formats... I don't know exactly what is a max entity limit...
I want to simplify the usage of Gamestudio, it's not only a Level Editor but a (atm in)complete scripting environment
Thats sad, Other engine have great workflow from years in their level editor.
Let's wait for some A8 new WED ? new worklfow ? or just a new skin ?
What is your point? What is sad? Why are you always complaining about Gamestudio resp. its level editor? Do you even work on games, at least prototypes?
I think sometimes everyone of us is angry at 3dgs workflow, especially when comparing to other game engines, where more types of 3d formats are supported well, or have better level editors. but 3dgs has its advantages of course, this is why we still use it. (atm I'm angry because of the not perfect .x support, what is e.g. a drag&drop in Esenthel engine and always fine beside another formats like ms3d b3d fbx dae... and these are little things could be solved easily)
that's a different category I think, 3dgs is targeting beginners and indies making only small casual games, while "Unreal Engine 3’s advanced toolset is designed specifically to accelerate developers’ productivity for ultra-complex, next-generation content." but 3dgs would be fine for nice indie projects with a revised workflow.
p_trace(Vec From, Vec To, var Mode, void* HitCallBack)//last parameter is optional
p_trace encapsulates c_trace. This function allows me, to trace even through portals. HitCallBack is called everytime it hits the endpoint or a portal inbetween.
I see no reason to use Gamestudio for Indie Games. No multiplatform. You got to do a lot of the stuff yourself. The workflow for non-programmers sucks. Plus, LiteC has some disadvantages vs C#/C++.
(But yeah, probably nice to do script-prototyping)
yes, it's not perfect at all (yet), I mean perfect normals are not perfect, so some randomization is needed... and playing with tree models is also necessary as I want lower the poly count.
Why not bringing some 10 years technology also. Your sprite shadow is good and fast, but for more little scenes, or non open world won't it be better this one ?
Working with view-space material maps. Also known as matcap-shading (greetings @ParadoxStudio), but I guess Pixologix will sue me if I call them so to advertise my game instead of processing the per-vertex normal for the cylinder hull (like ParadoxStudio), I calculate a per-pixel normal, knowing that the hull is cylindrical. That way, I get a per-pixel view-space normal, which I can use to fetch the view-space material map. It works nice! And is fast!
The caps are btw. quads:
I evaluate the per-pixel object space-position and clip the pixel, if it is outside the uniform circle. This technique guarantees me a perfect rendered circular cap + it reduces the amount of polygons by the half + a static amount of +4 polygons.
[EDIT] Plus, I store the diffuse in RGB channel, the specular & glow part in A, so that I can modulate with that grey material arbitrary colors!:
@sivan Overall the scene looks nice. But either the shadows are too dark, or the trees could use more contrast.
I would try 2 things: either increase the transparence of the shadow blobs and make them even more blurred.
Or try to alter the trees, in a way that the leave-patches are darker at the bottom and inside. And lighter on the outside. This makes the trees have more contrast. (outside shiny, inside more shadowy)
And the trunk of the tree should also be darker than the leaves, especially in the middle where the leaves are dense.
thanks, yes I used the mtl_hdr not too wisely, the truth is somewhere between the 2 scenarios, but playing with lights and materials to get proper contrast is not so easy (until I integrate it into my editor where it is in real time), the wind shader also decreases leaf contrasts a bit, and setting the leaf normals a bit differently also affects leaf contrasts... shadows are only placeholders, I'll create them from tree model view images (like in my fake tree shadow system for Unity free - sorry for mentioning the enemy )
No, not really. Basically you take the viewspace normal and fetch from the center of the texture, which I generated in Cinema4D. I came up with this when ParadoxStudio showed their work with cloning the MatCap technique from Pixologix (in ZBrush) - and they use it to render a complex material in e.g. 3DS Max with lights, ambient and everything the material on a sphere. The rendered hemisphere then "bakes" the material properties including lighting and specular and so on at every pixel, which has of course a certain view-space normal. In the approach, you use the view-space normal of your object to fetch exactly that pixel and voilá, you get the pre-rendered lighting from 3DS Max.
Problem is, that this is super-static and the illusion breaks, if the camera moves around the object. If the material is uni-color, you can use it, though, e.g. for car-paint materials or piano coating. You can also apply a texture pattern on it, but then it really breaks down, because when you move around the object, the pattern "stays".
Interestingly, I read yesterday a paper about rendering techniques of the Toy Story 3 game. They said, that they use spherical harmonics for the ambient term and they render that on a spherical map, with they then use as a lookup (see texture in the upper left corner of the screen). The benefit is that you calculate each term once and then you just do a lookup for each pixel, which has also the benefit that it can be instantly used in a deferred shading context.
I already prepared a scene in C4D with spherical texture map:
Since C4D can bake me the color, lighting, both, and reflection, I am going to try to combine them:
The goal would be to translate the viewspace normal to a corresponding UV coordinate and fetch there.
------------------------------- I'm working on a joystick managed inventory with a very noticeable Rage game inventorys' influence. Not really interesting but it tooks its time
When a new item is added, it is classified by a prestablished category, subcategory and alphabetic order in last instance. All the categories and subcategories that have not items are occluded and they result in a transparent navigation. Each category is a linked list. It is navigable item by item or subcategory by subcategory. The item navigation is accelerated when mantaining the joystick up or down. Each item has its own action menu navigable one by one. All the panel bmaps are rendered with bmap_rendertarget when something is changed.
@rayp thanks! I have more than one project always in my head and I work with modules, so it serves to more than one game. But the main project at the moment is that of the omnishadows I show you sometime ago.
Salud!
edited___________ I forgot to say that all the item icons are from Ryzom art assets, published with cc-by-sa license and the inventory is not screen resolution dependant!
@HeelX: I thought the idea is to speed up directional lighting by calculating the lighting for "all" possible view space normals only once per frame, e.g. illuminating a "sphere" in view space which gives you the material/lighting map.
@Error014: That's a cool scene, but what is the purpose of the beam lights ? And the scene lacks some ligtening ? (it is like it would need some lights or effects ?), perhaps it will look less monotonous in some vidéo ?
I thought the idea is to speed up directional lighting by calculating the lighting for "all" possible view space normals only once per frame, e.g. illuminating a "sphere" in view space which gives you the material/lighting map.
The idea of MatCap'ing is not the speedup, but more the idea to "bake" complex lighting and -material properties, produced e.g. in 3DS Max. In the following picture, I assigned a concrete material to a sphere, that is above a grass field, I added a sun with winter'ish light (cold blue) and a backlight with a warm feeling. When projected on the geometry, the complex lighting is preserved:
You can also do this in realtime by rendering a sphere in e.g. front of the camera and use that as a lookup for ambient terms, like I do here in a static fashion. -- The speedup is obvious, of course. Question is, if it looks fine, because point/directional light/specular directions are not corresponding anymore for objects - that is why this is suited only for ambient terms, if applied to everything. If you create local lightprobes, though, you could do a multitexture lookup and blend them according to the distance to the object. But this might create some overhead if done in realtime, which I don't want to do.
Of course it wouldn't work with local light source types like point or spot lights but for basic directional lighting where the position of the fragment doesn't influence the result it should work fine.
EDIT: And yes i know the edges on the terrains (wrong normals) And the thing to move the 3d-model (don't know the name) doesn't work yet, but i'm working on it
@mk_1: the left image is actually the result. @HeelX: jep, on the GPU. Concerning your missing waves: it's not tweaked at all, just the first results.
EDIT: ahh, you where searching for horizontal waves! Well, as you can see on the right site, the complex coefficients of the horizontal waves are blacked out, so this is intended. What do want with waves moving orthogonal to the wind direction?
Have you tried to apply a filter in fourier domain and reconstruct the image? How fast is your implementation?
Of course, that was the first thing before I threw the phillips spectrum at it: (right side shows only real component of the complex coeffients, not the magnitude (fault of stupid me two days ago))
Runs with 2.3 ms for a 512x512x64 image with FFT + iFFT. So it's nothing you want to do for a simple filter like a small/midrange blur. EDIT: on a nVidia GTX 560 Ti
Have you tried screenspace effects? The interesting thing would be to make a lowpass filter dependent on the screen-size, so that the resulting blur is constant in time and quality (blurryness) for different screen resolutions.
Not much experimenting done yet, since I implemented it for wave generation in the first place and that's what I concentrate on for now. Also the fft requires a power of two input image, so you had to rescale your render target anyway. But as far as I know 3DMark11 uses ffts for wide range bloom.
i use models and for the moment standard stencil_shadow (will change that later to a shadow mapping shader). There´re no seams cause the models have open edges. each side is cut off from the other one ^^
@Sammy: those textures are looking good, but a bit to steril.. Can xou add a more metalish look with Scratches and rusty parts under the paint?
I think they just need a scene where they look good in. If you use them for example in an active space station, there would be surely someone who cares for scratches and rust to keep it intact. Imagine the Emperor comes and all the walls in the death star are scratched and rusty under the paint...
I changed the way of downsampling the image before adding the blur effect for the bloom.
Before doing this I could't use more than 4x downsampling without getting aliasing problems (the bloom startet 'flickering' when the camera was moving which doesn't look good at all)
...but with my new shader I'm able to use more downsampling without much quality loss: I tried it with 40x and It still looked good
(40x means dividing the X and Y resolution by 40 - so the actual size is only 1/1600 (0,39%) of the original image!)
Here's a screenshot how the bloom looks with 40x D.S.
...oh I almost forgot: @Rackscha: looks very good. great job!
Thx! However, for deep water under the given lighting conditions (dawn/dusk) it doesn't make a big difference whether I do all these nice depth based calculations or just set a suitable water color as I did so far for testing.
@Hummel: Indeed a great job, but i had really no idea how to say that, so i found this nice picture which represents my face when i saw the video Would it be possible to create an endless water effect with this?
MasterQ32: I was actually quite amused by your "reaction".^^ The height field is perfectly tileable, but as usually the tiles become noticeable if you don't add additional randomness (there are four grids in the video sharing the same height field (I think Crysis uses also four tiles before they blend to their flat water)).
...After almost 3 hours I finally won the fight against my object shader and got normalmapping perfectly working. ( I need some coffee now )
Originally Posted By: Hummel
but as usually the tiles become noticeable if you don't add additional randomness
In this case I would use multiple displacement maps (or wherever you get the waves from) with different sizes which are unlikely to repeat very often, like 1, 1.4, 3.2 etc. and blend them together.
About 8.7 ms, e.g. ~115 fps. (moni-res: 1920x1200, grids: 4x512^2 vertices, ffts: one for height field, two in parallel for horizontal displacement, res: 512^2)
Thx guys. Foam is a really difficult to achieve effect, at least if you are aiming for a visual quality which is consistent with the water, but this is a good reference pic Harry (also for light scattering behavior). Regarding the animation speed, this is something I still have to tweak the way that it follows the rules of physics.
Hey Hummel, Screenshot is beautiful! Are you using this in a game where boats interact with the waves? I'm creating a boat racing game and am using a static bumpy plane with a water shader on top to make it look dynamic but I'm guessing yours is much better!
Oh god wizard attacks EVERYWHERE oh god oh god oh god. Guys, I didn't mean it when I made fun of robes! Robes look totally cool!
Quote:
That's a cool scene, but what is the purpose of the beam lights ? And the scene lacks some ligtening ? (it is like it would need some lights or effects ?), perhaps it will look less monotonous in some vidéo ?
Thanks for the compliment! There is lighting, but I admit that the scene looks a bit "flat". I think tweaking of the fog settings (which actually give the strongest impressions of depth in the game) should be in order! The tileset used is also a bit boring. It was just a quick test room to test the conveyer belts. The game isn't using the engine's lighting.
@Error014: As always a joy. :3 It's just an idea, but I think SSAA could heavily improve the image quality by smoothing out the aliasing introduces by transparent textures. SSAA is usually too slow for practical use, but since your scene doesn't use any computational intensive pixel calculations it shouldn't be a problem to render the scene at four times the resolution (2x2 SSAA). This wouldn't hurt the blocky style at all but gives you very consistent results.
Sorry but I've to show some of my graphics poser stuff too.^^° Creating normalmaps in frequency domain is awesome, since you can compute the slope vectors analytically and you can weight how much which frequencies contribute to the result.
Base texture:
Normalmap computed in frequency domain containing fine grained and coarse information:
And it's also possible to recover heightmaps from normalmaps accurately:
I think I've to create a normalmap tool using this techniques. The results are really good and it's damned fast: you could adjust a curve which weights the amount each frequency shall contribute to the result while watching the changes in real-time (as I wrote earlier, the whole process of bringing the image in frequency domain (fft) and back (inverse fft) takes about 2.3 ms on a GTX 560 Ti (fps = 1000/ms))
I'm currenty using a normalmap generator I found somwhere on the Wiki but the Results are not always that good and everything is calculated CPU-side so it takes much longer.
@Error014: Yes, that looks very good now compared to old colors/tiles Will you have some outdoor rooms ? like some trees among ruins ? Keep it up , and bring us some little demo one day ?
@txesmi : Could you put a REALLY REALLY TINY PICTURE in your signature please ?? Really it takes all my screen, and brings nothing to me, i prefer to have the place to read all users post instead of having to scroll a lot to avoid your BIG BIG signature picture ?
Thanks if you could od it, in the other caes, i'll pray to see not too much posts from you in the same page ?
this morning I wrote a shader looking for a leaking spray paint effect
Interesting! Are you drawing the figure by hand (and the shader does only the "leaking") or have you encoded the "drawing", too? How do you achieve the effect?
I guess you use a "spray paint" brush bitmap which is written additively into an accumulating floating point target. You then use another target which is the target for the shader and you just push the values above a threshold x down in Y-direction, in an infinite ping-pong manner.
Interesting! Are you drawing the figure by hand (and the shader does only the "leaking") or have you encoded the "drawing", too? How do you achieve the effect?
I drew the figure by hand and the shader does all the rest.
Originally Posted By: HeelX
I guess you use a "spray paint" brush bitmap which is written additively into an accumulating floating point target. You then use another target which is the target for the shader and you just push the values above a threshold x down in Y-direction, in an infinite ping-pong manner.
good eye! Yes, it is a brush bitmap rotated randomly each frame. The painting action is done over a floating point 32bit bitmap, as you guessed. The shader adds the brush values over it and glides down the overpaint. Everything is done in the same pass. I added a random left-center-right glide to achieve a saw teeth effect. I use an other shader to translate the previous bitmap to a colored and visible bitmap.
Here is the code, but remember that it is not a final version, just a workaround and not very fast indeed. enjoy!
Will you have some outdoor rooms ? like some trees among ruins ?
Yup, there's outdoor tilesets and decorational items you can place in your dungeons. Some things that could qualify as "ruins", too - though I admit not too many at this point. But we'll steadily add new content, the whole game is setup so that it can automatically update itself for that type of content.
use this minimalistic style, just different block colors for different towers and smaller blocks for the enemies
I thought about something like that, too
But I'm not sure... maybe I'll texture the blocks with very simple textures or give them slightly randomized colors to make the blocks look not all the same.
Anyway, that's not very important. Next thing I'll do is get save / load working and think about how I could solve the pathfinding.
not much to see here but: this is actually a Dungeon Keeper style map with a size of 2048*2048. Each block is a single entity. Running on a 3.4GHZ core of an i7
Iam NOT working on a DK game itself, just looking how far i can push it^^
PS: editing a mesh in realtime is a bit to much for me atm
this FOV thing could be interesting, probably storing pre-computed data could help if created when map is created, by storing farest tiles only to keep memory usage low, and flood fills are fast realtime - farest tiles are blocking floodfill.
its like you are suggesting Everything is precomputed and stored in the main data struct. The slow thing is the visibility computing. The grid shown in the video tooks more than half hour to compute. Too much for my spects. I am looking for a way to speed it up...
hehh that's really slow. what is the x,y grid size? maybe the computation should be less precise, like measuring only from tile middle to tile middle without taking into account slightly touched tiles...
128x96 cells. The main problem is that every line scan iterates with every wall line and I can't imagine how to filter them. My Known choices are multithreading and searching over the net xP
it's not so big, so imo it should be much faster even if c_tracing is used. and yes, there are some fine articles (I have never taken a really deep look into this topic, but will need something similar for my future AI developments)
Are you familiar with common visibility structures like BSP-trees, ABTs, etc? If not I would recommend you to read some papers.
Yes, I know what they are and how they work but never intended to code them. The documents I found speak about derivatives and integrals and it have past 15 years from I learned them. Right now I feel too lazy to relearn them, it totally looses its amusement
Hello there! 2 months of hard work and our 3d action / tower defense is 90% ready! There are some GUI elements that should be fixed, the high scores system and of course the sound effects... We're planning to release the first beta in the end of this month, or the beginning of the next! Here are some screenshots:
Well, you've got a point here, but if I use a texture with more elements in it, the blending between textures won't be very nice looking... The bad thing of a top-view game is that you have to choose the terrain texture very carefully. Even small inhomogeneities in the texture will be too obvious from the top view camera. Thanks for advice, anyway. I'll see what could be done here... Adding some more vegetation, rocks, etc. could make the miracle.
working on a system to draw lines that intersect each other and fits point and line positions for link creation. The struct screen draw can be scaled and moved maintaining mouse action.
not really a visual thing: inside my editor I'm just extending my AI and pathfinder functionality with real time building placement that updates area and tile data, thus the pathfinder will be able to handle more levels, more exactly ground level and 2 building floors (= gates, walls, and towers), with acceptable memory and hopefully minimal processing time increase:
so basically the 2D pathfinder will be a pseudo 3D system, just suitable for my purposes...
(the green flags mark newly created clusters: building areas or other areas cut off from their original cluster by the buildings)
sieht verdammt gut aus! Habe fast nicht zu meckern, doch kannste die Schatten vllt was dunkler machen?! finde die sind was zu hell geraten. Vllt die Sättigung von den Schirmen ein klein bisschen runter drehen, finde die stechen zu sehr aus der Szene herraus. Und evtl noch ein zweites Graß Model/Sprite erstellen, finde das sieht sehr künstlich aus , vllt hilft es auch bisse zu scalen, das ist einfach zu Perfekt von der höhe her
Danke für die Kritik, hab die Schatten und die Schirme mal entsprechend angepasst. Klar, viele Details fehlen bzw. es gibt noch sehr viel Potential, gerade beim Gras und allgemein im Bodenbereich. Wobei der Eindruck auf den Bildern auch daher rührt, dass in das Gras immer noch kein PSSM reingerendert wird und es deshalb etwas flach und homogen wirkt.
Im Spiel bewegt sich aber alles, Vögel fliegen am Himmel, Schwäne und Fische schwimmen durch die Gegend, Segelschiffe fahren durch den Fluss, Wind rauscht durchs Gras und durch die Bäume und dadurch wirkts lebendiger als auf den Screens.
I agree. The only thing you could change is making the lights a bit brighter. But don't make the rest much darker, it already looks very cool. Also, if you make it darker, all those nice details will hardly be seen.
I agree. The only thing you could change is making the lights a bit brighter. But don't make the rest much darker, it already looks very cool. Also, if you make it darker, all those nice details will hardly be seen.
If there is a flashlight ingame, I would make it darker. Then its much cooler if you light at the walls etc
Created a new SSAO shader. Uses only 5 samples instead of the usual 8-32. No Position Reconstruction needed. (Almost) no haloing. No Pixelshader 3.0 needed. Variable selfOcclusion (can be set by user).
And it's damn fast! ~270fps (no SSAO) to ~250fps (SSAO) on my old GeForce 9800GT
As I could not work on my sidescroller on vacation, I've worked on two other things: 1) Screenspace reflections, that means no additional rendering (except a depth view) is involved. They can be attached to arbitrary surfaces but have many limitations (only what is on screen can be reflected and so on) and, as a work in progress, there are still many bugs.
(a screenshot from my laptop, runs at 1024fps on my desktop)
2) A pretty fast 2D collision system that supports boxes, spheres and convex polygons with GLIDE support for my future sidescrollers:
@Superku : there is already BOX2D and other free solutions for 2D collision Perhaps you needed specific features, or had pleasure creating your own system simply ?
Yes, right, I forgot about Box2D. Nevertheless, I like to code most things myself as you have already guessed (including a quadtree-like structure for my movement functions and my own function scheduler, so there is no wait involved).
Problem with those physics solutions is that you have to hack them down to get standard collision/movement behavior working. I actually implemented a collision system based on the very same article Superku posted. This works good usually but also has it's drawbacks f.i. two projectiles with crossing trajectories (ger.: Flugbahnen) in a single frame will always collide if your implementation doesn't consider that case.
Because we can create our own framework. And also because we can optimize it to our solution. We don't need pssm shadows, ... But we need a simple solution for a flashlight, lightflares and many customizable lighting stuff in the video you can see some nice lightshafts through the door they are just some simple sprite model with one of our shaders attached (it's the same shader for the lightflares at the lights). the modification to use it for lightshafts took me about 2 minutes so i know my framework and know where to change stuff to get cool effects
to be honest, these winter pics are my favourites of your fishing project, and the water shader is also very nice, suits in the environment. I would be happy to achieve such a quality by myself...
Thanks a lot . The winter look is achieved by changing the textures. I did all winter-stuff by myself in Gimp and Photoshop Elements based on the original textures. It`s still not ready. The snowy weather particle effect is still missing and there`re some other minor things to do.
Originally Posted By: oliver2s
Sportangeln Nordeuropa?
Eastern europe. Featuring seasons and stuff. You can read about it in the Astragon brochure from the Gamescom (page 5).
Originally Posted By: Harry Potter
Iglo frozen fish?!
No, fish sticks. You can find them in dicarded, old moonbots
Implemented the ability to derive from a base objectshader in Shade-C. After that i tested the whole thing by writing a dissolve shader, including emissive color/bloom. Took me only 4-5 minutes, as i didn't had to write a shader from scratch, but only the dissolve part.
And what is it in reality? As I know you this is some fancy super mathematics stuff mixed up with shaders and PhysX - and the result just looks like a race track
Haha ^^ no, just eight models - two for the bumpers (inner and outer ring), one for the track, one for the sand, and three for striped parts and one for a dust part on the track. Handmade. No fancy stuff (except for the stuff inside Cinema4D, but that is a different story).
I knew it . Looks good. Reminds me on Wacky Wheels. The game was one of my favorites when it comes to battleing my brother when I was a kid. ( http://www.youtube.com/watch?v=0vVIiKp6xMs )
Originally, I wanted to make a mode7 game, but things took a different shape in my head. But whatever, I defined a milestone document for a first prototype of what I want to do - if it turns out that it is what I want and what I can do, maybe you'll see a Racing Game in the Future. The assets I am working with are just placeholders, btw.
Working on a SpaceShooter Game, like a mix out of the X-Series and Freelancer. Maybe i'll compile it in Multiplayer, but it's just a Prototype v0008 version. U can already click the asteroids and scan for some minerals or ore. If you do so, u'll get an Information panel about the value inside the Asteroid. Then you can take it out with this crazy wuhuuuu greeeeeeenlaser. Some shadertests already in there.
The Shipmodel and the Stationmodel ist NOT my work! (I'm no sucessfull 3D-Artist (u see the Asteroids? ^^))
started modeling the first Room. Not much done so far, there will be more Details on the Ground and some furniture Objects Hopefully i get much done today
Please ignore the ugly AO Seams on the ceiling, there will not be visible in Gamestudio (cuz i dont bake the AO into Texture :P) Also the windowglass is just a placeholder
Trying out some different modeling methods for getting decals on the track. I kinda like MoGraph, but it is very hard for a me to get used to all that advanced modeling stuff ^^
Sooo schön, hach... You should srsly consider to add SSAO since you'll have a lot of not directly illuminated scene parts, especially in situations without lights... .. smart!
@Superku: I have read the article. So for each pixel, the vector is calculated an the required pixel(based on depthmap) into the calculated direction is copied directly from the screen image?
So when you have a box floating above the water, you can not look under it by looking at the water? So minor perspective erros will occur?
@sivan: Thanks and yes, the water scene above runs at more than 330fps. @Rackscha: Yes, exactly, and there are many other issues and restrictions, too. It's rather a technique to enhance certain effects in a more or less subtle way, one should not put all focus on the reflection, otherwise all the limitations will become easily noticeable.
Two more shots, first one with normal mapping, second shot would normally require two additional views:
Yes, but as far as I remember they've added some random jittering to make the result a little blurry and to cover up noticeable flaws and mistakes of the algorithm. I don't know their implementation, but I think they only called it a DX11 feature because it came with the DX11 patch for the PC version of Crysis 2.
Playing around with a component based design pattern in 3dgs. An entity can have multiple components(scripts are also components). This is an example of a script component attached to an entity:
Php Code:
CPTransform transformation; //cache the transformation
//@TODO caching only works for one instance at the moment. need to add the possibility to store variables for every instance
void PlayerStart()
{
transformation = CPGetComponent("transform"); //CPGetComponent is always refering to components of the current entity(the one with the player script in this case)
printf("player initialized");
}
void PlayerUpdate()
{
transformation->rotation->pan+=1*time_step;
CPTransformTranslate(transformation, vector(1*time_step, 0, 0));
}
At the moment you have to attach those components per script:
The lights seems too much strong compared to reality where all the corridor is lightened even with more light near lamps, but perhaps it's a special ambianec or lightening ? There seems to be some black graphic artifacts on walls ? Yes texture projections to project lights or shadows, in some way as Doom 3 engine is a very fast way less expensive then real lights or shadows computation.
Finished my first runnable prototype for years. And yes, I borrowed the graphics from a certain game, and yes I know I'm a bad guy because of doing so and no, I won't go for retro textures and yes it will feature maybe robots and glass and laser lights and stuff
Wallpaper's bumpiness is rather small scale. Normalmaps don't help much to simulate them except perhaps you are using excessive filtering. Better option is to use a more suitable BRDF to simulate the effect micro facets have on lighting. But I think for their case it's totally fine to keep it that way.
Also my normalmapping shader can create a really good depth effect but we reduced the normal map strength because a normal wallpaper with a depth of about 5mm is not really real
@HeelX: Great Fzero circuit It was one of my favourite SNES Racer If you keep adding details, you can finish with some awesome track. Your beginning is top, just keep adding stuff.
@ratchet: F-Zero GX was the worst installment of the series imho. And no, I won't go for retro textures. Since F-Zero and F-Zero X are for me the best racing games ever, I won't clone them. If I will ever come that far, it will hopefully contain the quintessence of F-Zero - racing with high speed without weapons - but with something different as well robots and stuff!
@ratchet: F-Zero GX was the worst installment of the series imho. And no, I won't go for retro textures.
Have you really played it ? it was incredibly fast, really fun nad the challenges and cups were hard on the last ... I mean it's more some sort of personnal taste onec again
I didn't say anything about retro textures ... you use what you like I bet very well choosen or made textures and detailled enought for the track are the most important thing. And with some shaders (Glowing like night cricuit, reflective for futurist parts , specular for circuit outdoor in the sun etc ... ) you can obtain some great result.
But i htink you target more simpler and more accessible graphics specially if you won't make it in a team or won't recruit or adress level work to some skilled 3D artist freelance ? Keep it up
Yep the Circuit screenshots are the better. The jungle is oot much like some Rally circuit or Star Wars flying motorbike circuit , not too much futurist circuit ! I think you target F Zero style or some Wipeout style ?
You have a solid base, adding 3D models around the circuit shouldn't be so hard !
You have a solid base, adding 3D models around the circuit shouldn't be so hard!
The prototype was more a workflow test (Cinema 4D -> Gamestrudio), I fulfilled all requirements I wrote down and, no, it won't be recycled. I intend to make several prototypes for some key elements and then I will built the game from scratch, if I am convinced. That is the important thing for me.
Ty for your feedback. @Heelx: The main idea is to make 3dgs more unity like. You can extend your entity by simply adding components(and those have always the same syntax and can easily be migrated from project to project). Components can be nearly everything(particle emitters, controllers, ...). I now added the possibility to add singleton components. Singleton Components only run once per game(Input, Settings, Interface, ...). I also added the possibilty to set the execution order of components:
Last but not least, as you can see I started to implement a Input and frame based GUI component to 3dgs(will get like the unity gui, except that it will only run once per frame). The first parameter for the GUI functions is an id. This will be no more needed if jcl implements the __Line__ and __File__ macros for 8.6 (ty jcl <3 )
@Krial: really cool concept, there are fine things in Unity to follow and to migrate, and this modular system seems to be very practical! I use also nearly similar code snippets for my entities ordered into a few categories, only a few parameters differs, so your approach would really make easier my work
poor men's shadows in MapBuilder coming soon: added static lightmap generation (based on the Wiki example) in my editor dealing with terrain, objects without own shadows, identifying object alpha textures, blurring. there will be a menu to adjust a lot of parameters to look nice together with terrain shading and other shadows. especially for free/extra users, and for low end pc users where pssm is slow. I'll integrate next pssm in a slightly modified version. so I have to make soon another menu where the shadow scheme can be selected (decal shadows, decals, stencils, blurred stencils, lightmap, pssm in some compatible combinations)
Emre@ but that way, player's movement acts like PH_CHAR with physX, right? I mean, player himself doesn't have any collusions with rigid bodies, but can only push them (even big ones, which couldn't be pushed by player). Did you overcome that somehow?
rvL_eXile@ looks nice, but I have to agree with Superku about DOF effect.
Remove the DoF completely, IMO it has nothing to do in a first person game, it always hurts the visuals and oftentimes the gameplay. DoF can look nice in 3rd person, 2.5D, puzzle games and the like.
Emre@ but that way, player's movement acts like PH_CHAR with physX, right? I mean, player himself doesn't have any collusions with rigid bodies, but can only push them (even big ones, which couldn't be pushed by player). Did you overcome that somehow?
i don't know how is the physx but yes, player himself doesn't have any collisions with rigid bodies. that's not problem for my game. but if you want to collision for player, you can do that.
Here's a screenshot of my MMORPG tech demo. I'll hopefully have a public test for stability later today (sunday), so if you're interested keep an eye on the projects forum.
@eXile - How did you do the upward light flares? I first thought they were just sprites, until I saw one in the upper left corner wrap around a bump in the wall. I've been trying to get that perfect look, to no avail
it's just simple: use a spotlight with a really low falloff value + a simple pointlight this results in a soft spotlight. with adjusting the spotlight angle you get a really nice effect
Just managed to include the CEGUI system into a engine sdk project. Coolest thing is that there are graphical gui editors. I know that JenGs is working on a plugin for this, but his last post about this is was half a year ago. Are you still working on this?
If I'd only known how easy this was to include I would have never created my own xml-loading gui system..
Nice, will you publish a tutorial on how to include it? Is it possible to use different skins with CEGUI? And can you call Lite-C functions from Events of the GUI elements?
Yes, I still some how woriking on it. I have a full functional dll, with all helper functions included, for example for lists, events and d3d_switch and so on. I can upload it, but i was hoping Crazy Eddy would release CEGUI 1.0 in time. But, äh, you get it. I will open a thread in projects this afternoon (german time). I also will release the source code, because I never wrote a documentation.
Continued working on my component plugin for 3dgs. The GUI component got better. Here a Video: http://youtu.be/lk2JZ6bL2iQ (sorry for the crap music. First time using that videocapturing program and couldn't find a way to switch it off :P)
The code you need to do this is very simple and short:
I have finally managed to complete 99% of my SkullS game, after some music & end-level statistic tweaking I should be able to post the new demo here, on IndieDB (&ModDB) and also TurboSquid:)
yes, a tile is a compilation of up to 3 walls and a piece of the floor. (or just floor for center room tiles). every corner, every straight wall and every floor and roof plate is a single entity later, there will be 5 types of each tile, each one with a different roof formation, to get this wavely look of the DK2 room tops.
and i want to use complete rooms for different situations. like: instead of using 8x5 tiles for a library, i create a single model room with the same size (to save entity memory)
mk_1 is right though. Lots of draw-calls (lots of entities) slow down rendering imensely. You can get a lot of fps by combining models wherever possible.
My Test PC: Intel 2,5 GHz Quadcore 6GB DD2 Ram NVidia GeForce 560GTx 1024MB
It runs faster - each engine with an FPS increase of 200, when not using Normalmapping (except A8, cause we don't use it there right now)
The main Problem of combining entities here is the mapping system. From a 256x256 Block Map, we only show 40x30 Entities max, scroll them with the screen and morph them to another one if needed. There will be 32 different tiles for each room, rockway and hero-room. I calculate with a total of 640 different tiles i think it's not really possible to morph all them into often used block combinations
we will use a replacer system - like my library example. Build a 5x5 lib, and 25 Blocks will be replaced by 1 Lib model that covers that space, build it up to 8x5, and all blocks+the lib model will be replaced with the bigger one.
All those DK clones [...] make me want to make one, too
I think the game mechanic is very appealing to most programmers because the underyling mechanics are pretty easy, because the game world is very constrained and tile-based. Making it pretty is also not so difficult, given a good and talented artist.
Making it fast is a completely different story and requires lots of indian rope tricks. You can't blame Gamestudio or any other engine for not being fast enough for this, because they weren't written for it. Plus, the graphical detail used in the original game (in 1997) was very poor:
My guess is that they dynamically created, changed and skinned the 1) upper, the 2) lower and one or many 3) wall meshes. All other static 3d objects might be baked into one mesh, too and all other entities are animated sprites (sic!).
[EDIT] It also seems that they precalculated the lighting and put it into the vertex data.
Making it fast is a completely different story and requires lots of indian rope tricks.
Not true from what I can tell. I was fixed by this idea some months ago and did some prototypes. I did the straight approach and rendered a map of 256 x 256 tiles without any problems. For some of the tiles I even used 2 entities, namely for the lava and water tiles. The engine was running fine even when showing all 65.000 entities. With fogging enabed and a smaller clip range it ran perfectly fine, i.e. always > 100 fps.
i'm not creating a DK1 clone i'm working on a remake of the DK2 System, wich is Block/Tile based, uses single models for each Block and single textures instead of a texture map Problem is the high number of models to be created. My goal is the DK2 Gameplay + a Overworld Gameplay (wich was planned for DK3) + Graphics near the one in Dungeons (a game by Calypso).
EvilSOB is a great help here, and i'm proud of his help and the Progress of Development we made (he in the mapping system, me in the AI, builder, Modeling and Environment System).
But it's really hard to create it. I think we will think of a combining system and a precalculation method for the mapgrids and lighting.
@Uhrwerk: Since A8 uses an ABT-tree (and since you can turn some knobs to customize it's behaviour), your argument is valid as long as they run no function and especially when you turn off their dynamic flag and such tricks. Have you tried using shaders on the tiles and have you tried to throw in dynamic entities, that move and are animated?
@Espér: It doesn't matter if you create a DK1 or DK2000 clone, the underlying challenges are all the same and it gets even more complex if you want to throw tons of other stuff into it. But, I wish you the best
This real life thing steals all my game development time! Whats with that!?
I at least try to add a new tile/item/enemy or something small to the game everyday - the game was made so that this could be done rather quickly (so we can keep putting out little updates like that very regulary without spending much time on that - I want to work on other games sometime in the future! - and the client automatically checks and downloads those kinda updates. No biggie, yeah, but still neat).
So here's today's and yesterdays tiles (as well as previously existing grass tiles by the very talented TurricaN!). I already adapted the top potion of the wall so it's tiling is less obvious.
And yeah, that's right, I'm throwing snowballs in there because I can.
Because it snows the first time here this winter I'll post something relevant. I'm working on a weather system. The snow effect is based on the weather.c sample, though only the particle code is left in common.
Yeah, it looks like from 10 years ago but hey, it's dynamic and shaderless.
lemming@ it looks sweet are you going to provide source for community?
Thank you! When it's in a stable state I'll provide it as a lib, yes. Planned is snow and rain. Later maybe day/night transitions. All shaderless. I wonder how far I can get with that.
Nice mate! I'm looking forward for it to learn from. Are you planning to make collusions for rain? Plus, I've seen same snow effect in UDK, but it was made with shaders as far as I know, snow covers stuff even depending on wind direction, just as an idea for you
Nice mate! I'm looking forward for it to learn from. Are you planning to make collusions for rain? Plus, I've seen same snow effect in UDK, but it was made with shaders as far as I know, snow covers stuff even depending on wind direction, just as an idea for you
I havn't started on the rain code yet and I'm also learning, but collision detection is at least a target for rain and snow. The snow flakes "land" purely random in predefined areas by the model skin.
I find you put too much noise on your briket, you should put less lot caus it looses it's basic property that is shiny metal parts even with rusty parts !
Here is a lamp for you :
It's all basic hand painted, and just used some noise mask for the Blank part.
Still working on a project called [jiddisch three]
All it's game mechanics and rules are done, but it still needs a multiplayer and an interface. these digits at the lower corners are just temporary. I just want to use the lbgui.h GUI but it doesn't seem to work with my old A 7.02 (extra) and I'm still dont know how to update to a newer version, without to buy it. And so I've to reinvent the wheel.Therefore, I need some time to finish it. Im exited to present it next time to you with improved graphics, sounds, specialeffects and a gameplayvideo of course.
it's based on the gameplay of mill "mühle" in german.Everybody has to to put alternatly 27 edges of his color on the field. the randomly generated number in the center of the field shows the number of the star where you have to put the edge on it. the goal is to create a mill out of 3 edges. if this is done the triangle is conquered by his color as long as the other player reconquer it. if one triangle is conquered it can't be conquered twice by the same color but by the opponent, which prevents zwick mills "Zwickmühlen". Additionally if you conquer a triangle you're able to kick an opponent edge out of the game. if both player have set their 27 sticks they move the edges alternatly from one point to another, this creates new combos. the game is finish if one of the players got no more sticks left on the field.
Working on a sample Project for my CEGUI plugin. This will also be an editor in acknex. The Gui design is a win7 lookalike which I constantly develope further.
btw. The widget on top of the toolbar is a treeview.
I have released the wip version of the plugin. But you only have the header files as a reference. I am working on a full documentation. The wip-version can be found in my topic in "projects". but this version is not complete. I will upload a new version with some documentation in the next week. I added the support for treeviews, tabcontrol and some helper functions which are not in the current version.
@rvl_exile: I like it! There are only some things which don't look right:
- the table cloth of the upper table looks unnatural, very starchy. In reality, it just hangs down. - the knifes and forks are a bit to wide and flat. Especially spoons and forks are elegantly curved. They look at the moment like eating utensils in the Army. - the plates look very modern, but if those are for your hotel scene you should stick to a more classic look. - the wine glass on the pulled table cloth looks wrong for two reasons: first, there is still wine in it, second, if the table cloth was pulled by thrown chairs or zombies or whatever happened, it is most likely, that the glass fell down. I just would remove it.
Other than that it looks really, really nice. You develop a sense for atmosphere. Stick to your hotel scenery and keep working on it!
the rooms i created all in wed. random would be cool but as a non coder i have no idea how to do that for now this works and the workshop will be soon on my noob place. People can use it then to create there own games with. All i do is create small basic stuff to learn from
- the table cloth of the upper table looks unnatural, very starchy. In reality, it just hangs down. - the knifes and forks are a bit to wide and flat. Especially spoons and forks are elegantly curved. They look at the moment like eating utensils in the Army. - the plates look very modern, but if those are for your hotel scene you should stick to a more classic look. - the wine glass on the pulled table cloth looks wrong for two reasons: first, there is still wine in it, second, if the table cloth was pulled by thrown chairs or zombies or whatever happened, it is most likely, that the glass fell down. I just would remove it.
That's why i don't like and don't stick to realistic games, too too much parameters and real world rules to follow, specially for lonewolves.
Borderlands world and style for example make you free from reality, no more way to stick to real world ruleS. Same thing for cartoon games (WOW mmo etc ...) and other styles ... for indies it's lot more simple to follow
to follow up on whats been said here about the dining set, to fix the dark lines around the silverware, when you bake your ao, you might want to hover the silverware a little above the table, then put them back down when the baking is done.
That's why i don't like and don't stick to realistic games, [...] too much parameters and real world rules to follow, specially for lonewolves. Borderlands world and style for example make you free from reality, no more way to stick to real world ruleS. [...] for indies it's lot more simple to follow
I think that is a misconception. Those not-so-realistic games are just derivatives of what is known to most people, but with a twist to some things of your common sense-model of your reality. So if the game makes it plausible that you can shoot portals with a gun "onto" walls and you can walk through them, you still apply your well known model of physics to this exception, even if unconsciously. If those two things (new, non-realistic thing and known rules) fuse well and it is fun, you have a really good game mechanic or -look, like in Portal 1 & 2, or Borderlands. Compared to the shot RvlExile posted above, you find also some dishes in Borderlands, like in this screenshot:
Even though the scene is very strange, you can see a porcellain teacup and a bronze teapot on the table and handwritten place cards on the table. It looks like our super not-so-real freaks with those strange clothings had a nice tea party. Before that crazy girl blew the whole thing off. See? Even if you build a not-so-realistic game you have to stick to real-world things, shapes, dimensions and common sense to make it plausible. If very much of your game is not plausible and highly disagrees with the things the player is used too, he will most certainly reject it. Or won't have fun.
That accounts to all things and relations you know and therefore you can't say it is simpler to make a non-realistic game. Sometimes it is even harder if you introduce very innovative stuff that is in conflict with some fundamental principles people know and are used too. So, yes, making a photo realistic game is very much work for one person, but a non-photorealistic game can be the same amount of work.
Don't be fooled by simplicity. Making simple things really good is sometimes more time-consuming and much more work then throwing in a potpourri of the greatest hits.
thx chris... good example. i dont like the comic style much i prefer realistic artworks. maybe i start in a few days with creating the dinning room but before i have to model an teacup
But in the Borderlands style you can add some crasy item, without it sticking out weirdly.
Lets say a clown on a tricycle pulled by rats. Or a giant P standing for no reason on the steet.
It would not work in Crysis or Skyrim for example.
The more a game relies on (visual) realism, the more it will throw the player out of his immersion in the world.
-----
stylized graphics have the advantage to work with less detail and thus less chance to land in the uncanny valley when there is an error compared to its real world representation. Once the rules for the style are set (an followed) the workload reduces a lot.
For example making a character texture in Minecraft is possible even for non-artits without much hassle or experience. Doing that for a Mod in Skyrim would look off in 99% of the cases. Only a skilled artist can pull that off, keeping the visual style of the game.
Even if you build a not-so-realistic game you have to stick to real-world things
I see you are like a lot of people: limitated by all repetitive game patterns : FPS, RPG etc ... in Medieval, futurist etc .. worlds Game making is unlimited, your only limit is imagination in game design ... why not creating a world with giant tables ? Don't be so limited, and open your mind
-----------
Hopefully it exist people that bring originality and don't think like robots and follow traditionnal game design patterns and have good results !
Some examples :
just anotehr example :
Why not creating surrealistic world , with different gravity zones, only water would be dangerous , and you would be chased by chickens ? Following real world rules is not needed, it's up to you to follow it to some point. Realistic games makes that you must really follow up all real world rules : siezs, textures, 3D objects etc ... what noon realistic games don't have to care about at less or more degree.
I just hope game makers to keep some crazy ideas and make original worlds wiht entertaining gameplay.
And yes non realistic graphics can be lot more affordable for indie game makers (Zelda Wind Walker, hand painted WOW textures etc ...) !
Even if you build a not-so-realistic game you have to stick to real-world things
I see you are like a lot of people: limitated by all repetitive game patterns : FPS, RPG etc ... in Medieval, futurist etc .. worlds Game making is unlimited, your only limit is imagination in game design ... why not creating a world with giant tables ? Don't be so limited, and open your mind
I think you understood him wrong. If a game follows no realistic (and thus known) pattern you won't understand it. All your screenshots follow a realistic pattern. 1: Usual forrest. No blue/yellow striped trees, just normal green. Even the giant rainbow snail can clearly be seen as a snail. 2: Physics behave as you would expect it. I played the game. No new gravity rules or strange alternations. 3: ??? 4: Lego bricks. That pretty much describes Minecraft. What do you expect? You stick them together and they stay in that place. Nothing super "unrealistic" for a brick game. 5: A floating landmass. But besides that everything is normal. Grass is green, the player sticks to the ground and does not fall through it. The space is not 4-dimensional and you don't move in a giant frozen methane-block.
All the unrealistic things are just aspects, never the whole thing.
For example : some rythm games with some "Tron" style vehicle surfing on surface made by music notes are not realistic at all Action games using some sticks turning around where you have to avoid some sort of powers speheres is also TOTALLY Irrealistic
But yes i understand your point of view, sa&ying that a base can exist in non realistic games.
But i could make a game using only Green and Blue Spheres and your characte rwould be a red cube firing red globes ... where is the realistic ?
Well forget reality, exit from the matrix ... yes irealistic games exists hopefully !
I don't target more detailled than that for the moment :
Like Error014, i target simple retro style for the moment ! I'll add complexity by putting for example a second or three floors of 3D tiles and bigger 3D tiles (4*4 tile matching others)
Hand painting textures, modifying , adjusting 3D tiles = long hours ....
So much problems with edge extrusion, than recalculate normals outside ? unwrap on some faces, but the painting don't work, even directly on texture ? ? vertex problems ?
---------- More complex graphics = lot more time spend. So my rule now is : KEEP GRAPHICS SIMPLE (Like Dungeon Dities) ! I bet on gameplay, fun, cool world : places/NPC , puzzles, dungeons and fun than graphics.
Creating a small generator to generate maze-like Grids where entities gets placed on the maze-like grid which can produce mazed-paths for dungeons/ streets/caves/pipes or whatever else there is i have not thought about..in this generator i use tiles or cells if u will, i create random mazelike structure with it and find all cell/tile walls that forms corners and t-junctions and + intersections and straight areas to put the requered model in place as needed and the grid gets scaled on the size of the straight model segment and by the wanted number of cells ...so that segments join up nicely provided its moddelled acuaratly ..while experimenting i easily constructed a random generated small city with streets and buildings(on cell positions)
@ratchet: It looks like you don't have any ambient light. Thats why you get the extreme shading and black models with "unlit" . Check "ambient_color" (and maybe also camera.ambient) in the manual.
Also, I like your ground tiles a lot. I wanted my "testgame" to look something like this, but I'm so bad at texturing.
@Exile : the squarre peices on the floor are so cool, brings a real detail attention !
@SchokoKeks : Basic hand painted texturing is really easy (i made all painting under Blender until now) What you need is to know some basic techniques like what patterns to paint (wood, rocks, grass, metal) and how to give some volume.
-----------
I put my model wiht Unlit, and put a greatter ambient lightening, it's black ? (click to zoom in)
Sun , Ambient : Full White (255,255,255) Model : diffuse , ambient : Full white aslo With UNLIT The colors in game are darker than in teh editors WED and MED ?
MED preview engine preview : better
Solution : use lighetning for better colors popup !
-----------------
Problems and problems : i give up on Blender and 3D tiles for today and some days (tired)
Blender bug, :
I painted a trees, the texture were painted nowhere in no texture among the available ? but somewhere ahahaha ?? Texture painting mode : 3D models are displayed with textures, Edit/object mode : blanck textures ? ----------- the work is not lost, i'll have to take vertexs of each 3D tiles and save as object individually, before remaking a scene containing all of them !
What i learned : Save lot of versions of what you are working on !!!! Don't overwrite each time ! This way you have some history and can come back if something turns wrong !
@Pappenheimer: The puzzle pieces were originally intended to be "opposites" of each other. You'd place them in pairs, and then whenever one walked upwards, say, the other one would walk downwards, and so on. In the end, I just added them as normal "cannon fodder" enemies with an element attached. At least it's unique, and I like their look
Also, now you too can be unimpressed!
I may have to do some of those orb images again, but for the moment, it'll be fine.
Great How orbs system works ? you can choose one horb any time ? or are there some always active ? Do you find them trhought the game ? Remind me Follout 3 skills !
yes a flashlight is implemented yet and the flashlight is switched "on" on this screen. But the Light shading is not yet finished, needs more work on it.
Well I think it's time to show you a little screen shot of my current project. Still in a very early stage and the map's for testing only so I'm well aware of the fact that there's actually no bridge and that the textures look horrible and the sky's strange and stuff ^^.
But still I would appreciate some feedback. Thank You
Ehrm, I just realized that I didn't give any explanation ^^.
It's a Tactics Game like Final Fantasy Tactics, Disgaea, Tactics Ogre or Jeanne D'arc. Basically you deploy your heroes on the battlefield and then it's your turn to move them and attack your enemies. The Battle's won, after you defeated the enemies. But thats just basically the concept.
You'll be able to equip your heroes with weapon's and skills and stuff and you're following a storyline.
But well, it's in a pretty early stage. Most of the things i did so far is programming the map-handling, database handling, enemys and hero-classes and stuff. equipping items, map control and camera control. You're even free to play it with a Gamepad.
I just started to make a real testmap for programming the battle's.
@Roxas: Your level looks great, i loved a lot Final Fantasy Tactics and Disagea, and i don't have finished Jeane Or Arc that is great also ! I like the grass texture with shadows and lightened parts. Thats a tips making things look great (Trine games uses a lot of lightened spots on levels)
Will you use sprites or 3D models for characters ?
Thanks ratchet ^^ Well I'm not sure if I'll use 3D models or sprites, but I tend to use 3D models, because I really like what they've done with the models in Jeanne D'arc.
But I think I'll have to try out first, since I'm not pretty good at modeling and animating 3D characters. I've made a lot of sprites for RPG Maker once in a while though. But that's some years ago I suppose.
However thanks for the feedback ^^ I'll post more if there's anything worth showing.
Someone really tried to smoothly cut the metal bands. They should stick out a little. Other than that it looks really nice and I wouldn't mind if you sent me all the models
I really like your work eXile. Amazing textures, they really fit well together. I always have the problem, that the different textures of my models don't fit pretty well.
rvL_eXile, I know you quite from you start with gamestudio if I remember correctly. You always got better and better over time, but the progress you made in the last few month is awesome. I'm jealous.
And a small update of my project, since I had a little time I made quite some progress.
I reworked the texture for the cliffs. I just hand-painted it, but to be honest, I didn't take my time to fit it well, i just placed it there. like i said, it's a testmap anyway ^^.
I'm finally able to place my heroes on the battlefield and made an animated panel for showing their stats and stuff, which also shows the stats of the enemies when the cursor is placed on their position.
well ^^ I just started to create hand-painted textures, but I draw a lot of stuff. The face is made by me as well.
This is the Faceart in full size. It's a couple of month old though. Don't know if i'll use it for the project ^^. added it just for for testing the interface.
Thanks for your reply ^^ and I think I'll have a lot of stuff going down the road on the programming side before I actually start to make 3D models. But I think I'll start drawing some concept arts very soon.
EDIT: Don't be confused by the "(C) Dekurian". Dekurian is my main-nickname. But because of the fact that I love Kingdom Hearts and Final Fantasy I used "Roxas" here, because everywhere else that name is often already taken.
fewww : Your 2D art is very pro, nevre thought you could do such well done 2D work. The shadows on it are great.
That's perfect 2D for some narrative dialog above the tactic combat grid , or for some panel character showing stats, orbs equipment.
---------
I would like to know how works your levels ? It is made freely by hand , or using assembled 3D tiles ? Is the white grid made on top of a free made Level or made above some 3D cube tiles ?
Your grasstexture seems bigger , perhaps tilling, but not necessary aligned to cubes, but tilling along the whole level ? Or am i wrong ? it is 3D tiles with grass tile on it ?
Thank you . I think I have to practice first, before I'm able to create such drawings again. Didn't make any for months now.. I was just too busy with a couple of other things.
-------
Well actually the Map itself is one single MED file. Since it's Tactics-Game I don't need any real physics or collision.
So I started to make up a plane in MED. Each square is 32x32 units, 2 polygons. After that I just extruded the hills and cliffs out of the plane. Finally i made some simple touch ups by deforming some tiles so that it'll look more natural. I think maybe for the real maps I'll even add more polygons just to get more detail.
I also applied the skin in MED.
For painting the textures I use Photoshop CS2. It's a pretty old version, but I like to work with it.
The grass-texture is not tiled. It's just one big texture. It's 1024x1024. For more detailed levels I'll create tile-textures I think because such big textures need a lot of memory. But the big texture has a benefit, what's hard to achieve with tile-textures: it's pretty detailed and looks very natural.
The cliff-texture is a handpainted 512x512 texture which can be tiled. The gradient-texture around the level is just a gradient effect, 64x64.
the grid is just a 32x32 sprite.. and for each tile I create one sprite. Not very effective tough, uses up a lot of memory and entites. I think sooner or later I'll change this. But I don't have a good idea up until now.
aaand that's everything i suppose. ^^ nothing really tricky, no shaders involved.
i htink you'll gain lot of power using 3D lines drawing instead of sprites ? One trick can be to toggle on/off the grid also to gain some power? for example only show the cubes where the player can go.
Even your texture of 1024*1024 can fit and be good like that, specially if it was for some mobile game.
------------
I've played DrakenSang Online : They use 3D Tiles , they can put two or more layers of them in height. And for texturing it seems they have some system that paint multitexture in top of all 3D Tiles, so this is not a simple texture per tile, but like some big multi texture paint for the level made of tiles ? But i'm not sure.
------------
What will matter will be your characters, their attacks and special effects and all your 2D art, customizable things (characters equipment, powers etc ...) The level is already good.
Will you make Final Fantasy tactics combat style : zoom on the map, or Fire Emblem like : 3D cinematic fight ?
The grid itself can be toggled already, and for the moment I don't got any frame problems^^. It runs smoothly on 190 FPS, so I think there's no optimization needed at the moment.
I just think about adding more polygons, more detail like trees, blades of sprite-grass , bigger maps and as you mentioned above, special effects for magic and the likes. This map is just for testing, but also for testing my abilities and how I'll create the maps in the end. But I think some of your suggestions might be helpful for optimization.
For now I'll focus on the code. Up until now I think it's pretty dynamic and well structured. But I optimize a lot, too.
And for the battles, I think the battles will just be Tactics/Disgaea styled. so I think I'll just zoom in on the map, and the battles will take place on the battlefield. So there will be no extra battle screen or something like this. At the end, this project must be bearable by me, since I'm (almost) working alone, so I'll just focus on basic elements right now.
Don't get me wrong, your ideas and suggestions are welcome but a roleplaying game is quite a lot of work ^^ and I stopped working on several projects because it was just too much for one person. For this reason I restrain myself from adding to much features and things.
@Roxas : Don't get me wrong, your ideas and suggestions are welcome but a roleplaying game is quite a lot of work ^^ and I stopped working on several projects because it was just too much for one person. For this reason I restrain myself from adding to much features and things.
Yes this si the mistake we make a lot, we begin simple, but want to put more and more things and features. So we should make less features, less things in a first time , than complete two levels with menus, effects etc ...
Once done, perhaps add stuff and features ?
For mobile your level would be ok like that , perhaps some rocks and trees, but it would be ok.
I must also not forget that indie rule : Keep things simple
-------------
@rvL_eXile I would really see some game like that with your levels : Dark meadow But with perhaps old customizable guns and stuff ?
very nice, will look even better when you smooth and color the shadows, I'd suggest a orangish-yellow light, with a darkish blue shadow (low saturation).
@ratchet I planned some basic elements, that really should be in the game. ^^ But I also have some extra features, which may or may not be added to the game. This depends upon how the project will work out at the end, and if I'm able to add them afterwards. If I'll finish the project, who knows, maybe there'll be a second game. ^^ But for now I'll focus on this project, and keep you updated about the progress
Your level looks nice. But please, please, please, please, please do not use flat sprites for your chain model. Nowadays graphics cards can render millions of polygons.
Some years ago I made a chain model for a drawbridge. A chain model with 20 links only needs 1120 polygons, but looks much better than a flat chain.
@Harry_Potter Yes fully modeled chains looks much better but this will be to highpoly in my opinion. i need 2x 20 chainlinks = 2240 tris + the rest of the model. This is to much... the whole jail just have 5000 tris.
actually, I´d just go with modeled chains if it just means like 5-10k more polygons in your level, as without it looks soo 2000 like
Well.. actually a lot of professional games use sprites for chains up until today. And since the project of rvl_eXile is supposed to be some sort of horror game there'll always be some light only. So I think to go with sprites is the right decision to save some performance for dynamic lighting and stuff. I won't go for high end visuals just because there are graphics cards which can handle that amount of polygons.
I startet working on a small shader-pipeline editor. It is (or will be) a simple tool where I can create materials, render-targets etc. and link everything together by using an graphical interface and the program does the code part automatically in the background.
Just found an old project directory on my hdd. There was a small 3DGS-App: The Mandelbrot sample ported to a postprocessing shader for navigating in realtime over the set.
I created this long ago as I started to learn coding shaders:
//////////////////////////////////////////////////
//////////Postprocessing Helperfunctions//////////
//////////////////////////////////////////////////
////08.03.2008////////////////by Nils Daumann/////
/////////////slindev.wordpress.com////////////////
//////////////////////////////////////////////////
//temporary pointers
VIEW* PP_Temp_View;
//Add an effect to a views stage chain
VIEW* PP_Add(MATERIAL* Material,VIEW* View,BMAP* bmap)
{
//find the last view of "View"s effectchain and store its pointer
PP_Temp_View = View;
while(PP_Temp_View.stage != NULL)
{
PP_Temp_View = PP_Temp_View.stage;
}
//create a new view as the stored views stage
PP_Temp_View.stage = view_create(0);
set(PP_Temp_View.stage,PROCESS_TARGET);
//assign "Material" to the just created view
PP_Temp_View = PP_Temp_View.stage;
PP_Temp_View.material = Material;
//if a bmap is given, render the view into it
if(bmap != NULL)
{
PP_Temp_View.bmap = bmap;
}
//return the pointer to the new view
return(PP_Temp_View);
}
//remove an effect from a views stage chain
int PP_Remove(MATERIAL* Material,VIEW* View,VIEW* StageView)
{
//find the view with the material selected or "StageView" and the previous view
PP_Temp_View = View;
while(PP_Temp_View.material != Material && ((StageView == NULL)+(PP_Temp_View.stage != NULL)) != 0)
{
View = PP_Temp_View;
PP_Temp_View = PP_Temp_View.stage;
//return one if the stage doesn´t exist
if(PP_Temp_View == NULL){return(1);}
}
//pass the views stage to the previous view
View.stage = PP_Temp_View.stage;
//pass the render target
View.bmap = PP_Temp_View.bmap;
//reset the views bmap to null
PP_Temp_View.bmap = NULL;
//remove the view
ptr_remove(PP_Temp_View);
//return null if everything worked
return(0);
}
//remove all postprocessing from the view
void PP_Remove_All(VIEW* View)
{
if(View.stage == NULL){return;}
PP_Temp_View = View.stage;
View.stage = NULL;
View.bmap = NULL;
while(PP_Temp_View.stage != NULL)
{
View = PP_Temp_View;
PP_Temp_View = View.stage;
//pass the views stage to the previous view
View.stage = NULL;
//reset the views bmap to null
View.bmap = NULL;
//remove the view
ptr_remove(View);
}
}
//returns the pointer to the viewchains last view
VIEW* PP_LastView(VIEW* View)
{
//find the last view of "View"s effectchain and store its pointer
PP_Temp_View = View;
while(PP_Temp_View.stage != NULL)
{
PP_Temp_View = PP_Temp_View.stage;
}
//return the pointer
return(PP_Temp_View);
}
Of course, GameStudio Com or Pro is needed in order to run it. For the others I want to provide compiled binaries. Download
Looks much better now. The next thing you could do: think about PhysX.
Not for those chains... We will use PhysX mainly as a better collision system than gamestudios OBB (PH_CHAR + ent_move works much better than c_move) But those chains are just a nice graphics prop and most time (i think) you can't even get to them, maybe 2, 3 times.
Edit: @Kartoffel Will u share your shader (creater) app ? Just read again: material creator...means no shader creator right ?
well I think I didn't describe it very well... sorry for that It should be a program which allows you to create materials / views / render-targets and set all the values (like material.ambient_red) with an graphical interface and link the views to a final shader pipeline; all this without having to deal with the code - the program does this in the background.
This makes it possible that you only have to call one function (which got generated by the post-processing shader-pipeline editor) in the application you want to use it. Actually its main-purpose is to make it easier to connect pp shaders to a complicated shader-chain without having to care about the code-part.
...which means it is not a program to create 'click together shaders' - if you meant that. ( But that sounds like a good idea )
But at the moment I can't work on my projects very much because I've got a lot of other important stuff to do.
The basic principle behind it is that on every axis the texture gets "projected" onto the surface and the higher the angle is the less alpha it has. (enable the line #define DebugProjections in the shader to see what I mean)
@Kartoffel: Very nice and useful, though Hummel already did this some years (?) ago with his bacterium. I've had a look at the source and it's pretty simple and fast but could even be a little faster:
float zProj = pow(abs(dot(input.Normal, float3(0.f, 1.f, 0.f))), ProjectionPower); is obviously the same as float zProj = pow(abs(input.Normal.y), ProjectionPower); which can be shortened to float zProj = input.Normal.y*input.Normal.y; as ProjectionPower = 2 apparently gives the best results (the same goes for zProjx/y, of course).
I've had a look at the source and it's pretty simple and fast but could even be a little faster:
float zProj = pow(abs(dot(input.Normal, float3(0.f, 1.f, 0.f))), ProjectionPower); is obviously the same as float zProj = pow(abs(input.Normal.y), ProjectionPower); which can be shortened to float zProj = input.Normal.y*input.Normal.y; as ProjectionPower = 2 apparently gives the best results (the same goes for zProjx/y, of course).
oh, of course... didn't see this Thx for pointing this out
EDIT: about Hummel's 'Ehec' shader: I saw his shader but I didn't knew he was using the same technique
for far awaya soldier you could use some simple Blob projection shadow, you'll gain lot of FPS. Shadows are what east lot of power in 3D ! Some games you put them off/on and you can gain/loose half FPS sometimes.
yes, with decal shadows it is much faster, but combining pssm with decal shadows is possible only by ent_decal that requires a short c_trace for each unit in every frame, which becomes also expensive as unit count increases. I want to use c_trace for units only when they are walking not on terrain, I mean normally they are using stored surface data. but maybe there is an optimal combination I have not tested yet. maybe pssm shadows should be somehow clipped over a certain distance, but until now I could not modify the pssm shader successfully, maybe I need a more simple shadowmapping system.
they have no alpha because they have in this video static shadowmap created by c_tracing, only moving units, some models and bushes have pssm shadows.
of course, I can make the full scene with pssm, which looks much better, but that is a bit slower, and I'm trying to find some solutions balancing between performance and quality, to have finally a couple of options to be selectable.
sure, just check the alpha value of the hit pixel, which can be done using the hit structure and pixel_for_bmap or something like that. Then either decide depending on that already if there is no shadow, or start another trace ignoring the object and so on. I did this a long time ago and should be really easy to realize.
yes, there are some possibilities, I also made a version for it, checking hit texture type and dealing with multiple hits, it is a bit slower, and I simply used the faster for testing. but it requires large texture size like 4096x4096 (depending on terrain size) to look well with a standard resolution pssm. pssm looks nice if has a resolution of 2048 and 4 views, it works fine if there are only a few moving characters (max 50), but in case of e.g. 120 on screen characters like in the video, c_animate results in fps drop what I have to compensate somehow.
another way to apply decals as I did before, I can make acceptable quality from tree model preview images, but it requires setting angles according to Sun direction finally used in the level (it would be more easier if x and y decal size could be set separately).
so I don't know what shadow system mixture I'm going to use to ensure optiosn for lower system requirements too...
About the animations: do them on the gpu, gamestudio offers some easy to use functionality for that. That should speed things up a lot. You may have to modify the shadow shader a bit for it to work right, but it is definatly worth it, as it will be nearly as fast as without ent_animate then.
There are a few things to do though. The face-art is not changed according to the unit and the menu-entries will have the face art on them as well later on.
That's great. Strange when the camera moves at begin of video, it's like frame rate was very low ? perhaps too much entites, sprites and HUD stuff for 3DGS ? You should make some english menu for character actions now I prefer icon based menu , and menu at bottom alos it is lot more natural then left (for my taste) :
the framerate is not the problem ^^ it's the video recording software. I got about 260 fps during gameplay and everything runs smooth and soft. It's just I use a trial version of screenpresso and it's enforcing 20 fps and low quality.
thanks for your reply though.^^ I'm not planning to release an android or iphone/ipad version or something, It's playable with a Xbox 360 controller and the resolution is set to 720p by the way. I like icons too, but since i don't need touch panels i think it's fine this way.
^^ well I'm on it. at the moment I'm just preparing the menu for dispatching. But i'll use the same code for the character actions later on.
There is old games like Phantasy Star 1 to 4 on Genesis using icons, lot more clear and intuitive at first sight, instead of words (golden sun on GBA also and lot of others)
Well dod as you like But for a begin, it's a solid base you have there : Menus, character portrait, 3D tiles and grid.
you'll just need good characters classes. What will you choose : - Making all characters on your own - Buying some on 3D sites like Dexsoft - Buying or creatign a base character, and adding armore stuff yourself to make the different classes of the game ? - Using sprites ? (why not , but i really prefer 3D characters smooth moves)
Well as for the classes, I don't know if I'll implement character classes. In the moment it's planned more like Jeanne D'Arc with individual characters that can wield various weapons and skills. But that's not decided up until now.
I think I'll create the 3D models myself, a friend of me is helping me with the models, too. He made a test- / base-character today which turned out quite good. The textures will be made by me aswell. Just started to use a graphics tablet today.
Just started to use a graphics tablet today. yes, that's the best thing for painting, and i use Blender with Tablet also (mouse just to zomm in/out) , 3D coat and sculptris i sculpt with tablet indeed.
For characters you can use some basic body base, why not minimalist and put more effort on clotes items something like that :
If you want specific bodys, faces you'll have more work indeed ! if you use Low poly and good painting skills , yes you can also follow original game design or some Warcraft character like ?
Just started to use a graphics tablet today. yes, that's the best thing for painting, and i use Blender with Tablet also (mouse just to zomm in/out) , 3D coat and sculptris i sculpt with tablet indeed.
Yeah, i dont know what I'd do without my tablet. I could use a bigger one though, Ihad a bigger one, but it didnt have as many features as my wacom.
@Slin - yes, but that would require the Pro version to be bought, and I do not plan it until its price is not lowered...
@Roxas - use the free MSI-Afterburner pack, I just tried it and works fine. you can capture the video as uncompressed to keep frame rate high enough and balanced, and convert it by e.g. the free Microsoft Movie Maker.
@ratchet: ^^ like i said before I focus mainly on the code right now, but I already drew some basic concept and textures, for testing. Since it's isometric view, I'll use lowpoly models. detail will be added with sprites and well-painted textures. I'll take my time for that I think
@JustSid: really? how? using the matBones bone matrix array in a vertex shader is said to be Pro only. is there any other way to effectively animate a character by a shader on the GPU? I have not found any example of it... (currently I'm using vertex animated models because of better speed, but of course they are converted from bone animated ones)
^^ No It's not stolen either. I don't steal anything from any game, not even for prototyping. This model was made by a friend of me. He actually got some inspiration on low poly character models he found on google, but it was designed completely by him.
I did the textures so far, but I'm not finished yet. Poor girl only got some hair, face and boots texture.
I like doing all the stuff by myself, because this is what keeps improving your abilities: Training ^^
The tiles floating in the air were supposed to be a bridge. But since this is for testing only I didn't took the time to actually place one there. ^^ But I like the map, so eventually I'll rework the textures a little and place some more details including that bridge and use it for some kind of tutorial level maybe.
That's great. We just want to play the game for some of us So it will be some good platform gameplay ?
The breakeed columns are great you could have made somme highter ? I hope you'll put more special 3D models on levels above the tiles ?
-------------
The orbs power could you mkae them more clear ? for example the one displayed, it's hard to see what it is ? perhaps some more bigger or precise with more colors power icon ?
I have some technic question : Your 3D tiles : are they MDL models, or do you use WED blocks ? There is no visible seems, do you use some shaders ? Or does it works if you just put UNLIT flag (no light) ? Do you use several lights per level or one ambient one ? Does the lightening works on 3D tiles without showing the seems ?
Because i plan a 3D editor also for my game , if not too much complicated to do. I'm borred to loose so much time under Blender.
The breaked colums could perhaps be higher, thats true. I'm unsure whether the bigger ones should be seperate objects, though, or if the game automatically chooses an appropiate height. Perhaps the former is the better way to go.
You can put any decorational item (and non-decorational item, actually) on the higher tiles. This room looks rather "empty" right now, but there's nothing in the way that'd forbid that. Since all these rooms are testing rooms anyhow, I usually don't try to make them super-pretty, but you're right that they should be tweaked if they were to be finalized.
You're right on the orb thing. The yellow one (=the active one) is the boomerang orb. Since it's currently in use (see the boomerang flying?), it's "locked" from being used again, which is shown by making the graphic transparent and small, so thats why it is difficult to read. In normal gameplay, it looks the same as the violet (=the passive) one, and that is large enough to be playable in our experience.
Quote:
Your 3D tiles : are they MDL models, or do you use WED blocks ?
The tiles are MDL models. I have a nagging suspicion that the rendering could be sped up considerably by merging the (static) meshes, but so far, the game runs smoothly on my old not-too-great laptop. I do want this game to work on even the worst computers framerate-wise, and also try to minimize the memory used, though. It's nice to be able to reduce the nexus by quite a lot!
They're really just normal with the UNLIT-flag set, though the "sides" and "top" aren't merged (=connected). I color them if needed using the red/blue/green-parameters.
There's no engine lighting involved, all you see is very simple static "lighting" by using the pixel-instructions once on room-creation. Dynamic lights I simply "fake" by using bright sprites - there's no big use of those anyhow, so I get by with that simple approach. Entities can check that lightmap and I do so to color them if necessary.
The room editor is actually a php-script; you edit those in the webbrowser. You also move your rooms together there to form your dungeon.
I also prefer hight and smooth frame rate and super entertainment. And your 3D retro art is great.
-----------------
For your levels i see it's blacks screen behind on background Why not putting same color as the level or somewhat darker on background ? Or some very simple painted background :
I agree it's lot more work also !
----------- Yes UNLIT seems to work well with 3DGS, Perhaps i'll make some level editor also in some other easy language.
Working on a new Specular Bump Mapping shader for my editor.
It uses a specular map to add glossiness to the defined parts on the object. It has also LOD. The bump mapping and specular mapping is smoothly faded out in far distance to keep the framerate high.
nearly finalized alpha support for static lightmaps with blurring:
the only thing I don't know is why some textures are disappearing when converted by bmap_to_format from compressed dds to 8888... Edit: I know, bmap_to_mipmap is also needed...
now static lightmap with a tweaked alpha texture support looks much better beside character pssm shadows. I think I found all necessary parameters to be adjustable in my editor (accuracy of alpha texture shadows, blurring strength etc.). But it is quite long to make a nice hi-res lightmap... I hope I have enough time and patience to do the gui for the new features (mainly shadow related things, and more support for Commercial users), and release MapBuilder update before XMas. Downloads of the last version is over 100 within a few months, I think it's not bad...
Quick update: This Modelpack will be release (if i have enough time to work on it) end of the week by Dexsoft. This Package contains: -15 different Tiles (destroyed / non destroyed) -Small Objects (Pipes / Room signs / Roomplants /Doors /switches ...) -different rooms -Stairway Tile
This program im working on is pretty much a personal education endeavor. Turns out this crazy scale I "came up with" is actually the hungarian minor scale! I use it to write a lot of my music, as well as the augmented and half tone scales. I'm closing in on 3800 lines of code. As you run your mouse over the selectable scales they preview on the fretboard.
@Stansmedia : really cool stuff, it would have been great to have when I started to learn guitar playing, but it would be useful evem now! can I try it? (and the piano is also fine, as I plan to add some midi intruments to my music)
For sure man. I want to release a fully functioning free version. The foundation is pretty much done, I just got to finish a bunch of little things which really amount to a lot of work. Options menu, string count (4 - 8 strings), preset tunings (for open chord tunings, drop d, c etc... so you dont have to manually click your tuning). Saving "fretboard scenes", possibly having multiple scales on screen. Agh, it seems like it will never end. I also hope I can figure out a way to export the scale to a txt file tab. But if not saving a screen shot is just as useful.
maybe a reference tuner. like sound clips of notes to tune to. i would have no idea how to program an actual tuner. and yeh, im currently in drop G# haha. i might as well get a 7 string guitar. Speed picking gets hard when your strings are like jello, but bends are easy XD
Still working on my Lite-C CEGUI editor Video There are still a LOT of bugs, like the one you are able to see when I am selecting an element withing a FramwWindow.
Why I create an editor, while there is the unified CEGUI editor? Because that thing is ... crap ... . If you want to do anything more complex, it will crash. And you are not able to fill container like a tabControl Panel with it. And the xml files only can save flat hirachies. My editor is able to save Lite-C code (implemented!) and optional xml code.
I still don't have a documentation, and still have not enough time to write one. If someone want's the actual gui plugin + my Win7 looalike skin, contact me.
Just uploaded the video of my new project. Its a first person shooter made with the ANet plugin (big thanks to darkinferno) Hope you enjoy it! Also please crtiticize! http://youtube.com/watch?v=Na2kUdu4d_k
Thanks to me? for What? o_O? I don't recall sharing my FPS code with anyone, so am a bit curious to what's going on there but if I did with you, i'm sorry, I don't remember. Either way, the game looks quite clean though, so, good job.
Sry Ive written it wrong again. I always mix up your name and Dark_samurai... I meant him, peter soxberger for making this awesome mp plugin xD Thanks btw for your words and thoughts!
@Ch40zzC0d3r - would you mind showing this to both Treyarch and Infinity Ward, so they can see what a First Person Shooter is supposed to behave like?? It would be much appreciated. lol Good luck on the rest of the project and beautiful work so far. Absolutely amazing.
what do you think, if I have 85ms function processing time on my 5 year old core2 duo 1,6GHz laptop with 1MB RAM, what could it be approx on an up-to-date pc? (I have this when 600 units are moving in 20 groups using a common tile accessibility map)
Now I get 35ms on a i5 laptop. I know nobody else is interested in it
Yeah sivan totaly intrested , gs and its terrain screams rts game dev.. And besides , i dont see anyone else providing a base for rts gaming for free to others so you should be proud and F them if they dont show intrest in my opinion gs is perfectly suited for rts and your the only one i know of that also see it and only one i know of that explores the reilms of rts for gs , i myself are going to need something that can handle so much npc/units for rts , what would a battle be without Being able to send hordes of warriors on a war path
thanks for the kind words, I really need it sometimes. I'm not sure I would share my full pathfinder script package for free, only the editor that is suitable for making such game levels, and ready for rts game AI processing, or maybe a limited version, I don't know yet at all, there are a couple of possible scenarios... to be honest, I see now that the pathfinder scripts (a high and a low level) are the easiest part of my work , the map data handling, unit entity memory containers, group movement and behaviour parts are much more complex, and mainly responsible for getting useful results and acceptable performance. this part almost ready, but nearly useless without a unit level battle behaviour system, and without a powerful non-player AI decision making system, which is another heavy task. so I have things to do in future, won't get bored soon and I'm also continuously learning the engine features and how to create a working 3d game environment, to improve graphics etc. which is essential to identify the limitations I have to deal with to maximize e.g. unit quantity, level size and complexity. so have to badly count those evil milliseconds that sometimes drives me crazy!
Best way to learn is diving into a project ,guaranteed ! yes projects sometimes get to much but a little break away from it and all is fine again , whar ever you decide commercial or commersial and advertising versions , it simply has to show intrest because i am sure many love rts and need ssomething to point them in the right direction especialy when someone does not know where to start or how to do something , although i would encourage you to try and develope your system to be open to accept some custom ai because i think many ultimately want something that they have plenty of controll over to support their game vision but then again this is your baby were talking about here so dont lose faith in it and perhaps you dev it for your personal vision and create a great rts to show of what you learned and made all by yourself so far
I wrote a system which use a mdl file as a template. The mdl file contains a model with width 0. I place the template model along a spline and set the vertices to the right positions. After that I merge all template models to one model. What do you think, would it be faster to use bones instead to set the vertices with ent_setvertex?
When the system is more stable and faster, I want to share it with the community.
Dion, that looks great! I especially like the second screenshot, although a flat plane for a forest looks a bit too.. flat
@Bone: Merging all template models to one model is a good choice. I don't know if it would be faster if you use bones, but I guess it would be more flexible during ingame-editing. But I don't know either if you support that feature...
this afternoon I made a small menu for the lightmapper, what I made bug free last week (bottom right corner). hopefully in the followign days I can assign all the parameters, and make the save/load functions. the 2nd trace length slider is blurring treshold in reality, copy-paste issue and some rearranging is needed...
what I plan to add in future is delete/redraw of static lightmap only under the selected area. it won't be so hard but can be useful (but blurring can cause minor problems on area borders)
I put a lot of stuff into this shader - actually everything I know at the moment, from triplanar texture mapping to colored ambient lighing (threre are currently 33 #defines for all the settings...)
On the screenshot I'm mainly showing the specular reflections.
They are calculated with the lyon-algorithm. It also supports a phong, a blinn-phong, a torrance-sparrow and a trowbridge-reitz specular calculation.
I saw them somewhere on the internet, converted them into hlsl and optimized them.
If you like the specular lighting, I can share the code with you
Coded a simple Zombie crowd simulation 2 days ago (based on my pathfinding code from the official goodies pack). The Zombies still block each other in front of doors and the like but it's a somewhat dynamic pathfinding with obstacle avoidance and region+path based static pathfinding that works on huge maps:
The problem is that writing all these shaders takes a huge amount of time and work even for "experts" and therefore they don't want to give it away for free - if they ever finish it.
My shaders are far from perfect and not easy to setup / use. ...And I'm pretty sure that they're also not very efficient
The problem is that writing all these shaders takes a huge amount of time and work even for "experts" and therefore they don't want to give it away for free - if they ever finish it
Even with all shaders some indie would like to have, imm' not sure he will make and finish a game using them
You need a solid base : 3D models, great textures, level design (just look at Guild Wars 2 worlds) , characters , animation, 3D effects, good ideas etc ... Shaders are only put on top of that to bring candy visual stuff.
If you have bad textures, bad levels, bad characters, well ... shaders are useless.
Some other Engines bring to you a library of complete shaders, others propose some shader editor ... some are affordable , some expensive ... some for 3D artists ...
3DGS have a good shader base ready to make little next gen looking levels. It's to people to work
Even the best shaders won't make a game look good at all. As you've already said the base is the most important thing. And if the base is done well shaders can make it look much better but it's important to have something solid. ...And besides that, good gameplay is always better than shaders ;D
But i really don't like even good gaemplay with very bad or too much amateur graphics, specially whenelements and level don't match in colors and brightness ...
So lot of gamesin IndieDb ... Not all have a GOOD RETRO design and textures, or GOOD PRO visual quality ! Lot of games don't have enought attention on graphics, some feels like made by a programmer non skilled at 3D !
For people working hard there is room in IndieDb, Steam GreenLight
-----------
Yes, even normal map/specular shader of 3DGS can do marvellous things if you have a solid level , textures ,and characters ... We have enought basic standard shaders with 3DGS, but where are the games using them .... I only see SuperKu and SuperCan ...
So adding even Shade C next version or lot more shaders ... won't bring games !
But i really don't like even good gaemplay with very bad or too much amateur graphics, [...]
Yes, too bad graphics can make a game unenjoyable - even if the gameplay is really good.
On the other hand, good graphics don't make a good gameplay ( I wish developers of 'big' games like Battlefield etc. would think like that ) ...but if the gameplay is great, good graphics can make the overall appearance and 'feel' of the game even better
its the same damn conversation everytime.. *shakes head*
shaders are not required, but they'd be nice to have to fully realize environments that may be in the head of even a hobby dev. Every time someone brings up how nice a shader package would be someone talks about how shaders are useless. No, if you want to dynamically change the way a surface looks, they are required, pure and simple. cartoon games can even use a good normal map if done in some ways.
It would be nice if someone did make this, but as said its a lot of work, so i guess we can all agree that it would be great honor to someone who did. for instance boh_havoc, slin, etc. they have all done it, and im pretty sure they have the respect of everyone here on how much work they did for free.
Thank you guys! Feedback is much appreciated! I've watched a lot of Youtube-videos on stats-screens, I tell you that, and I keep coming back to Nintendo-Wii-videos, surprisingly. They really know how to design sleek interfaces.
Also, how did we just skip over this?!
Originally Posted By: Roxas
And it's me again
current developments: - successfully implemented a-star path finding, so that my units will find their way to their given destination.
This is just a first version, so there might be some tweaks or changes to it later on.
please let me know, how you like it.
cheers Roxas
Your project is super interesting! Forgive me for not replying earlier. I like seeing the progress - can't really comment on the music other than it sounds good in structure, though it does sound a little "MIDI-like". I guess the samples you're using here are at fault at that. Do you plan on making the music dynamic? I also like what I'm seeing regarding the art and style of it all. It seems like you do have a pretty clear vision of it. Also, I do want to implement differently shaped tiles for my game for such "cliff"-like structures, it's good that I can see that this would actually work!
~ ~ ~
There's not too much crazy progress in Dungeon Deities, but here's a shot:
Featuring: Debug text, debug lines, and a debug dungeon. Also, I've - once again - fiddled slightly with the camera settings.
Yeah, there's still a black background, but I haven't forgotten that suggestion - the user can actually select from a few presets, it just so happens that I never really do that and thus use the standard gradient (there's a very slight gradient from black to grey)
Your gradient shouldbe like sky and moutain : inverted. You have black at bottom (moutain) then go to grey at top (sky) , and yes some other color would be great.
The light blobs appear to be placed a bit off (to the left). Have you tried a smooth gradient for the light blobs already? I think it could fit nicely.
Btw. during the last semester break I worked on 2d stuff too but didn't manage to finish this concept scene:
Uni is killin' my time since then.
This is an older shot (before I've doupled the resolution) which occured after a failed shader compilation:
Thanks! \m/>.<\m/ Just to make that clear, that stuff is still in the unfinished state. Originally I inteded to show nothing until it is actually finished but since I've no time to work on it...
I presume it's programming boats and not graphics the goal ? it's os black, no colors , models looks very simple, well so dark, no real lights ... but it's programming ... so my opinion don't matter
Mobile game , shows the engine don't so much count some times
ratchet, why are you oftentimes so negative? Try to create such a scene yourself (with or even without the gameplay) and you will see how hard it is. Why the comparison with a mobile game again and why do you tell us/ him that "the engine does not count too much"? How is that relevant? Additionally, when your opinion does not matter why do you post it then? And please don't call every second thing (shader/ some system/ component), contribution or not, "useless", I really hate to see that word.
rayp nice stuff going on there when will i be able to test and kill your zombies ?
Graphics is nothing wrong with in my opinion. It fits the scene setting good. The problem with critics is always they forget we are all unique persons there for it's a matter of taste. \Someone's taste does not mean it's allways the right taste and visa versa respect for each others work is the greater picture i think
The screenshot is missing contrast, it just feels... empty and blunt. I don't feel like constant "oh my god so awesome" is bringing any benefits, we can only grow by critique. But still, what ratchet is doing - calling every submission "useless" - is way more harming than the most harsh critique. No submission is useless. But a lot of ratchet's posts are.
@Superku: and you will see how hard it is I already know, the incredible amount of time, of design if you go further than retro or some minecraft style. But Rayp's screenshot is very black, even textures looks so grey, non detailled, there is no lightspot ... well it's juts my taste.
@Rayp : I have some very old models i've done in the past : Soldier futurist Goblin thread They are not awesome, but they have details and uses normal maps and have some good details. And i've done one or two for you that wasn't so bad for some abandonned cartoon mobile roject.
Sorry, but even programmers can find good textures and models for free sometimes ..., but i understant you put some screenshot showing you are on gameplay ; even if gameplay is hard to see on screenshot
------------
Ok i should not really post for nothing and criticize , even if the goal is to push people to improve their project. That's my goal, but i see that i can demotivate people, specially when graphics is not their main purpose and targer gameplay first. So i have to post lot less more, show what i do instead. And only say: "Great, good etc ..." when i will see cool things, and i'll pass on projects that will not attract me in the graphic part, like lot of people do
work in progress. just for fun. ~3hrs in Blender 2.64, 1140 tris + 2x smooth modifier (neck is not connected) My first-ever face model. note that i am not a 3d artist and doing this for fun. Constructive criticism welcome.
@ratchet: I think it the main problem is that everyone has its own style. Imho the screen looks great. And it IS the lack of contrast giving the screenshot some kind of reality. Were you ever out at night with low light? You can't see that much, only schemes, no hard contrast (in case of equal lighting) The screenshot looks much scarier than the one you posted, so @rayp: Good Work!
Hi! I have been asked for help for a mix of two shaders with a pixelated effect, and I saw a great oprotunity to apply all my studies about volumetric lights. Given its low resolution and its foggy taste I thought it should work and it works!
My NvidiaGTX285 spends less than 1 millisecond to render 5 lights and no more, but more than 6 ms an AMD card
VIDEO with all the effects in motion in a higher resolution
Restructured polygons on forehead,cheeks,chin and eye sockets. remade lips and nose. added the back of the head and neck.
1290 tris(eyes not counted) --- Blender question: Model is mostly made of quads, ctrl+t or exporting fbx triangulates the model. It's mostly fine but some quads are triangulated in from incorrect corners, do i have to triangulate them manually or is there a way to ask blender to triangulate them in a way so that the triangulated quad is concave not convex?
I made some planets. Pretty happy with the results from the workflow that I created. Some planets are not that pretty completely zoomed in though, was still toying around with textures. All planets are created using only free concrete and stone textures as resources on texturez.com.
I'm doing some tests again: 300 units in 10 groups are moving smoothly, but for 600 units in 20 groups, with pssm shadows, I got now about 14 fps when all is moving (when standing, it is over 40). Vertex animation is used, it's not as bad as before, if heavy lod applied: 400, 250, 125, and 75 tris. I found it faster to use pssm than decal shadows, over a certain quantity.
maybe I should use physx instead of c_move to get better performance?
Î Saw this technique in other Crytek-games, too (Crysis: Warhead for example), mostly used for vegetation. It is a really good idea and there's almost no visible difference between actual polygons and a sprite.
EDIT: just noticed that if you're animating a high number of models it's going to be really performance eating. Think about disabling the animations for distant models - or, as ratchet suggested, use (not animated) sprites
Started to work on an RTS myself a few days ago as a pet project next to Superku, pathfinding and group movement already work on a playable yet imperfect state (so does fog of war including hidden units, the minimap, group to hotkey and so on):
@sivan: The framerate problem in your screenshot obviously is a result of the script, not the rendering (29.4ms for code, 7.1ms for rendering). You shouldn't use a while loop for every unit but a single loop that executes all unit movements/ actions, this will speed up your game notably. Then you should not use c_move too for performance reasons. Luckily you don't need it because you can simply use 2D sphere-sphere collision detection instead.
thank you all for the suggestions, I'll try them. morhping into a sprite also costy in 3dgs, considering fast and flexible camera movements, but maybe I can solve it for vegetation and units too. from Fragmotion it is also possible to export animated sprites...
@Superku: I don't fully understand how to apply a 2d sphere-sphere collision, but you are right, it woud be enough, even if units are moving on complex 3d buildings. and yes, my code also needs some performance optimization (I've already done a lot in this area to minimalize pathfinding costs). at the moment I cannot see how to solve to avoid the usage of while loops for each unit, but it seems to be a walkable way too.
That's a good point to use one loop to manage all units indeed. For collision, non oriented box to box collision is really enought (this is not precise fighting game collisions !) That's lot more faster than math sphere calculations.
For 3D models, and sprites you should have pre calculated sprites indeed, and not calculate them, that's better performance indeed, and you'll boost a lot the speed using sprites. (this is 3DGS ,not a AAA speciallized RTS 3D engine that can run thousand of units)
(this is 3DGS ,not a AAA speciallized RTS 3D engine that can run thousand of units)
Also Gamestudio can do this, you only need to optimize all stuff. Also circle colission shapes are much easier than bounding boxes to calculate but it is a lot of work to implement all of this
... Vertex animation is used, it's not as bad as before, if heavy lod applied: 400, 250, 125, and 75 tris. I found it faster to use pssm than decal shadows, over a certain quantity.
maybe I should use physx instead of c_move to get better performance?
Bone animation is better imho, try with bones.fx shader, like in the crowd.c from sample folder, but you need the com version or above.
yes, i forgot, bones animation are done on 3D graphic card and not CPU. Reduce bons per character is a good solution. I don't know if 3DGS supports different bones number by LOD ? If not you'll have to manage LOD yourself (a pool of 20 pre created entities on very low LOD), than display and move them from a non visible position to terrain position only when need to display them).
units are in multi-dim arrays currently, organized in groups, so iterating them is easy, but it would require rewriting completely the unit loop, but definitely it would save some time of similar calculations. maybe limiting further getpath function calls and cycle lengths based on actual performance would be also beneficial, allowing better quality when performance is okay...
Well caus i don't use lot fo characters at same time i never noticed that ... Pro only ... hummm ... hummm ... a so basic technic that exists from lot of years ... i prefer to shut up ...
Yep for paths , you should calculate them every 5 , 10 frames or more ... like people using navmesh, you don't callfunctions each frame. It won't be noticeable at all and it's not some racing game or Fast paced FPS game also ...
Ah, sorry I suggested from crowd.c header, but the manual say A8Pro for the gpu bone animation.
Quote:
//////////////////////////////////////// // Crowd demo // (c) oP group 2010 // Requires shader support for good frame rate // (A8 Commercial Edition and above) ////////////////////////////////////////
no path calculated in each frame, only when units step into a new tile, or when an unexpected and unresolved collision happens. and paths very short for 80% of soldiers, depending on formation, they follow their fellows in the previous line, normally in 2 tile distance. only the first line units, the leaders, make a bit longer path searches, about 20 tile distance. and each search is cycle limited too, in case or errors. so only a minor part of the full high level path is processed by units as tile level searches.
it is implemented. but I found a bug in my script. the getpath calls are limited by proc_status , but it does not work as expected. now I use a global counter that shows the exact result. if limited to 25/frame, tha 25-30ms drops to 15ms. but it requires a priority value based queue to handle well whuch unit can start a getpath and which cannot.
Joozey, the two planets on the left could do with more distortion of the texture. Have you checked out the gimp Planet tutorial? http://mygimptutorial.com/the-ultimate-gimp-planet-tutorial I found it to be pretty awesome, the results are great and it's quick and easy.
I used it to make all planets in the background of our site: (They change once every month.) http://csis-game.net/
further testing: without pssm, 600 moving units results in 30-35 fps, depending on camera position (it would be 13-15 fps with pssm):
so I modified the pssm shadow function not to process far shadows, and together with reset/set SHADOW flag of units based on their camera distance resulted in a minor improvement (3-6 fps):
as I realized, the displayed milliseconds do not show exact values when pssm shadows are used. if I calculate the fps from them, it should be much higher. but there are significant differences in polygons per sec, and in the quantitiy of visible entities and triangles. I don't really understand why time_client+time_server values are increased when pssm applied... I should learn a bit of optimization. but meanwhile I started to deal with code clearing/optimization of the unit and group update loops.
The palisades' color should match the color of the window shelves of the ground floor. Since you decided to have that one wall demolished, you should put some of that demolition on the ground.
1) clipping decal shadow dots enables a bit lower system requirements than in case of pssm shadows, together with static lightmap (unfortunately there is an engine bug when mirror view applied for water, resulting in unupdated shadow positions):
2) using lod3 stencil shadows with static lightmap is also fine, maybe even better (unfortunately there is an engine bug resulting in too early clipping, at max lod0 distance):
Germanunkol: yeah the tutorial I follow is quite close to the Gimp one. I used too low resolution for the textures on the lot of the planets. The ice planet on the right isn't too good either. It's best to have the texture at least twice as big as the final planet result to get a good quality texture after spherize effect. Thanks for the link! Your planet looks beautiful on CSiS .
I still haven't found a nice way to create continents though, the tutorials you linked and I followed result in one solid texture, and the colour of the planet is a separate mix of layers which might give a faint hint that there are oceans and planets but if you look at a picture of earth, the tutorials never come close to the variety that a true earth planet has.
You see lush, deserts, oceans, cloud patches, shallow waters, mountain ridges, it's all well defined in reality.
@joozey, Im thinking you could get the effect of all that definition with some brushes made from rock and cloud textures. continents could start with a base color with a polygon selection tool, and then brush in the defined features, then a layer over the top for clouds brushed on, maybe even turn up the contrast on the alpha mask to make the edges more defined.
Joozey, I know exactly what you mean, I had the same problem back then.
One way that springs to mind is: - Download a rock-image like this one: http://cgtextures.com/login.php?&tex...bg7ra0eg1m1flg6 - Duplicate the layer - Blur and increase contrast on one of the layers and then use it as a "layer mask" for the other layer
That doesn't give you a perfect planet texture, but I think it's a pretty good start. Result (I also played with the "curves" tool before setting the image as a layer mask):
The scene above uses 3k Particles. Anything more screws the FPS < 60. Though the problem here is that 3dgs particles need a function..or is there a way to create a particle without an always executing function7event and never dying?
Looks great! With A8 pro particle instancing is used so that you can show approx. 30k particles in your demo. Besides, try to pass NULL as function and step each frame with ptr_first and .link.next through all particles, updating lifespan to something >0, e.g. 100, and updating the .xyz position.
I am not 100% sure, but since wait(1) is not allowed in particle functions, I guess they are executed at once in one frame roughly as I described above. So, my tip can also likely be no improvement. I guess the main reason why your performance breaks down is just the casual overdraw.
Getting the first particle is a bit tricky, because you need a particle instance for ptr_first. I see two ways: 1.) create a dummy particle that is alive all the time and writes it's pointer into a global one, this can be used with ptr_first:
or 2.) you give the very first PhysX-particle an event that overwrite that pointer. If the particle is destroyed, you have to reset a flag or so to indicate, that you give the next PhysX-particle the function again.
Nevertheless, you have to indicate the PhysX-particle somehow, because it might be the case, that there are other particles as well, which don't belong to the fluid. I would set each's Physx-particle's skill1 or so to an identifier reserved for PhysX particles. In the loop, in which you update the particles, you check against this skill, and if and only if it matches the ID, you update it.
But i think in this case instead of using the LinkList, i'll just put each registered particle into a Particle array, which allows faster iterating(no check for non Phys particle). This list is then directly given to my DLL to do the stuff. Might even kill 1 of 2 copy loops(resulting in a speedup)
working on an very easy shooter template that holds the pure basic stuff. I got help from my friend Arrovs a few minor things need to work perfect and then it will be uploaded a.s.a.p. For me the reason to create this is that in the old days of A5 i liked the simple click and play thing as a Noob. Soon A8 users will have this simple click and play to i hope
after correcting my stupidity, and lowering cluster size thus enshorting unit path searches, lightly blurred stencil shadows work fast enough for low poly units and buildings, and looks fine together with static lightmap of static objects. it seems to be a good alternative of pssm shadows for slower pcs.
it seems to be a good alternative of pssm shadows for slower pcs.
Talking about performance I really reccomend you to work on the script / unit-management! 610 functions running at the same time are clearly too much. So before adjusting the balance between graphics and performance to get as much fps as possible you should optimize your code. That would result in better performance without worse quality
thanks, of course, it is also under development, I'm just cutting up my spaghetti, and making it much more modular, then I'll be able to decrease loop quantity as it was recommended by others too. only group and map updates are optimized. I simply had no enough time for it because of my job, and had only time to play with some graphics performance optimization solutions (few-minute-long modifications and tests). moreover, I also have to redesign my entity and database structure to be more flexible, which is still in brainstorming phase. the current system is a bit rigid and okay only for a simple RTS battle game, not for an age of empires 2 like RTS.
@RealSpawn : i like a lot your simple and ready to use Template, ideal for quick level test, or to begin some FPS base game before improving it !
@Sivan: I would want to say : go for C++. Just Call 3DGS renderer and level loader ! An better why not C++ plugin : Call to Direct X hardware animation handling as you don't have the pro version.
And yes animated sprite for soldiers indeed on far distances ! It's only some quad , incredibly less CPU or GPU than a real animated mesh (even if low poly).
@ratchet i am tweaking the last few things. it will be soon online as it holds all the basics with simple code to use and study.
@rayp I got the joke but you know me, i always want a few extra's to make it more appealing so i made some crater sprites and some simple explosions. Put my own cool picture in it have ammo number and health pickup and real bullets working elevator and gates so all to create a very simple shooter and learn from it
Changing material settings and coding some ambient and athmospherical switches to increase the athmosphere of a level.
But now I'm too close to see if I'm making progress or not. I'm grateful for any feedback and suggestions! Keep in mind that I'm interested in suggestions for all kind of settings, not only for a forest at night!
Top: Before Bottom: After.
I guess it might be too dark already. Augh.
EDIT: Second shot:
I'm also having some slight trouble with draw_obj, but the bigger problem is using PP-effects on a camera-view when camera.bg is set to a transparent value - it goes insane there. Has anyone encountered that problem before?
It's way too dark, I have to get really close to my monitor to see what's going on. Besides, esp. when I look at the first screen I think the black background could use some work. Maybe you can make it as a dark circle gradient with low contrast and its "bright" spot in the center (that is covered by the map/ room), then tint the gradient by the average color of the map's tiles? I am thinking of something like http://en.wikipedia.org/wiki/Ambilight.
@HeelX: Nice! Is the shader physically based in some way?
@Error: The fog does it's job. I would recommend you to use a simple ground fog too, though. However, the fog is a good tool to give feedback about the lighting conditions in those scenes, f.i. your bright forest at day fog settings imply that there is a clear sky where for your night scene it feels like there is not much moon light reaching the ground (cause could be a thin cescent moon or clouds). Darker fog for a day setting could therefore imply an overcast sky and rainy conditions, white fog for night scenes a bright full moon.
Thanks! And nope, it is not. It's a bit tricky to explain, but basically the screen is divided into squares and each pixel checks it distance to the square center. The distance measures the brightness of the "CRT-dot", which is controlled as well by the luminance of the pixel and optional gamma. This is not only done per pixel, the results of the surrounding 3x3 area are summed up to achieve blooming between CRT-dots.
In the past weeks I spend my time on shaders (making new ones and improving old ones) with the goal to get some tasty graphics. And I have to say, I'm really happy with the result
It seems to be a simple normal map/specular shader ? ( No offense but i see the same result, perhaps you have put some special stuff or optimisations ? )
I like the fog effect, add a good ambiance I thought, you could do some "Cannon Fodder" style game instead with some RPG elements ? (less characters at screen etc ... )
it's just a default scene with default settings and with a few test objects... I don't really like action games controlling a few squads, after a few missions I get bored. but I won't have problems with a few hundred units, especially because when my game will be ready, pc-s are going to be much faster than now
further minor editor improvement (beside some others), load/clear material effect for easy integration of own simple ffe-s or shaders (but later I have to change the material management to be more flexible):
Thanks for the feedback, guys! You were definitely right, it was too dark. I was unable to get a better feel working without resorting to shaders. So now I'm a little more flexible, AND have to offer a new option in the menu screens. GOSH.
Anyway, I'm still struggling a little with the night-settings. I assume this is still too dark, though I found it perfectly playable. Then again, I've been playing this game for a while now, so perhaps I'm not the best judge of that.
Here are two other settings. Here's a fire dungeon:
And here's an evil dungeon. Everyone knows that violet is the most evil color.
@Superku, thanks also for the gradient feedback! I'm planning on trying to use simple skycubes (also mostly gradients, I suppose, but we'll see) now. I'll have to create a mockup or two to see your ambilight-idea in action, but I'll keep that in mind!
@Hummel, Yeah, I wanted to try ground fog anyhow. I'll post a shot when I get around to adding that!
@Joozey, Thanks! I am not using draw_obj at all right now, but I plan on using it for the bright spots that mark room exits, entrances and the golden treasure chest.
Looks much better now, but IMO the blue tint is a little too strong. The fire setting looks fun but the violet dungeon looks even better, I really like the colors except the pillars, they really need some texturing or even modelling work.
finally a nicer sprite lightmap creation is ready, replacing the random one, as Jcl suggested by creating temporary models to get exact image alpha, taking into account x2 x3 sprite property too. there are a couple of tricky settings e.g. accuracy of multiple hits (limits length of c_trace-chain of a terrain pixel), alpha treshold value, shadow darkness, 2 types of shadows generated atuomatically: a darker for non-alpha hits, and a lighter for terrain and for too many alpha hits (it can be modified by alpha accuracy). another feature completed for the new release.
@Error014: You don't need to have all level in same color or lightening for the fire dungeon. You could put some sprites moving here and here to simulate some moving red fog ? so we could see original colors of level with some red tones. Well it's good like that already (it's just an idea and perhaps not that good).
For the pillars , they look like 4 glowing pillars, because they seem to have some luminosity stronger than violet ambiant color.
The "project" I wanted to work on is cleaning up my hdrr-code to make it easier usable and more efficient... but somehow I ended up with completely re-creating the whole thing.
I just finished the bloom shader ( the most complex shader I've ever written ) and I'm pretty happy with the result:
I've still got problems with the blending in bright scenes but for this style it already works well
was this a serious question about what's the box supposed to be? ...I made an object shader with self illumination support for ratchet an I thought using his box is a great way to test my bloom shader
EDIT: the blending seems to be okay now ...some more screens:
EDIT#2: Just made an fps test 600fps - which is pretty fast! compared to my old bloom shader (which was less complex and looked worse) this is a huge improvement (~5x faster)
But now I've got the next problem: making a game and using the shader properly
sometimes I have the feeling that a lot of users here do not really deal with making games, just does some shader and graphics stuff, which is of course also very important, but real game development information is relatively rare on this game engine forum. (I'm developing my game within my editor )
while children have been sleeping I heroically overcame my lazyness, and made a main settings menu to set and store in a text file screen resolution (only valid is selectable, thanks Ayumi for the AUM link), window mode, engine version and a lot of other data. maybe a reset option would be fine. and a cancel. anyway, editor update nearly ready, what I want to add a lot of free graphics stuff to the package next week (e.g. vegetation based on nobiax's package). any contribution welcome
Thanks again for the feedback! You're right, the blue is a bit too strong. I'll try to reduce it while still keeping the night-feel of the scene.
@ratchet, some structure in the overlay is something I've already experimented with, but it didn't give too great results. I may come back to it sometime, but for the moment, I'm rather happy with how the "color correction" turned out
Yeah, those pillars... You guys don't like 'em, huh? I'm not sure whether it's because of the code (it may set it's unlit-flag?), or because of it's texture. I'll have to experiment. I've put that on my list, thanks!
In the meantime, here's King Red with a humble request
(DungeonDeities doesn't have a real tutorial, but if it detects the player not moving for a while (20 seconds), that speechbubble will appear, guding the player to the monarch, and explaining that they have to press space to interact with him)
... Yeah, I know the font goes past the outline of the thing.
@Error014 I really like the style and idea of your game did you already think about using anti aliasing?
I'm not saying that I don't like your retro style but it would be nicer to look at without the aliasing.
--- I spent the weekend with working on my HDRR-shaders and I'm finally happy with the result. I really like how it handles and displays the brightness even if it's out of range.
It turened out so much better than I expected it to be:
Click to reveal..
(click image to see in full resolution)
The brightness of the lights in this scene is far out of range (6x brighter than maximum)
Everything above maximum usually gets cut off but HDRR can handle this brightness and the bloom does a great job on visualising it:
and some sun reflections
...now that I have absolutely no idea what I could improve I'll finally start working on a game which uses these shaders Any ideas?
I'll finally start working on a game which uses these shaders Any ideas?
Here is my idea:
Click to reveal..
You could make a modern looking version of the old 'STAR WARS' games X-Wing or TIE-Fighter. With realistic looking cockpits like in Star Citizen. You can find many free 3D models of space ships in the internet.
Or you could make your own spacebattle game with YOUR OWN spaceship-models.
Your shaders are optimal for engine glows, or laser beams, planets und suns, windows of space ships, cockpit instruments, etc. etc.! And the coding of a spacebattle game is not as complicated as other games. You could use PhysX for moving your space ships and weapons (laser beams, torpedoes) and to simulate realstic collisions.
but as I said: first I'll make a simple fly-around-and-shoot-some-stuff - game. Since I started with 3DGS I've just completed some extremely small games (pong, tic tac toe)
I really want to create (and really finish) something bigger than pong now but still nothing huge.
first one: no idea second one: same as first, just used the AStar lib linked in the post with the image Third: Nope i don't want to do this (because most of the tiles have special passways and stuff, you need to configure special arcs between nodes to be active or inactive (because of doors or such things)
It's like Dungeon Dities, Error104 uses a PHP editor but the games is with 3DGS. I could also make tools under other languages to make some file containing game stuff to load under 3DGS. For a 3D tile level editor this is in what i thaught also, lot more easy to make in C# and with a vry simple to use 3D engine.
Was bored today and made a screensaver with some monster fishes (models not made by me of course). They swim around and attack/kill each other . For the ocean movement i took shade-c heathaze effect.
What i can say is that they have some great 3D textures and models, that's what makes great the game before shaders. The level design is good with diversified themes, but not outstanding.
imo they put huge-huge amount of professional 3d art into the game, concentrating on all details: models, animations, background, menus, cutscenes, sounds, talks, effects, textures etc. the levels are very different, so each of them requires long design time, not just reorganization of a similar environment to get a new one. it includes a couple of game play type, so probably it's fun to play. I really don't know why they don't make an english version...
Yes, you must be right about pro 3D assets, and perhaps they baught custom made models from freelance 3D artists also. For english version perhaps it's a matter or rights caus it uses some Marvel license in that country, fro other country the rights are not perhaps allowed or costs more ? Their game is not based on some animation movie also ?
beside posting stupid and non-stupid comments on this forum, I just found that I forgot to update my editor's terrain deformation due to A8.40.3 that finally auto-updates hull correctly for all the chunks, so I just modified it to show shading changes also in real time (vertex normals are not auto-updated after deformation):
simple brush based height increase, updating normals only in modified area for speed, and then updating existing entity positions manually (it means pressing a hotkey or a button):
adding some noise to a flat surface by selecting a proper brush, a proper area size, and a small enough height adjustment value, then only one click:
still learning to code and trying to make a simple racing game a lot to do but it's fun to create it step by step it will prolly turn out to be a little workshop
@Sivan : Yes, that's i think lot more better to work on individual parts ot have lot more speed. Does the terrain neightbours adjust their height smoothly with the chunck you modify ? You should have chunckc with more resolution perhaps ? we see a lot the polygons, just y taste and perhaps your game won't need that ?
@Realspawn : Their was a simple racing game without physic contribution. (But i don't remember where it was). Good work, looks pro.
every vertices are handled similarly fine, no chunk border problems, it was bad only in A7 (but I had a workaround for it using d3dcomputenormals).
what you see rough is because of vertex distance i.e. terrain resolution. if it is set to denser you can get a much smoother terrain resulting in performance drop (I use 4 tiles as vertex distance i.e. 128 quants which is okay for me). and I used here high raise values to see changes clearly. moreover there is a smoothing option to post process your rough terrain by a few clicks to get something really natural.
Edit: also the selected brush, and the manipulated area size have effect on the final smoothess of distorsion.
Yes that's different from indie engines havong some advanced terrain , big and smooth with good frame rate, and you would need the pro version that contains LOD on terrain.
@ratchet I have not found any car game example without the physics in c-lite. As I am a non coder i try to create all from scratch step by step. so all you see is the base foundation of a full game. The cars do fly sometimes due to the path thing when i use path spline the models get spinning around instead of staying at their path so i keep experimenting Next clip i will use shadows
but grass sprites are mainly okay for pro users. 80% of them are clipped at lod0 distance, and the remaingings at lod1. I decreased their resolution but still performance eaters in large quantity. this is why the grass model category exists (I will add new models soon to them too), they work faster.
shrubs sprites consist of a group of _x3 sprites, you can define the quantity from 1 to 5 resulting in max 15 sprites per shrub. they have a special lod clipping, the first sprite is clipped at lod2, the second at lod1, all the others at lod0 distance, thus they can be used effectively due to my experiences, but they are mainly for fast design, models can be nicer and more fine tuned.
You should make sprites more little no ? The character seems tiny ? And if they are too big, with immobile units passing throught them, it could look weird until you add physic movement of plants ?
I disagree, I think it looks more natural... I despise these games that ... well for some reason people think lawns got mowed in medival times or somethings???
please don't forget, it's just an editor screenshot, not a game. things are approx. in good scale, but there is a scale slider. maybe hard to see on this image the 3d distances... I decreased sprite resolutions today afternoon, and it has a sensible effect on speed.
shrubs are not passable, I mean the sprites itself are passable, but each plant gets a collider entity, by default a cube, but if you load it in your game you can change it by script easily to any shape.
in my pathfinder, what I plan to adapt to more game types, heavy units (like vehicles or tanks) are already set to be able to go through shrubs, while infantry and cavalry are unable. once I need to implement smoothed vehicle movement (I have the reference material for it that suits my system)...
sivan@ man, to be honest, I was amazed.. and if not that guard model on the middle of the screen, making his was throw the fields.. I could never believe that this is Acknex... Great work dude.
thanks, but they are just free stuff from nobiax.deviantart.com I spent some time with organizing and converting his works to my needs. simply better than the loopix collection, and totally free. I'm going to add some models with his nice textures, better trees than the current ones, and other plants, rocks, all having lods etc.
the guard model is a 3dgs standard reference object, I will use it forever if I have time I will rig it to add mocap skeleton animations to it, I think it would be a big mistake to make it nicer
"sivan@ man, to be honest, I was amazed.. and if not that guard model on the middle of the screen, making his was throw the fields.. I could never believe that this is Acknex... Great work dude."
after vomiting throughout last evening, today I added 7 new grass models consisting of 21 new grass textures. they can be used effectively below pro edition, but not in extreme quantities...
yes, using lod models is an easy way to manage proper clipping. I simply use 80% clipping at lod0, and the remaining at lod1 by default, but if a model having lods detected, in that case no clipping set, and the the lod models determine clipping.
here's a little demolevel with the water shader. The shader doesn't support skies, yet so the it renders incorrect if you look at the sky. But at least the depth and refraction renders correctly even if you're underwater
In the demo you can move the camera with WASD, Space, Ctrl, Shift and change the waterheight with the mouse wheel.
@Ch40zzC0d3r I'm sorry but I didn't finish it yet. It's a bit harder than I thought and I've got some problems making it work properly.
Thanks for doing demo! :-) It looks pretty good. Unfortunately it renders pretty slow on my machine. I guess my GeForce 8800 GT isn't really up-to-date. I noticed there is a pixel seam around polygons before the water. It's pretty well noticeable when you look down though the arcs on the water surface. Are you aware of this?
I guess you meant the refraction of objects above the water and the pixel noise it causes on its borders like on the left side of the pillar in this screenshot:
(click to see in full resolution)
It usually isn't possible to fully compensate this because the shader is just rendered in screenspace. ...but I was able to almost compensate it (right side of the image)
It looks good but is also very slow on my machine. I wonder if you are clipping pixels if they are not on the water plane and/or not visible? Because if I look against in the middle of the scene along the XY-plane, I get no fps increase (it stays the same).
To turn that into an object shader you would have to render the scene like usual in the camera-view but without the water-plane. And in a second, staged view you would render only the water plane and turn NOSKY on. By this the z-Buffer won'tz be cleared and the plane gets clipped against the z-Buffer of the camera view. With TargetMap you can then access the previous rendering result.
Thanks for the information HeelX. Yes I am clipping the not-affected pixels and yes the shader is slow but I'm pretty sure that this isn't caused by the watershader because it 'just' uses some clever math, but nothing too complex.
I think it's the complex & inefficient objectshader I'm using there and the textures because they don't have mipmaps.
EDIT: I just tested it, in the worst case the textures and the specular lighting reduce the framerate by 50% (~120 fps)
First version of my street creator. You can't edit the street curve but that will be the easier part i think I will support also rivers carved in terrains.
@Kartoffel: Just tested your pp water demo, surprisingly the water isn't visible at all on my machine/ 5870. I can see the rather nice material on the curtains and the two colored lights but no water.
caution! new editors are flooding us! (taking into account IceX3 too) congrat for this fast work, really cool, and if anybody lacks the gizmos in my editor can find it in yours
I just created 6 new trees (working on lods now). the pines probably needs finer vertex animation...
That looks cool! I am suprised how fast you are! I wait for the first demo, I wonder how stable and versatile your tool is maybe I didn't catched this.. are you going to sell it or are you going to give it to the community for free?
I want to sell it. I don't know how much it will be but i'll release a demo where you can try all stuff (except level loading, saving and export) I also thought of two versions: One with the basic features (The stuff i've implemented so far + export) And the full featured editor with more stuff
@Kartoffel: Just tested your pp water demo, surprisingly the water isn't visible at all on my machine/ 5870. I can see the rather nice material on the curtains and the two colored lights but no water.
the water level is just slightly above the ground, you have to increase the height with the mousewheel.
Ok got it now. Looks pretty good (from beneath the water surface, too), I like it. Seems to run smoothly but I would have liked to see default.c included.
ratchet, would you kindly use quote tags to quote. That would extremely improve the readability of your posts and also make clear whom you are quoting.
Ok got it now. Looks pretty good (from beneath the water surface, too), I like it. Seems to run smoothly but I would have liked to see default.c included.
@Uhrwerk : Perhaps it's better for people to not understand what i say or mean , i critic too much sometimes.
@MasterQ32 : Well , why not making a free version ? I don't know about 3DGS future ? will the new WED be as good as your editor ? Anyway great work.
**************
I have seen that on manual : Without a bones shader, bones animation needs more memory space and more rendering time than vertex animation. When you need huge armies of animated models on the screen, but have only the Free or Extra Edition, you might prefer vertex animation. Using a bones animation shader (Commercial Edition C or above), bones animation renders remarkably faster than vertex animation. I thaught bones with GPU was pro only ?
********************
I've just thaught of the mini game contest and how to import animations and animate in MED and well ... We would need some tool to name animations in an easy way : 1-20 : walk etc ... I'm loosed, i don't know how to do as i have a simple FBX character with all animations together ? It was some Blender script to make name for frames, but i don't find it , and i'm not sure it would be compatible with Blender 2.6 ?
We have used it from years (A6 from my part). Well, there have been GED, but it crashed too much, too often by just simple tasks like manipulating models.
Yes it becomes pressant, to have some lot better workflow and editor when the engine is not pure API, but a Game Studio.
after some texture, lod, and pssm optimization, I got at least 40 fps with 700 trees (7 different ones, 100 of each) in the level (one tree is one entity, no groups):
what I don't know why visible entities are more than total entities.......?
Thats maybe caused by a shader (just and idea, I've faced this with shaders before) or additional view? BTW, what fps you get without PSSM? Are you going to make groups out of trees, to speed the draw calls? All-in-all, your screens looks awesome man.
yes, right, it is a shader thingie (4 views of pssm), thanks. without pssm it is at 60 fps, but it cannot go over in fullscreen mode (1366x768). I cannot check it in windowed mode. but I'm cheating a bit, because the last split of pssm is not processed.
later I plan to add placment options for groups of obstacles, but that needs some data in advance, to be handled fine by my pathfinder, and to put correct collider entities...
Finally I put myself together and started a demo for my CEGUI plugin.
This is the unified CEGUI editor developed by the CEGUI creators. You can create the gui in it and load it as xml into acknex. Sadly the editor is very much of a beta and doesn't support filling containers like listboxes or comboboxes. And it crashes a lot.
Äh, to clarify this, the editor is not made by me. It is the work of the CEGUI team.
I just wrote the plugin dll, to use CEGUI with the acknex engine.
The picture was meant to demostrate, that it is very easy to create some small guis.
The plugin is now complete, but the problem is the documentation. I have no time for a whole documentation, so I am building a demo, to show the most important features.
I am working on an acknex intern editor, but it needs some time (complete Lite-C code generation without xml files).
With the xml system it is possible to realize some kind of model view controler system. Of course in a script language like Lite-C it is not realy possible to outsource the gui complete.
BTW:
If someone wants the actual plugin, send me a message.
Looks nice, indeed. The reflections of the flames don't feel like reflections of light sources, though. Some additive glow sprites for your torches could increase the illusion of light in general. And perhaps some sparks particles...
@Error014: This is one screen of your game, which is not as awesome to me as the other ones. Here are the reasons why I feel so:
1.) the reflections are a nice thing, but they break the design of your game. Your game is made of pixelated graphics, so that the visible texels are sharp and the outlines of the geometry are sharp. The reflection in return is pixelated, too, but in an unsharp and blurry way. Even the low-res sprites of the torches are made even more low-res in the reflection. I would massively increase the resolution of the render target so that the outlines of the geometry are at least sharp again and, the better, reflections of sharp texels are visible, too.
2.) This is an ice cave, isn't it? I'm sorry, but it doesn't feel so. The first reason for that is because the surrounding level background is black. If it where white or ice-blue, it would greatly enhance the frosty feeling. Plus, it is too blue. It is simply astonishing that so many games try to make something look icy just by adding a blue tint. Sure, an ice-blue color should be dominant but a reflected and unabsorbed light should illuminate the scene more and make it more white-ish. In your case, the blue color is really dark for an ice tone.
3.) The snow flakes are very soft, have you tried to use pixelated ones? One pixel snowflakes maybe?
4.) The same for the fire around the player: it certainly breaks the design when you use soft-fire flames for the fighting and then you have pixelated flames in the same scene.
5.) I don't know if you are using drop shadows, but the shadow of the floating "thing" at the lower left of the player has a drop shadow that is certainly too small
Your game is great but I think this screen reveals some design flaws, I'm sorry..
[EDIT] One solution to the reflections could be the "fake-way" by doubling geometry beneath the floor like in the good old days. Removes the necessarity to re-render the scene and you have perfect relfection in an instant.
All is a matter of taste : - I like a lot the blue colors - The reflections are great - The new power attack effects looks great
We should need to see a video to really appreciate it in motion !
I agree : -Background : some white/Blue grading color skybox would be better. -Color should be more white snow on top of blocks, full blue is lot more for Water temples and dungeons
It's a retro game so no need to stick prefectly to retro conventions like pixellated on anything, it's a game so no rules, and your own rules
He could use even bump maps on some elements if we wanted.
Glow Sprites: Yeah, I've considered that. We've recently added that glow to highlight the entrance and the transitions to other rooms (You can see it around the staircase at the top), and I've wondered whether or not to add that to bright objects, too. But I guess I'll stick with normal sprites (or maybe particles) for such a thing, but it's definitely on the list!
@HeelX:
Thank you for the detailed feedback! Much appreciated. No worries about offending me, criticism is always welcome!
Reflections: Since I'm not the one who designs the rooms (the players are), I can't really go with the "copy-the-geometry"-way. I've thought about doing that, but here's the thing: It's possible for objects or particles or other effects to fall through the geometry. If I simply copy everything and abuse PORTALCLIP, things would look weird with them, wouldn't they? I dunno. If I don't use PORTALCLIP, then dungeons will look weird if the creator doesn't place tiles around other ice-tiles. So that's why I used a reflection shader. It's hardly an ideal solution, though - the bigger problem I see is that for different tile heights, I *should* have different views, right? But I don't want the scene to render five times in the worst case. So that's an unsolved problem right there - I may still end up faking the reflections (which is, however, difficult - consider what to do if the player stands on the edge of a higher tile). Or removing reflections. Hmm...
The resolution, to me, isn't that big a problem - I could increase it, though personally I kind of like the look of it - it seems as if the reflection isn't "clear", which seems reasonable. If anyone else has an opinion on this, I'd be grateful to hear it
Background: Yup, there's no background right now. Sinve I've added Color Correction (more on that in a minute), I apparently cannot use view.bg anymore. There used to be a slight gradient in the background, and I still want to add that (and, later on, more different kind of backgrounds), but I haven't yet got around to turning that gradient into a skycube. The CODE is already there, though - it's just the graphic that's missing. The blue tint might be too strong, you're right - I've quickly adjusted the parameters before the shot. That's a lame excuse, I know - it sounds more defensive than I intended. I agree with you, and it's easy to fix is what I'm trying to say - just have to adjust fog and color correction settings. I better let someone else do that (who knows that stuff) anyhow!
Snow Flakes: Nope, haven't tried pixelated ones yet - I might try that! Though I'm not too much of a fan on simply adding pixelated graphics on modern effects. Still worth a shot.
Fire: I used to use pixelated flames, but I couldn't get that to look right - given the scale of the player on the screen, it'd either look as if the player was barely burning, or just be a confusing ugly mess. That's why I went with these flames (which look very nice in motion). My general rule is to make effects fancy, which makes them stand out a bit, and which in general I like (see also the lighting, which is also smooth instead of pixelated). So maybe I should adjust the torches' flame instead.
Shadows: Agh, yeah, they're scale is wrong. When they jump, the shadow is made smaller (that's how it was done back in the days! :D), so that's why it looks so super-small right now. The usual size is also too small, though - will investigate!
Gah, this all sounds way too defensive. It's not meant like that - I truly appreciate your feedback and comments! I just wanted to explain why some things are the way they are - not as an excuse, but rather an explanation!
Like any game, all people have their own tastes. Some will love a game, others will hate it or say it's to their style of game , style of graphics etc ...
I sould say, take advice from serious problem design people could have and see if it is really justified yourself.
And keep your vision of the game on effects, mirrors etc ... Don't be so defensive , do like some studios, put originaly and YOUR OWN VISION OF HTE GAME GRAPHICS
Could you have time to make a lil' video of this ice level ?
The reason why scaling dropshadows were used in older games like Mario 64 was because it adds a measure of how far away the player is from the ground. Since realtime shadows were not possible back then, it was a good trick. With a realtime reflection, you don't need that trick anymore, because the reflection is a more sophisticated cue to the player to measure distance between ground and the object itself. I think this might be tested through trial and error; I suggest to fade out the drop shadow on realtime-reflective surfaces.
Faintness in reflections are especially dependent by the distance of the reflected object to the surface, otherwise it looks dull. I made a quick mockup with a 50% reflection-surface. In the left you see a perfect reflection in the right a fainted one; please take a look at the cube near the plane and how the reflection gets blurred the farther the cube is away from the surface:
To achieve a fainted reflection in realtime games, you could use a high- or mid-res RT and do the blurring in a postprocessing step. If you write the depth of the render view of the mirror image into a seperate screen-space render target, you can use this as a cue for a dynamic blur shader, so that you get the depth-fainting of objects far away from the mirror surface.
[Troll mode on] @HeelX : I don't see any difference between the two pictures !
Again to annoy you on reflections : It is justified to spend so much time on some mirror effect that works already specially for a retro game And people won't stare at screen looking if the reflection are good or not, they will be in the fire of the action
It is justified to spend so much time on some mirror effect that works already specially for a retro game. And people won't stare at screen looking if the reflection are good or not, they will be in the fire of the action
I think that should be of Error's concern but in general, it depends. If you only have one or two rooms in your game where you have realtime floor reflections, you can go away with a cheap trick, but if your game relies on such a feature, you should invest time in it.
If they are done right, reflections can add a subtle degree of "polishness" to your game. It is certainly no great feature which promotes your game, but it is in general a pleasure for your eyes. That is why you wash your hair, add some perfume and put on some nice clothes when you go out on a date - even if you are handsome, this extra polish will make you more attractive to your girl's eyes and other senses. Subtle polish is always attractive - to gamers and the opposite sex as well
It's hardly an ideal solution, though - the bigger problem I see is that for different tile heights, I *should* have different views, right? But I don't want the scene to render five times in the worst case. So that's an unsolved problem right there
You could disable the reflections on older hardware and use screen space reflections on PS3.0 graphics cards to support an unlimited number of different tile heights respectively reflections on those.
@MasterQ32: Are you saying that you like the style of Dungeon Deities or that you like the style of the game ratchet posted? I think that the hidef VS lowres approach in ratchet's screenshot is very appealing. Best of both worlds, but in the end the beauty is in the eye of the beholder
This is where we can shine, little games with special graphic style and gameplay
Here is a video fo a super addictive game i played today, it's a rogueLike, simple 3D tiles levels, but it 's well done and the mix of strategy and Dices for gameplay is just excellent (specially powers to re roll, or influence the dices)
...just making some performance optimized example levels (using 5000 pcs of larger grass models, 500 pcs of bushes made of sprites, and 700 pcs of trees). fps starts to be okay for denser vegetation too, but there are some things to do...
looks very nice sivan! But are you using single models for every grass bush?
I'm just asking because I see that there are > 8000 entities and that are a lot of draw calls causing a lot of render overhead which has a great impact on performance.
If you combine multiple grass models to one you can improve performance a lot.
The grass polygons delimitation is too obvious on terrain. You should have grass less "packed" texture , more alpha and more holes in the textures, this look more like bushes not grass or it's very dense grass.
Under grass you should paint on terrain the same very green grass color and not put brighter green grass texture , caus it looks like the grass is just posed on top of terrain.
now one large grass area model is of 144 tris. 6 times larger than I used earlier based on a model from Loopix. it can be bigger on flat terrains, and lod could be used too. I will make some more trials and include it in my editor, so everybody can play with it. what I found not to be an effective thing is to use sprite based bushes (500 bushes means 1500 _x.dds sprite entities), I will combine them into single models with 2-3 lod models. in the sceen above I used 4 split pssm shadows, processing only the 2 closer splits. in a RPG/FPS game it results in a fine fps boost, while camera.clip_far can be 10.000 quants to have a far enough visibility range. moreover using UNLIT and OVERLAY and no SHADOW for grass and bushes is also a speed increase. only the emissive values of the unlit material should be adjusted to level lights...
I am remastering Crytek's Sponza scene right now in Cinema4D (not yet finished) for Gamestudio. I touch all objects and architecture parts, breaking normals, fixing polygon errors, making some new textures, adding missing normal/spec maps and so on.
This is currently a render shot, but I hope that I can switch to a realtime demo soon. Kartoffel is helping me with some shaders in the meantime.
@ratchet: I hope you know the purpose of this scene... And you CAN actually make a game out of everything. let's take Kartoffels water shader and build a racing game within the sponza lower level: water racing midlevel: airplane and car racing
you would've just one track but various vehicles...
The Atrium Sponza Palace, Dubrovnik, is an elegant and improved model created by Frank Meinl. Over the years, the Sponza Atrium scene has become one of the most popular 3D scenes for testing global illumination and radiosity due to it's specific architectural structure which is particularly complex for global illumination light. It is donated to the public for radiosity.
what sort of shaders ? Mainly an Objectshader, Bloom, and some Vignette/Noise/etc. stuff.
what bitmaps (norm, spec) ? Currently a Diffuse texture, a Normalmap and a Detailmap is used. Later it will be possible to give specular intensity and gloss information in the normalmap, too.
How many dynamic lights ? 8, limited by DX9. With custom light management theoretically unlimited. (as much as Shader Model 3.0 supports)
fixed directionnal light ? You mean the sun? -> yes.
Does it take account of WED sun and light ? Yes, but levelblocks are not supported at the moment.
...added some bush models (small ones to grass, bigger ones with lod to osbtacle category), and replaced our test hero with a new one. if I have time for making some example levels, I will publish the update this week.
@MasterQ32: Cool, but remove the ".zip" from the project name please How does the structure of the "level1.owl" file look like?
@Sivan: You seem to have found some nice ways to speed up the level rendering process (LOD, PSSM, ...). It would be great if you write a tutorial about how to design large outdoor levels properly!!! (We could ask George if he offers a "tutorial contest"!)
I think it doesn't matter how .owl is saved I will provide export functions to different formats (Direct Lite-C Load-Script, WMB, ...) Also the zip filename is just some prototype. i need to clarify what will be an AssetContainer and a Project (Lite-C or C# Project)
I've ever wanted to make nice documentation and tutorials for my editor and how to use some external tools with 3dgs, like L3DT, Treemagik, TreeIT, Tree[d], Milkshape, Fragmotion, and some game modding tools, but it's very time consuming as I experienced when I made the Terragen skybox tutorial, and the current documentation.
fortunately, as all the settings are in a few panels in MapBuilder, if I make some short descriptions with a couple of useful hints, you can see in realtime the performance and quality changes very soon. and playing with the assets I provide for free, also can be useful, to see how texture size and model lod poly count affects these 2 things.
do you know any nice building models with lods for free what I could include in my package? maybe I have some in my old pc, but always forget to seacrh them when I'm at home...
vegetation requires max. a few thousands particles, so you have enough for rain and snow, and for some other special effects. but it is only one option, alternatively, if the vegetation is not destructible, a lightmap can be generated, but it needs time and a large texture.
these are the 2 solutions for free/extra, you have pssm shadows in commercial or above, what can be also used with good performance if everything designed and set correctly (I "hacked" shadows.c a bit for better fps, what I'll include in the script package after a few weeks of my editor update, but it will require an engine fix to really work as I want)
Mabe controlling your particle's duration "manually" might be an option... By setting your particle "lifespan" dependant on "max_particles-x" to make some of the new particles dissapear instantly before getting dangerousñy close to the limit. That way you could keep the shadow particles safe by removing the oldest "effect particles" before reaching the limit.
yes, decal shadows are fast for non-moving entities, but I found it slow in case of lot of moving ones. I mean entity.shadow, if shadow_stencil is 1 or 2, their only drawback is the disappearing on terrain chunk borders. ent_decal-s have no such an error on chunk borders, and can be used beside pssm shadows, but they are much slower... I just made a sample level with 3 different shadow systems, and imo it can be not ugly sometimes to use decal shadows, 3dgs free can do this smoothly over 50fps (more exciting than simple pssm shadows but if examined carefully shadows are added and can be unnatural):
I'm not a graphics expert, this column is simply nice. I hope, beside quality, you can achieve a good performance too.
@ratchet: yes, game making requires a lot of fake solutions, and I absolutely agree, in some cases nobody will take care of some lower quality graphics. but optimization requires to discover a lot of possible solutions.
Character with armor...This mesh has a strength morph added to it to deform the body to make him look like a warrior, which can also be reversed to make him look weaker, as well aging morphs. Still working on deforming the armor to match as their is some clipping but other then that.
a group pathfinder demo will be in the editor, lod models have lod highlighting textures. in case of free/extra edition feature set it looks like this: when using decal shadows for environment, and stencil shadows for units:
and if using a static lightmap with unit stencil shadows:
That looks good! Are you going to publish your port?
(O.T. Ch40zzC0d3r, I remember when you started using Gamestudio and asked beginner questions. Now you are posting lib ports, great progress man! )
Thx xD If you wish I could release the library for free Also, GWEN is a GUI project by the creator of GarrysMod
Ahh, and sorry for the wrong awnser, the image is a gwen panel directly loaded froma file! But I think I could access the filebuffer directly to write a BMAP to the gwen buffer
But I think I could access the filebuffer directly to write a BMAP to the gwen buffer
That would be awesome! Because I would be interested into A) using BMAP* as bitmap-sources for images and B) abusing that to render a VIEW* into a GWEN image via the BMAP*.
Just imagine for example a small window with parameters for a procedural shader applied via bmap_process to an image or a 3D-file preview of a selected player character or something
But I think I could access the filebuffer directly to write a BMAP to the gwen buffer
That would be awesome! Because I would be interested into A) using BMAP* as bitmap-sources for images and B) abusing that to render a VIEW* into a GWEN image via the BMAP*.
Just imagine for example a small window with parameters for a procedural shader applied via bmap_process to an image or a 3D-file preview of a selected player character or something
Woah nice idea! WIll work on this later Reminds me on the good old days of quake engine and Enemy Territory clienthooks with nice killcam-windows
Thank you all for your feedback and the super interesting discussion on DD's artstyle & reflections. The lack of reply afterwards is not because I don't value that - in fact, I'm very, very thankful for all your comments (including the negative!) - but rather because I had to think about your suggestions. And I'm still not done thinking and deciding, but I guess I should reply. So, in no particular order...
Backgrounds - DD now features Skycubes. Woo! The game used a simple background image before, which I deactivated due to problems with the PostFX shaders. But they're back now, in form of a skycube! With the additional advantage that when the camera moves around now, the background now also slightly shifts - a neat effect! At this point, it's just a rather simple gradient, but nothing stops it in theory from using something else there - the plan is for the user to choose from a variety of options what background he or she wants for their dungeon. The problem in practice is the lack of time (and money) - there's so much to do! So overly detailed skycubes might not happen too soon. But that's something that can easily be added later. But for the moment, tinted skycubes. DD really uses the red/green/blue-parameters of entities a lot.
Reflections - I really enjoyed the discussion on this. Thanks to everyone who participated! Right now, there's two tiles that reflect the environment (to different degrees) - the water tiles (shallow water you can walk through at reduced speed) and ice tiles (which are slippery). So it's definitely not something the game depends on, though - well - if you look through the screenshots I've posted, you can see a rather large part of them features reflecting tiles, it's obvious that I like the reflections as well! In order to solve the problem with reflecting tiles at different heights, I think I'll go with Superku's solution and try screen space reflections, even though I'm not too happy with the PS3.0 requirement. I realize it's not a huge deal nowadays, but it still bothers me that DD with it's look requires PS3.0-hardware But I'm sure it's worth it.
Dungeon Defenders - Yeah, that's a pretty project, huh? I also like it's looks. HeelX comment here
Quote:
I think that the hidef VS lowres approach in ratchet's screenshot is very appealing.
is interesting to me - I'm trying to use a similar reasoning when I said:
Quote:
My general rule is to make effects fancy, which makes them stand out a bit, and which in general I like
But I'm not going to pretend DD's effects are as fancy as Dungeon Defender's (alright, so that could get potentially confusing with the abbreviation not being clear). I'm probably not going to try and compete with their pretty lighting - DD's super simple solution can result in some surprisingly good looking caverns, I found. I don't think investing more time in that would be a good idea - there's other stuff to do. Thanks still for sharing that!
Alright, I hope I didn't miss anything. Thanks once again for the comments and feedback! It's very much appreciated!
Oh, and before I forget - yeah, The Walking Dead videogame has an amazing story. I really, REALLY liked it. Some very depressing moments in there, though.
Error, what do you want us to say? The project is absolutely beautiful and god damn well executed and it just keeps getting better and better with every update, please keep up the great work and your attention to details.
Error, what do you want us to say? The project is absolutely beautiful and god damn well executed and it just keeps getting better and better with every update, please keep up the great work and your attention to details.
I actually think this should be moved to Showcase asap
Good start, some multitexture to bring diversity on desert ellow texture could be better.
I don't think it's about the textures. Also I can't fucking stand your belittling attitude anymore. That guy just single-handedly converted GWEN over to Acknex and all you have to say is "good start". ALL MY FUCKING RAGE
But I don't want to calm down! I wouldn't give a damn if it was only this post. Or only a few of his posts. But it's like this all the time. All. The. Time. There's a difference between constructive criticism, harsh constructive criticism, bad criticism you can probably still learn something from and belittling shitposting like this.
Also, maybe I'm curious if one can actually get banned around here.
I think the realtime-rendered scene won't look worse than this, except for the shadows and ambient occlusion. (not saying that the rendered screen looks bad in any way!)
I created a plugin for the full usage of the libnoise library. Why? because I was sick of that programs like worldbuilder or other procedural heightmap generators cost money, when there is a realy powerfull library out there that can do all the stuff.
I tried to do implement the sourcecode in a shader and do the calculations there to speed up the rendering process of the bitmap, but since d3d9 doesn't support bitshiftig and other bitwise operations, it is a bit tricky. But I made some progress... The problem is, that my knowledge of math is, hm, not deep
Mythran@ PM me, I can help (actually as I see, you've been already told, how to fix this problem, whats the problem then? ask for an explanation, if you need one!)
Coding some vegetation tools for IceX3 editor. At the screen you can see some simple volumetric fog (work in progress) and faked model instancing. The vegetation models are merged together via script. I'm still testing what's faster, this faked model instancing or the engines built in sprite instancing.
This means you place thousands of vegetation models and they will be merged together to a few one.
A example: if I place 4000 grass models into a level (60k polygons), the framerate drops down to 15 fps on my pc, because of the huge amount of draw call. A function creates a single model out of the 4000 entities. So the engine have to do only one draw call. After that the framerate is over 500 fps. If all grass models in the level exceeds a poly count of 65536 polys, they have to merged into several models.
[EDIT]: It seems sprite instancing is A LOT slower than the model merging method (if there are many many entities with very low poly count, like grass, little stones or low poly trees).
A example: if I place 4000 grass models into a level (60k polygons), the framerate drops down to 15 fps on my pc, because of the huge amount of draw call. A function creates a single model out of the 4000 entities. So the engine have to do only one draw call. After that the framerate is over 500 fps. If all grass models in the level exceeds a poly count of 65536 polys, they have to merged into several models. You know you have a GEM here ?
Your system could be used to group together static models of a level (houses, static structures , static 3D assets etc ...) I can't imagine the big boost in the games we make if we would have such function to deecrease the draw calls and entities
Your system could be used to group together static models of a level (houses, static structures , static 3D assets etc ...) I can't imagine the big boost in the games we make if we would have such function to deecrease the draw calls and entities
Your system could be used to group together static models of a level (houses, static structures , static 3D assets etc ...) I can't imagine the big boost in the games we make if we would have such function to deecrease the draw calls and entities
What's a GEM? As I wrote the framerate raised from 15 fps (4000 small entities) to 500 fps (1 big entity). And you will only get a big boost if you have a lot of low poly model (like grass). For "usual" models this method is useless, there you have to use LOD models.
Btw: to use this method for vegetation is no secret and not new. I just didn't see any useful tool or script on the forums which makes this easy to use/realize.
@oliver2s: a few users did model merging, but keep it for themselves or want to sell for money. I also experienced in the trial version that sprite instancing is slower than needed. especiallly when you use shader for wind animation. probably you have to create an own wind animation shader for the merged vegetation models... last but not least, nice work!
I try to teach my stupid folks how to identify and move on building floors in my hierarchical pathfinder. in simple cases it works fine and fast, but it will require to add some more data for complex buildings (e.g. by marking stairs and ladders):
Nothing eye catching or awesome, but it makes working with TEXT a bit easier:
COLOURFUL TEXT! This is how you create it:
Code:
CLTEXT* mytext = cltext_create(_str("[clff0000]Guard: [clffffff]I used to be an [cl2233ff]adventurer [clffffff]like you. Then I took an [cl22ff33]arrow [clffffff]in the knee."), font_create("Arial#18b"), 20, 320, 0);
cltext_show(mytext, 80, 360, 1);
Not finished for now, I miss automatic word wrapping.
@oliver2s Will you share your merging-code evantually?
Sure it's not THAT hard to do ( - for people who have some experience with the direct use of vertex- and indexbuffer, etc. - ) but if someone has already written a script like this ( and if it's working ) it would be really nice to have it. And It would save some time for those who need it.
thanks for the info, I'll try also this merging thingie, seems to be very cool. but it apparently modifieas vertex normals, which are very important in case of vegetation. another solution for grass placement is dynamic mesh creation, where you can make simple sprite-like meshes and make from them an entity like this
Coding some vegetation tools for IceX3 editor. At the screen you can see some simple volumetric fog (work in progress) and faked model instancing. The vegetation models are merged together via script. I'm still testing what's faster, this faked model instancing or the engines built in sprite instancing.
Not yet. At the moment it's just a few sprites that fade in/out on distance. But this was just for testing how it looks. For speed reasons this will be done by a shader later.
some small progress on the post-apocaliptic project:
The idea is that you will have to find food to eat and water to drink, otherwise after some time you will die. Also you have to sleep or you will be out of energy and unable to move fast. I think all of this will give another dimension to the game: you have to "survive" as you can. Of course, there will also be a day/night system
some small progress on the post-apocaliptic project: The idea is that you will have to find food to eat and water to drink, otherwise after some time you will die. Also you have to sleep or you will be out of energy and unable to move fast. I think all of this will give another dimension to the game: you have to "survive" as you can. Of course, there will also be a day/night system
Will it be like "survive as long as you can" or with a definitive goal?
How to copy something you don't know? Also, what is so evil about copying good ideas?
Ideas belong to everyone, don't claim anything. They aren't your ideas anyway. They are from someone else. They float around in the ether. The only thing left to do is to be in the right state of mind to be able to capture them.
Will it be like "survive as long as you can" or with a definitive goal?
the idea is having a main quest and secondary missions to complete, but you won't be able to do that if you can't survive
Originally Posted By: HeelX
Originally Posted By: Hummel
Quote:
Google it to make sure you dont copy game ideas
How to copy something you don't know? Also, what is so evil about copying good ideas?
Ideas belong to everyone, don't claim anything. They aren't your ideas anyway. They are from someone else. They float around in the ether. The only thing left to do is to be in the right state of mind to be able to capture them.
@Dico: It looks nice, especially the movement of the car, but sadly I have to say you destroy (!) every project of yours visually by an insane amount of bloom or blur, particularly your older projects.
I have to agree with Superku, the bloom overbrightens the image a lot. It's caused by the blending mode of the bloom (additive-blending). When using bloom you have to adjust it carefully to not mess up the final image.
...and if you have some experience with shaders I suggest you to stay as far away from additive-blending as possible because there's nothing which creates more overexposure than this blending technique due to the limited color-range which can be displayed.
Very smooth, would need buildings and trees on side roads to make a real racing game.
Originally Posted By: Superku
@Dico: It looks nice, especially the movement of the car, but sadly I have to say you destroy (!) every project of yours visually by an insane amount of bloom or blur, particularly your older projects.
Originally Posted By: Kartoffel
I have to agree with Superku, the bloom overbrightens the image a lot. It's caused by the blending mode of the bloom (additive-blending). When using bloom you have to adjust it carefully to not mess up the final image.
...and if you have some experience with shaders I suggest you to stay as far away from additive-blending as possible because there's nothing which creates more overexposure than this blending technique due to the limited color-range which can be displayed.
I wasn't talking about the gameplay, I was talking about the grapics. (the bloom shader)
What I and Superku meant was that the bloom creates too much overexposure. As soon as something gets brighter than the maximum it results in a clipped and bad-looking color.
Here's a little example: Left = overexposure, colors aren't accurate anymore. Right = colors are limited smoothly. ( I hope the effect is visible good enough )
I wasn't talking about the gameplay, I was talking about the grapics. (the bloom shader)
What I and Superku meant was that the bloom creates too much overexposure. As soon as something gets brighter than the maximum it results in a clipped and bad-looking color.
Here's a little example: Left = overexposure, colors aren't accurate anymore. Right = colors are limited smoothly. ( I hope the effect is visible good enough )
yes i understand that , i mean by gameplay more clear , that i mean shaders on screen , i will to make it more clearly
While I had huge fps problems when using more than 100 lights and also incorrect lighting with my old workaround-method I'm now able to use a couple of lights more:
It's a whole shading pipeline and no it doesn't work with official shaders unless you have enough experience to implement it, but then I suggest you to write your own, better shaders instead of using the "official" ones. Besides that it doesn't work like "normal" lighting calculations and there are no "near lights" because that's a part of the light-management system of direct-x 9 which I'm not using here because it's extremely old and you simply don't need a 8 lights per entity limit for today's games and hardware. What I'm using is much more flexible. There is no lights per entity limit and there's also no global light limit.
it's not only for more lights. you don't need to rewrite the whole lighting system if you just want some new material with emissive lighting in addition (in a "normal" system you would need to copy all your light code to the new shader) also your calculation complexity goes from O(n,l) = n * l to O(n,l) = n + l where n is the object count and l the light count
Have you ever tried to put some lightsources (like campfires) on a terrain? The light limiting is really annoying there. Also, deferred rendering makes it much easier to use shadow mapping techniques.
The main advantage of deferred rendering is that it's really flexible, fast and shades all objects with the same techniques / shaders, which results in better performance and a balanced looking scene. Another benefit is that deferred rendering is geometry independent.
...and of course it's possible to use any post-processing effects with this. I already implemented my HDRR pipeline and added screen space reflections.
I'm Not sure if I'll release it. I implemented shadowmapping for 1 light but it still is nothing very special. Also, if you want to use this ligting techinque I suggest you to use shade-c which has more features and is more efficient.
actual game project: The level is 196608x196608 Quants large, done some tricks with rendering. The white cube is de default CUBE_MDL with 16x16x16 Quants size, player is 64 Quant high
if I was more advanced, I would try to use multitexturing only on close distances, and show a blended one on far, and maybe make some tricky sky cubes imitating far terrain. but extreme high view range alone is not enough, and causes other problems, because the landscape should look alive and densely populated.
Added modified default water shader, now going to build some "level editor" for placing entities into the level and creating their sprite pendant for really far view
okay, got the first part of my LOD system finished the reflected texture of the terrain needs to be edited but everything else works fine. After a certain distance the trees will only be sprites so the render time won't be too high
Since the var range is -99999.999 to +99999.999, a total level size of ~200000 quants can be reached. The only problem raises the camera->far_clip range, because it can be only max=99999. Anyways, why do you need such a large terrain and what meter scale do you use?
I only use a clip_far of 15000 for the main view. there is a large range view to render this "huge" distance. i can get max. viewrange of over 400 000 Quant (in theory) Atm i can't publish the project as is but i will make a techdemo of this tomorrow so you can fly or walk around in the terrain. My meter scale is the "default" meter scale with 1.8m = 64 quant (so i don't have to adjust my current player and network synchro code)
In fact the var range is -999999.999 to +999999.999, i.e. 10 times as high.
No offense but in general I would like to see less frequent posts here, one post per day per user is more than enough I think and adding a default water shader is not so huge.
If people want to post more about their project I encourage them to open a thread in this projects forum, so it feels more alive and this thread stays a little organized. Just my personal opinion, of course.
...after some headache caused by debugging, the n-level pathfinder works. building floors can be allowed for large units (bridge, gate) or restricted for infantry (stairs, walltops), and also deals with tile height differences too. everything destructible, new buildings can be placed runtime, and database is auto updated.
They just expressed they're opinion. In fact i believe that each project should have it's own topic.. I dunno why this topic exists... It is somehow a projects section inside other projects section.. Why make a fight over each other opinions?
In fact i believe that each project should have it's own topic..
If everyone would do this it would be extremely chaotic.
This Thread exists because you can post some stuff you're currently working on (even if it's some little progress) and you'll get feedback from the community.
Also a lot of the projects here are just small and simply don't need a whole thread.
Then whats the point in having a project Folder? Projects are projects no matter the size or importance.. Chaotic how? I dont get it... Every project would be inside its own thead..
Quote:
Also a lot of the projects here are just small and simply don't need a whole thread.[quote]
Are you saying there should be another thread called "Shitty Stuff"? Just kidding...
I really don't care if this thread exixts or not.. But at least it should have some rules implemented.. Or so i think. But it has not, so...
Well that update with multi-layered pathfinding is really cool and the visualization is neat, too, I'm perfectly fine with such posts. Only sometimes this thread is flooded with updates/ posts of only minuscule changes or additions that don't do justice to the remaining posts. Posts of let's say a simple textured room with like 2 cubes (i.e. 10minutes of work) that is supposed to be a cool *insert genre here* game one day are pretty much "useless" to use Ratchet's words. As a hobby dev you simply cannot generate enough content on one day to justify multiple what are you working on posts without a severe lack of quality in both the post and the work itself.
Again, this is just my opinion and I'm only one regular forum user of many but I think it's hard to argue with that.
In fact i believe that each project should have it's own topic..
If everyone would do this it would be extremely chaotic.
If used to be like that for years, and there was nothing chaotic about that! As Superku said, this could make thread a little bit more ALIVE, if we would USE the forum as it should be USED. I have nothing against sivan's posts or anyone else stuff, but let's keep ourselves a little bit modest And I have to say, that I really love this thread, cause each time I have bad thoughts about my projects and time I've spent on making anything, I go to this thread, to see, that I'm not alone in the boat! And that really makes me feel a bit better So, I still support Superku! If your project is going really fast, and you can't keep on posting updates, make a UNIQUE thread for your project, that's why "Projects" thread was created for! Anyway, same as was said above, this is just my opinion, my two cents I'd say. So keep up the work!
Not sure how I feel about it. on one hand it is nice to have one thread with quick updates in it. on the other hand... the "hey! look I added a sphere to the project with 10 updates on it" posts are kinda annoying. then again, I dont see them on here really that much.
a storm on the potty fortunately the conflicts on this forum are so rare and really light, like this which is imo a great thing. I just re-read the first post, because I have not known who started it and for what purpose. any tips? the decipher:
Originally Posted By: ratchet
Post any screenshot,picture of anything that you are working on with A8 , it can be anything like a model, a 2D level design, a texture, a game etc ...
(It's a standard thread in lot of 3D artist and some 3D engine forums, i just wanted 3DGS to have one, it can be interesting)
...after some headache caused by debugging, the n-level pathfinder works. building floors can be allowed for large units (bridge, gate) or restricted for infantry (stairs, walltops), and also deals with tile height differences too. everything destructible, new buildings can be placed runtime, and database is auto updated.
Wow , very nice project , it look like age of empir 3
@Dico : thanks, I have not really played with AOE3, only with AOE2 and AOM, but yes, once it's going to be an RTS, but a bit different... the buildings are only fast made placeholders (wmb map entities), what I plan to share for newbies in future as an easily tileable castle builder kit, after some improvements. @Mythran : I've never posted water shaders here.
I think this message was for me And i have to show something later (isn't in a good shape actuall, also with default water shader this time, but with a lot of other stuff)
Implemented variance shadow mapping for up to 16 spotlights in my deferred renderer. Maybe I'll increase the number of shadowmapped lights to 32 but rendering the shadow-views is already a bit slow.
got my water working with background view also... the solution is quite dirty because i need an additional shader and 4 additional render targets to get it working
Improved the rendering part with slins help (Now uses only the main camera, but with a clip_far value of 400000, also added distant fog).
Imroved my ingame level editor with a commandline which allows faster spawning, moving and scaling models. Also added vehicle support to the multiplayer game base
Also you can see a kind of performance measurement in the lower part of the screens. Red one is entity render time, blue one is function update time. Good to see update time peaks or periodic actions
Also you can see a kind of performance measurement in the lower part of the screens. Red one is entity render time, blue one is function update time. Good to see update time peaks or periodic actions
great idea, especially if saved into a file with some other relevant data, I will definitely implement something similar for myself...
for the far view rendering you could somehow bake the very far entities like trees into the terrain texture. I don't know how, but I have this idea for a while maybe using it together with a heightmap. it could be useful for landscapes like this
Integrated Shade-C EVO into IceX3 Leveleditor. But for the upcoming version it will be disabled, because first I have to make the IcX3 terrain shader compatible the Shade-C EVO renderer.
BoH_Havoc did really an amazing job with this shader framework.
@MasterQ, I also created an Awesomium wrapper couple of years back, but at the time i concluded that(dont really remember why, there was couple of different things) rather than using awesomeium it could be a better idea to use the Chromium Embedded Framework(used in many mainstream apps, like steam). I do not know how far both are advanced now but anyway you could also check it out. cef:http://code.google.com/p/chromiumembedded/
@Quad: I think the main reason was the licence (Rei_Ayanami asked me some hours ago^^) I am trying to get the interface working with the Lite-C class support (and i just got an idea which could enable us this wonderful feature) And thanks for chromium embedded, i will take a look at this also
EDIT: okay, didn't work... But i got transparency working
Currently I'm working on a flashlight code for our Meteor Mess project.
It's still in progress, but works quite fine so far. At the moment the flashlight is just attached to a left hand's vertex without adding a special "hold flashlight" animation. This is something we have to add in the next steps.
Nice! Have you tried to attach Lite-C function pointers to click/touch events and so on? Have you tried to overlay the HTML rendering over a camera view?
The text is drawn by gamestudio Just some debug information.... I don't think dds is supported (Chrome cannot open it so i think Awesomium won't do it either) and smooth alpha fading should work also...
Added the first version of a street system: Streets need a lot more segments (atm this street has about 16 or 20 polys... Good thing is that the designer will be able to build goodlooking streets in realtime and then save them into a file. This allows a good editing performance and later a good game performance
yesterday I corrected the shade-c 0.90 water shader for 16:9 screens. worthy to add to MapBuilder (I will make a water menu for easy tweaking) with default pssm and hdr:
nothing, it is a single skin test terrain, but I normally use a modified terraintex.fx ffe that works with mirror waters too. I use it because it is simple to use with L3DT, and works for any texture quantity, but will replace with a shader, when I find a good solution, maybe the shade-c shader, maybe Heelx's contribution, maybe a custom one (if somebody can make a good one and wants to share it, I'm open to it ). I've never used the terraintex3 shader because it is very limited (and I think currently excluded from 3dgs installation).
Interesting, is this a 2D or 3D simulation? How do you calculate the interactions, directly (but apparently parallelized) which means N² (or N²/2) calculations or with a more advanced method (N*logN or even -> O(N))?
@sivan: Looks very very nice but you should think about removing the white foam/ the shoreline, game developers like to do this but there would not be any such thing in your scene if it was real life and it IMO always (!) looks horrible.
@sivan: Looks very very nice but you should think about removing the white foam/ the shoreline, game developers like to do this but there would not be any such thing in your scene if it was real life and it IMO always (!) looks horrible.
Well if this was a reallife-scene and if there were waves there would be a shoreline but when using something like this in a game both the waves and the foam have to look really good.
Kartoffel@ in real life, lake doesn't have any kind of a white foam shorelines (as far as I can see from the pictures, it's a lake), if it was an ocean, then yes.
I just made the basic things to work with shade-c default settings mostly. there are several parameters in the water shader what can be set, including several foam variables, so it appears over (or under) a certain depth value i.e. at low water around shore (or you can use a depth texture instead of the dynamic one). my plan to make them set in real time to see what they do exactly...
if you turn off Dof, no foam appears (only if you use static depth texture), but in that case water depth looks not so nice. what I want is to eliminate Dof because I don't like it, imo distance fog is much better.
later I will deal with shade-c 0.91b too, since its water is the same (nearly). unfortunately the water shader is not lightning fast, but acceptable (but reflection, refraction, and dynamic depth can be switched on separately to check their performance cost). unfortunately in latest shade-c evo neither water nor terrain shader included.
today morning I got the idea how to implement pssm fade out, after lunch I have had some time to do it. currently the last split is fading out, but using 2 distance parameters is also possible. works nice with my pssm split processing limiter, which results in a significant preformance boost, and now not so ugly because of the hard clip, and much better with distance fog.
Basically ray-traced low-res depth fields with geometric antialiasing. I'll provide in-depth information as soon as I open an extra thread. The underlying rasterized geometry is a box, but it also works with a quad.
Atm: red - depth, green - offset id, blue - normal id The offset can be used to give a vertical face the normal of the adjacent texel or to prevent it from using the modified normal of the own texel if the adjacent texel doesn't use one. There are no real sloped faces, they didn't pay of performance-wise. The hardest part is actually the AA. It's also by far the biggest part of the shader.
The terrain looks really nice, I like it! The differences in texture scale are cool, too. The knight probably could use a more fancy material but the screen shot is only about the terrain, right?
thanks! @3Run: The tent is not finished yet. I'm going to change the straps but probably will be triangles. @Superku: Yes. It is about the art of the terrain, but all the elements are for same project so they should fit together. I have some doubts.
I think those additional thin and long triangles should be ropes, too. For the knight I would remove some detail from his texture, maybe even reduce it to simple plain colors and then define his look with the help of an object shader (for example like a cross of your current look and Super Mario Galaxy).
I will probably redraw the knight texture. The problem with a rope is that it is just a small line in the screen renderer. It needs almost 2 pixel thickness with a great antialiasing to be visible in a aceptable way and it looks more a mooring rope than a strape rope. I felt those fabric strapes as an elegant solution to ensure that the strapes are rendered in any distance and I can't imagine what problem you find on them.
I felt those fabric strapes as an elegant solution to ensure that the strapes are rendered in any distance and I can't imagine what problem you find on them.
txesmi@ well.. that is just your taste (same as I do have mine, or Superku has his own one as well). I don't get too much into details (as you do, with antialiasing stuff, I only took into account the visual look), but for me, tents, which is setup like that (with stripes only?) looks a bit weird (and not elegant at all), but as I've said, just my taste and my two cents
@3Run: Sorry if it looks that I was fooling your taste, totally out of my intention. I was just trying to explain the main reason of those triangle strapes. Actually i like both, but triangles wins.
rvl_exile, have ever heard about modular texturing? i think you could texture that building with just 1 or 2 textures. a nice example : http://stefan-morrell.com/oe2/oe2wip12.jpg
yes, heard about this, i use this technique for every single objects (doors / windows / pipes / whatelse) But for Walls i dont use it, maybe i could reduce the texture count a bit, but not really much more.
EDIT: im using 5x 512² Textures for the walls / 1x 1024² Map for all objects and 1x 512² for the decals
I'm working on a pipeline i thought it could work fine. It consists on goraud lighting with diffuse based specularity, shadow casting isometric projection and ambient occlusion. Everything deferred and computed in view space.
@HeelX: I like the rendering (/ materials) in your second screenshot very much, looks really nice (except for the bright sky)!
Thanks, but I pass the compliment mainly towards Kartoffel, who did a great job on the lighting equation of the shader.
For the sky, I implemented a small tool to import skyboxes from other free sources of the internet:
I already found a nice dark sky and I am writing a pixelshader atm to change the skybox orientation (changing the pan of the sky entity has no effect for me?).
It is a simple (and slow) 4x8 depthmap sample comparision. It adds or sustracts from a starting 0.5 value without taking account of depth difference. Since the camera angle mantaines the polygon depth into a small range, it works good and looks plausible. I have to improve the algorithm in order to reduce the depthmap samples, something like the pixel coord based vector election of your SSAO shader
edited. I move the samples center in sun light direction for each step too.
No I'm not (but I'll probably switch to this technique).
At first I used a 1x256 texture which stored the modified color for the 3 channels: Color.r = tex2D(..., float2(Color.r, 0.5)); Color.g = tex2D(..., float2(Color.g, 0.5)); Color.b = tex2D(..., float2(Color.b, 0.5));
...but this wasn't good enough because when doing complex color modifications in gimp (which I want to look the same ingame) output.red, for example, isn't only affected by input.red but also by the other channels. (I hope it's clear what I'm trying to say...)
Now I'm using a 2D texture (256²) like that: Color.r = tex2D(..., float2(Color.r, <average color>)); (for every channel) So now the output color is also affected by the average brightness which gives a great variety. It's still not as flexible as a 3D-Lut but easier to create and works pretty good.
Edit: I just switched to 3D-LUT because now I know how to save 3D textures and I've got a very easy way to create them while editing the screenshot to how I want the scene to be.
As first project this looks very good! But the teleportation when touching some of the checkpoints seems irritating if you don't know what exactly it does.
As first project this looks very good! But the teleportation when touching some of the checkpoints seems irritating if you don't know what exactly it does.
well its a powreup that you can use by clicking the mouse button so it shouldnt be a problem
making a little demo (too lazy for a tutorial) of customizing default 3dgs pssm shadows. manipulating splits and their distances for performance optimization, added shadow fade-out feature, optional per-entity bias and smoothing (makes initialization very slow), handling bad shadows on high slope terrain etc. don't remember of all what I put together, I need some code cleaning - hopefully it will be easy to understand. if I have time I'll add some sliders for real-time adjustments, but probably only code commenting will be the only guide (it will be definitely available in my editor, the current split limiting is performance effective but a bit ugly method).
and you get a water shader cut out of shade-c 0.90 (only water specific code is present, thus a bit faster).
as I mentioned at another thread, code optimization lead to a huge performance improvement of my pathfinder. I simply tried out what you suggested, and it really works using one update loop for all the units resulted in processing time drop from 35ms to 17ms, and more than 10fps increase in case of stencil shadows, by eliminating separated actions containing 600 wait(1); in total (beside some other fine tuning and extended functionalities)
with stencil shadows for units and buildings, and static lightmap for vegetation:
with pssm shadows for units and buildings, and static lightmap for vegetation (much lower fps):
@sivan: Nice, however I think the function processing time is still unnecessarily high, what are you calculating in that scene at the moment of taking the screenshots?
@txesmi: I love it, and it will look even better when the rest of the ground gets a subtle detail texture, too. I'm still not a fan of the knight's material though, have you tried something in the direction of the Super Mario Galaxy look (/ shading) so it fits better to the grass (and does not look as antiquated)?
@Superku: yes significant time, I'm sure there are further possibilities, and I need time for new features too. after a short question, a long answer is coming , currently it does:
4 clearance map updates (spread over several frames), one for each capability. it is used by 1x1 tile sized units (infantry), larger units create an own small map (shortly: a light unit can access different tiles than heavy units or ships. these 4 maps show for each tile the max size of a unit can access them of the given capability). it could be eliminated sometimes, but in case of high number of 1x1 units it is currently faster.
20 long group A* pathfinding among areas taking into account area properties (map is about 500x500 tiles, an area is normally 8x8 tiles). its update frequency is high, thus very dynamic, but should be decreased a bit, but it has very small effect on performance. it can now handle a few building floors too, and senses bottlenecks which results in different group and unit behaviour. it will be different for AI and player. a basic path smoothing is also applied that is coming from the mutual effect of group and unit level targets.
the most power eating is the 600 unit movement target calculations and pathfindings, but only a few runs per frame (simple but fast depth first) using a common map or an own map, checking only a small area. group leaders has the longest search loop, the 1st line of units also has a bit longer search limit, the others are followers making very short searches, only a few tiles. units are keeping the formation taking into account actual group facing direction, and they "know" if they are far from where they should be. the formation calculations are not flexible nor fast at all at the moment, it should be rewritten totally.
units use c_move (I found physx slower but will test later again), which enables collision detection on path errors, which can be high in case of several crossing groups. they try to avoid each other in 2 ways, to avoid continuous colliding. far units can cheat. shadows are switched off for non-close units, aimation is also disabled for very far units.
what I found a bit too high, is the negative effect of shadows on performance. so probably I'll test shade-c 0.91 shadowmapping next as it produces double fps for my test level compared to default pssm. currently only the static lightmap for static, stencil shadows for dynamic objects configuration is fast enough for adding further functionalities, like combat behaviour.
Sadly in it's current form the geometric AA breaks badly at greater distances in certain situations for outer edges. Also the bright inner edges at the right site of the block are due to AA which is currently done for color independently from the shading.
EDIT: For the case it wasn't clear, the actual underlying geometry consists of six four-sided pyramids only. The rest is shader magic.
I do not modify the actual geometry. Imagine a pom shader which does not miss/step over texels. With other words, I send a ray through a regular grid. But as I wrote in an earlier post, the by far biggest part of the shader is related to the geometric antialiasing, which is still far from perfect. I need really good AA due to the large amount of details pixel art produces.
@Oliver2s: Amazing work. What kind of vehicle is this?
It's a harvester, a forestry vehicle for cutting down trees.
Originally Posted By: sivan
maybe agricultural simulations?
Actually for my current project (some kind of rail track vehicle simulation), but you can create any kind of usual vehicles within a few minutes (if you have the 3d models).
I just took the planetshader from the wiki and edited parts of it. It now has better (and more realistic) lighting, better fog and it uses fake atmospheric scattering.
If someone is interested I can upload it, but I have to make it more userfriendly first.
Thank you very much. I guess three weeks (3 hours +- a day). The animations of the player model limited my script but iam pretty happy how fast it is, and how the attacking of more models at the same time works pretty well too.
I've already got a new system which uses a 3D-Lut to store all color-modifications (color-grading, tonemapping, etc.) but I think that's not very userfriendly.
HeelX, what are you trying to do there with the depth? There has to be some way (maybe vendor specific?) to render the depth into a special depth texture and use it in a shader, as that is what everyone uses for shadow mapping. In newer graphics APIs you can even use such a depth texture as a depth buffer in another rendering pass. Also, at least somewhen in the past, gamestudio sorted its objects from front to back, to reduce overdraw to a minimum, so even in a worst case scenario, a z prepass might not be as effective as one would expect. The easiest way for a depth prepass would be, to write a shader just transforming the vertices and the fragment shader setting a single color (with clip or alphatesting, for alphatested textures), the render stage would be set in a way that no color are writte, for example in your shaders technique definition, you could set some flags. And then before rendering again, you just don´t clear the depth buffer.
I've added a menu of environment setting to my level editor. You can even add music files to play them in the final level without any line of script (it's all done by the icx3 script framework).
Hope I can release it soon. I just waiting for the next engine update and hoping that one bug (PSSM in combination with shaders causes render errors) will be fixed then.
I wanted to clip pixels for which the depth is below the previously sampled depth to avoid complex light calculations. But I wasn't able to measure a significant performance boost. I guess this is due to the fact that I sort most of my geometry into the solid pass already, I guess.
just added the custom version of pssm shadows to my editor, onwater shadows are also fine now (mirrorwater and shade-c with or without depth), here is the actual menu (Apply button at wrong place) - maybe the blur settings should be moved here too?:
I'm waiting for the new 3dgs update before making an update patch. any news on it? I saw only a little increase in the number of posts on the closed beta thread (but my reported errors seems to be fixed due to the bugs page)...
thanks. I reset some water shader settings, so shore foam is now really smooth, nearly invisible. any news is news. earlier I thought Jcl postponed the April update because he had known that my birthday is around mid May, and wanted to give me it as a present, so now I'm a bit disappointed. ( )
looks very moody. I like it a lot. I remember the old shots like the one on the left.
The new pictures looks so much better I think. Reminds me of recent "Alone in the Dark" game.
Especialy the camera angle is way better now... close to the action. I remember you also had a video. Would love to see a video of the reboot if possible
I'm in the middle of my pathfinder code's restructuring, the rewritten A* core (based on some ideas from this article of Marco Pinter) resulted in saving another few millisecs in case of 1000 units on a 500x500 tile area without obstacles. in a forest like scene it is slightly slower. a peak situation in pathfinding and in target finding can be seen here, that should be spread over several frames as a next step to smooth and improve the performance:
@MQ32: great! once I'll try to do this too (by taking into account terrain texture masks) - hopefully I can generated from triangles and merged into large entities?
okay, I thought of creating dx meshes with texture, as somewhere in the forum I saw in an example, and forming an entity of them. or they are remade in every frames?
okay, I thought of creating dx meshes with texture, as somewhere in the forum I saw in an example, and forming an entity of them. or they are remade in every frames?
I do this in my editor. While adding vegetation I creating new meshes at run time. If the editing is done, the whole new mesh is saved into a MDL file.
okay, I thought of creating dx meshes with texture, as somewhere in the forum I saw in an example, and forming an entity of them. or they are remade in every frames?
I do this in my editor. While adding vegetation I creating new meshes at run time. If the editing is done, the whole new mesh is saved into a MDL file.
i improved the whole system. It's still buggy, unusable and needs a lot of performance, but hey, i've got about 6.5 Mio polygons AND waving gras. I would like to use LOD, but i only have commercial
1) generate the grass-vertex normals the standard way, fetch the terrain normal under each blade and merge them in a 30%/70% fashion to get decent grass normals; as discussed @AckCon
2) Try to add weights to the grass vertices to let the upper vertices wave more than the lower vertices. If it is one mesh, you cannot predict if a vertex is an upper blade or a lower blade-part, so you could write that apriori into the vertices during creation! Then, evaluate the weight in the vertex shader and use that as waving-factor.
3) to get a denser look, try bending the grass-quads along the viewing direction in the vertex shader. That sounds like an awful idea, but looks good in practice.
4) if you reached some kind of density, you can fake AO if you assume that the lower parts of your blades are occluded while the upper parts are lit and pass light, since grass blades are by nature translucent. So no need for costly SSAO here.
5) if you want to use SSAO, then do it on the scene while the grass is clipped and render the grass after that with the trick mentioned in 4. With a screen space shadowmap, you could even influence the grass shading by a lookup on the shadowmap.
I don't use any kind of alpha maps or normal mapping, i don't even do lighting in any way. The whole gras uses your point 4 already, but that's the only kind of lighting i do atm. To point 2: They already have this (i do it with the height factor i also use for the AO part) To point 3: for more density i would need to create more gras blades
I'm not sure what is sky and what clouds (seems like the clouds are shining?), but it looks like jcl himself is coming from the heavens to earth. ^^ Beautiful!
The problem is: The clouds are really shining. I need some kind of "second sky set", where only the sun is visible and everything else is not. or at least not that hard
has been a long time since i've posted an update on my project. I'm still working on the Tactics-project.
Here's a little video and a list of my recent changes. Finally after finishing my apprenticeship as a software engineer for application development I've got more time to work on the project again.
The following things have been added by me since the last update: - Skill Handling - Main menu for equipping skills and armor to the units - Item handling - Pathfinding - attack-ranges, physical attack handling and damage calculation - damage-number handling - using items - effect-handling for skills and items - some minor tweaks and code-optimization - some really quick & dirty animations
I spent today some time with shade-c 0.91-s1, converting some of its parts I'd like to be A8 compatible. water is okay now similarly to 0.90 (there are some minor differencies between them), shadows also work fine (incuding alpha texture self shadows). Depth of field, hdr and some other stuff are removed, standard 3dgs hdr and fog works fine. next I'll deal with the terrain shader to not to present such ugly shots like this:
Posted By: Anonymous
Re: What are you working on? - 06/15/1321:06
@Roxas This looks really cool. What is going to be the max number of units on-screen?
@sivan if this is ugly I'm never posting project shots..!
EDIT @sivan I showed the screen to a family member and they said if it was framed they'd hang it on their wall. I agree and can't wait to see 'pretty'
@Malice: Thanks! Well to be honest I didn't think about that yet. I guess it'll depend on the map how many units can be dispatched. Theoretically there's no limit of units on-screen though the space is limited because a map is for example 20 by 20 tiles large ^^.
And update: I added some cinematic-camera-effects and reworked some animations. TacticsProject #6
and here's a video of the main menu. here's where you'll manage euqipment and skills of your units. it's not finished yet, but it's almost complete.
thanks Malice, but terrain is ugly from close distance, I want to use shade-c terrain shader, in a bit different way, currently I find it not the easiest to handle. and soon anybody can do similar things very easily (if you have commercial), I'm just waiting for the engine update to update my editor which includes shade-c water, customized pssm, and in near future I will add some other shade-c features like its own shadowmapping system (it will be selectable, allowing the usage of default 3dgs shaders too). plans for the summer beside a cool pathfinder I'm reworking alongside
after correcting the shade-c 0.91 terrain shader, I transformed it into a 8 texture blending version, but resulted in less than 20 fps, so I made a new one, blending only 4 textures with normal and specular maps + base texture with normal and specular map + lightmap, but the performance is still not good enough at all... maybe removing some things will help. or Heelx will make a better one, if I send him finally some graphics stuff what I always forget
Edit: removing tiled normalmaps and specularmaps helped to improve performance. but shadow blurring is still very ugly
by commenting out some lines, I transformed the above mentioned shade-c terrain shaders to work with standard pssm and other shaders without any shade-c stuff (only a simple material setting and one fx file required). not surprisingly, it is faster than our lovely terraintex.fx Probably I'll put it into my editor to be easily manageable, and later will also make a small example project as a contribution. entity skins: 1 base texture + base normal-map + (1 or) 2 rgba blendmaps + (optional) lightmap material skins: 2x4 tiled textures (using 1x4 is a bit faster). they are not quad textures like in shade-c, the 4 textures are merged as a row, that can be easily done by Gimp (make a film without borders).
yes, it is the very same I made for shade-c, I just commented out the shadowmapping related parts (I modified the original shade-c terrain shader what I found a bit illogical in connection with its skin arrangement, now it is much more easy to use. and it was too slow because of its high quality features I removed partially, but I made several separate but compatible shaders to enable choosing the desired complexity).
shade-c 0.91 applies variance shadow mapping, its shadow area positioning is okay for 3rd person cameras by default, but I need a very free rts camera, where it behaves very badly. finally I could achieve an acceptable positioning, but in some cases black artifacts appear that can be eliminated if shadow fade-out is not used. moreover it requires further post-processing to get smoooth shadows on terrain (it is very pixelated in some cases on higher terrain slopes), and better shadow edges. but probably it is not so easy, and maybe has a bad effect on performance, so I don't know what I can achieve... currently it is okay for a shooter in a flat world with sharp (but not pixelated) shadow edges.
sivan@ I couldn't adjust the brightness of the shadows... I wanted to make them darker, but I failed (cause of my poor shader knowledge). I wanted to mix them, with per pixel lightning, to have smooth dynamic lights, have you ever tried anything like that?
in the terrain shader: float shadowAlpha = 0.25; but it can be set between 0..0.5 instead of 0..1 it should be parametrized by using a global variable, but shade-c codes are sometimes not really organized...
I can imagne how big work it can be, as I have recently "fought" with shade-c updatings/corrections/modifications/extensions to make a test environment suitable for my needs... imo start it now, until it you don't forget what is what in your code, later it will be more difficult.
huhh, finally, shade-c smooth shadows are made to be really smooth, without any speed decrease. fast as hell, but not so easy, and flexible as pssm. it was tricky, required a new blur shader, and blending the blurred depthmap with the variance shadow mapping output, some cheating, and tweaking a lot of mutually dependent parameters. nearly ready for fun
yeah those sprite based far tree lods are not really fine, could be reworked a bit, but I'm too lazy. and found no better solution yet, which is fast enough. moreover, in some commercial games you can find more ugly ones but you normally concentrate on close view action, so as Jcl would say: it has low priority.
@rvL_eXile - Great looking models. It’s really cool to see some art work being posted here. Can’t wait to see more.
@Sivan – The editor is looking awesome.. This project is so very needed in the engine. It will help lots of newbies who want to create awesome starter projects but don’t have the skills to code all this "coolness" from nothing. I can’t wait to see more.
@Kartoffel - Your shader pipeline is wonderful I was happy for what you released to us and can’t wait to see how far you are going to take it . Really beautiful work.
@PadMalcom – Procedural anything is cool, it reduces the work on the user. 3ds max had this idea in this materials setup and through out its work pipeline. It was one of the best parts about the package. I personal would love to see as many procedural plugins for the engine as can be made.
Awesome work ALL, and looking forward to the next updates
@oliver: thanks, but I just reworked the shade-c stuff, most of the work was separating the scripts I really use, and reading and understanding the view/material settings and shader codes. I always do only what I need at that moment, this way my work is very efficient. I needed good water, faster shadows than the customized pssm, and a bit better terrain multitexturing (later I will deal with the all-in-one shader too, because once I will need a few features of it). now all is ready to put into my editor (water was done in March) to make settings menus to really utilize these features, checking in real-time everything, because without that it is very difficult to use. it requires some further code reworks, and attention to keep max compatibility with other stuff, but does worth the time. sometimes I really don't understand why shade-c 0.91b was left as is, it has great features, maybe even too many and too complex, it requires only a more user friendly environment. I know shade-c evo, which is more orgnized but seems to also lack user friendly setting possibilities, and furthermore not free for released apps/games.
@rvl_exile: your models are great. my only crit on the couch would be that it seems like it is made out of a hard material based off of the scratches on the texture... not sure how to explain it but the scratches make it look like it was made outof wood or metal
@lost & rvl to me it looks like worn-off painted hard leather. the thing that it makes the couch look like wood is the shape of the cushions but i guess that's as good as it gets with this poly count.
hey exile. im curious. when making surface normals what process do you use. I usually make a depth mask and run it through a normal map calculator, but I've considered actually sculpting just for a simple texture and baking it. wondered what your workflow is.
I don't really like dof blurring, like in shade-c (I started my shade-c exploration by removing it), but it is much smoother, I mean better I saw only one fps game where it looks really cool on outdoor scenes, but don't remember its name...
@oliver: the pipeline is a deferred rendering (only opaque) and forward rendering (opaque + transparent) hybrid so yeah it works but only with ugly lighting
@sivan: some current games (mostly AAA) have pretty good dof ;D But in older games it was just annoying in the most cases
@oliver: the pipeline is a deferred rendering (only opaque) and forward rendering (opaque + transparent) hybrid so yeah it works but only with ugly lighting
How does other deferred pipelines handle this problem? As far as I know CryEngine3 (Far Cry 3), GTA 4, Skyrim and many other uses this technique.
I decided to totally rewrite shade-c 0.91 object shaders to be really compatible with 3dgs material settings (specular did not work at all), and to work fine with dynamic lights. finally, everything is okay now: map entities, models, vertex animated trees, and multitextured terrains with or without lightmap. terrain got a simple per pixel lighting option because vertex based dynamic lights can look very ugly (but slightly slower). they are my first own made shaders only the shadowmapping calculations are kept (but corrected too):
yes, the basic stuff is absolutely ready, I will no post more similar screenshots (maybe the sdadowmapping part could be tweaked further but I'm not sure) later I will make a 8 texture blending terrain shader beside the 4 blending, but it will take only some minutes. and later modify the ones using normal and specular maps, but it will also an easy job using existing scripts. the main task will be to create a user friendly integration into my editor (it will help a lot to use it outside of it too).
@sivan: looks great, you've really put a lot of work into this A good visual improvement is tonemapping (but it needs an hdr image input) Maybe that's something you want to implement. But it needs to be used right, otherwise it won't look that good.
@oliver some use a second deferred pass for transparent objects (only ONE layer works with this) but atm. I don't know any AAA game which does that, however the most are doing exactly the same as me. With the right object shaders it actually looks good, but with acknex-default-lighting it doesn't
yes, later I will probably need it, I'm targeting tonemapping and bloom, the standard 3dgs hdr just works. there is always something new to learn... anyway, I'm very happy with the current features I hope this shadowmapping and shader system works fast enough with the huge onscreen character quantity I need, in large open scenes. in the other case it goes to trash. or maybe combining pssm with variance shadow mapping could be another nice thing to explore...
@PadMalcom: no, it is nearly similar but a different project. I will once share everything, just have to strip down and organize a couple of things. and I want to make a version that works with 3dgs pssm shadows to make everybody happy
today morning I modified standard pssm to psvsm (pcf smoothing is replaced with variance shadow mapping calculation). next thing is to blur the depthmaps.
@sivan: nice, but the screenshot looks like it uses no texture filtering. maybe the filtering in the sampler is set to point or none... or the range is wrong (should be [0...1] afaik)
yes, you are right it was not implemented, everything looked like little shadow squares. since the screenshot I added blur to depth views, now it is better, but I would need a better blurring... new pic with debug colouring:
still does not look like texture filtering is on? check your samplers for the shadow depthmaps and activate texture filtering by changing something like: sampler sDepth1 = sampler_state { //... MinFilter = point; MagFilter = point; //... }
to for example: sampler sDepth1 = sampler_state { //... MinFilter = linear; MagFilter = linear; //... }
The biggest problem with vsm is in my opinion light bleeding, when shadow casting objects overlap. A better alternative would probably be exponential shadow mapping, it is very similar, but needs only one depth value instead of depth and squared depth and has less issues.
thanks, I put it in my code. in this project I use only depth and a light bleeding calculation, no depth^2, based on GPU Gems 3 (now free to read), but will try out other methods too.
basically I need a fast shadowmapping solution over large distances and for lot of entities, so I have to try everything the shade-c based one is better now (I had a lot of work with its corrections), but I think this solution also can be okay soon... I know I have to learn a lot of about shadowmapping...
[ToDo] -No Details yet (coming soon) -have to create stairs to get on the first floor -need to create various floors (more broken) -add more Details at the Planks (broken!) -Texturing the Corridors
After countless tries I've finally managed to implement bad single-split shadowmapping for the sun with wrong variance-filtering and ugly streching everywhere, yay.
(I HATE writing shadowmapping-code ._.)
Posted By: Anonymous
Re: What are you working on? - 07/23/1305:21
@Kartoffel: Soon you will have a shaderframework much like (or even better then) Shade-C I hope you can/want to pull this thru, because i just lost my hope that Shade-C will ever be finished...
Finally got my native Lite-C hashmap working (STRING*, void*). Converted a terrible written c library into the terrible Lite-C c like language. Especially the variable types in Lite-c have given me a headache. Why not supporting unsigend datatypes, why is no real short type implemented? Why can I convert a char value to an int, but not vice versa?
shadowmapping is close to black magic, especially the variance calculations.
imo shade-c is great but will be always rewritten this is one reason why I use the old 0.91 and take out what I need. the other reason is that the new Evo is not free to use. it has great features, but also there are a lot of bugs, and lacks a flexible interface, and a lot of important things are simply missing. but definitely okay for learning purposes, and as a base to make your own shaders.
@Kartoffel: Soon you will have a shaderframework much like (or even better then) Shade-C I hope you can/want to pull this thru, because i just lost my hope that Shade-C will ever be finished...
however, I don't think I'll release the pipeline... It took me very long (should be more than a year by now) to get the results you saw on the screenshots I've posted.
Originally Posted By: sivan
the other reason is that the new Evo is not free to use.
If you contribute to shade-c evo you can use it for free.
If you contribute to shade-c evo you can use it for free.
it's a good news. probably I'll check evo in the future, I hope I can do some useful things... but currently 0.91 is complex enough for me to explore step by step, and evo lacks water and terrain shaders I badly need, which is fine in 0.91, after I partially rewrote them...
however, I don't think I'll release the pipeline... It took me very long (should be more than a year by now) to get the results you saw on the screenshots I've posted.
That looks very atmospheric, nice work! Only the front of the left wall looks a little bit off, esp. the texture (there should only be "full" blocks visible, i.e. you should use a different texture for those sides).
hey chris_oat, nice looking scene so far... but you should improve the edges (Wall->Floor), it looks so "setted up". Maybe you could create some decals or low poly "dirt planes", which you place around the floor in the edges, this will look really cool!
One Question, why are there orange lights at the lower part of the walls?
@Ascalon: Yes, MasterQ32 already implemented a Bezier algorithm for me so it will only be a question of time till a random amount of roads will be curved. Thanks btw
Posted By: Anonymous
Re: What are you working on? - 07/30/1306:55
The shot above is WIP so let me explane; Dirt: Dirt is probably the last Detail I add to a scene. Once a scene/corridor is finished and I have the right texture for it I will add dirt either to the Texture itself or by decals. Orange Light: This should be like an emergency light you know. Yes the model is still missing, all my tries to create one ended up pretty ugly
By now, the scene improved already a bit, maybe I will post a Screenshot today if you guys want
@PadMalcom: a month ago somewhere I found a more easy curve method than Bezier, where all the defining points are on the curve, so it handles better the endpoints. it is called as Catmull-Rom spline, and has functions in DirectX, available in 3DGS d3d9.h, as D3DXVec2CatmullRom, D3DXVec3CatmullRom, and D3DXVec4CatmullRom. (I planned to use it but finally made a more basic solution)
Hey sivan, great information, thank you! The advantage of catmull is, as you said, that it goes through ALL points on the track which is really useful to me.
@PadMalcom You should think about finding points in your voronoi net that are close to each other and weld them to create intersections or roundabouts (for 5 or more connecting roads)
@Marius: Yes I know I already implemented a variable in my voronoi generator to define a minimal distance between intersections... but I didn't use it yet
EDIT: @Sivan: I found it quite difficult to calculate the position value of the Catmull. It is not only going from 0 to 1.0 but has to be set individually for all ways.
I dont like your color matching, sorry, but the floor is much to orange and the Walls have two different grey shades, this looks not really good to me. also the colors from your Pipes... sorry but this is to much contrast. The Dirt looks fine so far.
Why do you use two different wall types? (Bunker /Brick wall Just use one Style (i would prefer the brick style). Cut some bricks out of it and show a bit of the bunker wall behind the bricks which are broken.
chris, which editor are you using? Surely you're not doing this in WED... or are you? Could you describe your workflow, show some screenshots? How did you end up with that scenery? I'm having so many problems creating good looking levels, and if I finally get to hack together something with everything in its right place, the lighting completely fucks up...
Come on you guys, I need some sense of achievement soon or I'll become tempted to switch to Unity
@chris_oat: The pipeline in this screenshot will not be released in public (it is for a [hopefully] commercial project me and a friend are working on)
However, as I've mentioned in another post I've already released a simple shader pipeline which I'm still working on from time to time. It's not as complex as the one from above but way more userfriendly and free.
If there's enough time I'll post a new version within this week.
Posted By: Anonymous
Re: What are you working on? - 07/31/1321:11
@Feindbild: As a matter of fact, I do use WED for all my levels and even to create some objects, like shelfs, crates or vents etc. Altough, i think when im done with a Level I will convert all in models for speedresons. I simply export the levels/models via WED to a .mdl-file.
You need alot practice for WED. I started with WED (Gamestudio in gerneral) back in A5-times
@chris_oat: I am pretty sure the scene would look 10 times better without any normal mapping. Despite the common belief adding normal maps does not automatically increase the visual appeal, instead it more oftentimes hurts the look of the game.
The problem is, that it's not easy at all to create good normalmaps. Especially generated normalmaps f*ck up the results.
When created and used properly they can increase the visual quality pretty much. ...but in the most cases users just use a generated normalmap from the actual texture - which is the worst normalmap you can get.
(heelx' normalmaps from the previous page seem to be really good, just to give an example)
@chris_oat: The pipeline in this screenshot will not be released in public (it is for a [hopefully] commercial project me and a friend are working on)
However, as I've mentioned in another post I've already released a simple shader pipeline which I'm still working on from time to time. It's not as complex as the one from above but way more userfriendly and free.
If there's enough time I'll post a new version within this week.
I have to ask again. Why not sell licenses of the pipeline?
Posted By: Anonymous
Re: What are you working on? - 08/01/1309:42
Originally Posted By: Superku
without any normal mapping
Are you sure about that? Do you know know ugly nowadays a brickwall-texture look without normalmapping or hightmap? Beside that, I do not overuse NM, but NM is (imo) a must-have for todays games that target a realistic enviroment.
@oliver: I've already said once that I'm not going to do that. Also, the pipeline isn't THAT good. (no shadowmapping, unoptimized, partially buggy and not userfriendly at all)
I'm testing schade-c 0.91 water shaders and shadowmapping in my editor. a water editor is under construction. the terrain shader is compatible with terraintex, supporting 7 tiled textures and works with the same texture editor perfectly (but a bit slower than the optimized quad texture version, maybe I'll make a little tool to convert them).
I see now a little chance to reach the desired performance and quality by my RTS game system, but I will need some pro features, so probably I'll put my editor to the contest i.e. it will be open source soon. I hope I can get some votes...
600 soldiers finding their path in a dense forest, some areas are blocked by lakes (everything cast real time shadows but no self shadows applied for speed reasons):
@rvL_eXile: The construction is not logic. The huge basin wouldn't be required for that fountain in the middle. The water would silently flow down the pillar. I would imagine something with "more action". ^^
Well knowing that I might bore you but some new screenshots of the final street generator.
So far there are 2 algorithms to generate a road network. One based on voronoi, one based on a simple rectangle generator. You can easily generate your own streets just throwing in some lines in a list.
Example:
Code:
List *points = roadnetwork_from_voronoi(50, -1000, -1000, 1000, 1000);
List *intersections = roadnetwork_calculate(points);
// Delete intersections which are too near to each other (optional)
roadnetwork_join_near_intersections(intersections, 100);
roadnetwork_build(intersections);
Next step: Generate normal and specular maps, apply some simple shaders and let streets react on weather effects (rain, sun, fog).
EDIT: Quickly added a parameter to randomly delete streets out of the rectangular algorithm.
Currently working on a remake of the Warlock. (Very WIP at the moment) Main purpose is to have a model with bones for TUST (and fun/practise). I'm also pondering about the style. I really like low-poly, but it truely looks like from the 90s or early 00s (which I personally like ^^).
As the years passed by the warlock grew stronger and stronger. So its appearance slightly changed:
The pendulum has way too few polygons for my taste, esp. because it's the thing the players focus will probably go on (when it's moving). The glass should be transparent, too, otherwise it's looking fine.
Some kind of voxel system for TUST: Level Size: 512x512x256 Voxels Level Usage: 256x256x10 Voxels The system will feature automatic mesh generation with chunks a 16x16x16 Voxels. Texture mapping will follow....
Yes, it is not that performant. I'm trying to increase the performance but i am bound to the vertex limit Here is an updated screenshot, rendered on my NVIDIA. The previous screenshot was rendered with an Intel 4000. I've added texture mapping support and improved block updating. Now you can change any block any time and the chunks will be updated correctly. A worst case update includes 4 chunks, in best case only one.
EDIT: I had fps_max = 60; in my main loop... It runs with more than 60 FPS i already started to wonder...
I've added support for custom block types. A custom block can be created with all DynamicModels functions, here an example with copying an existing model into a block.
Me and my brother are currently working on a custom minecraft client (he does the main scripting in V.S. C#, and i do all the graphics / shader stuff), and when using another variable-type you can increase the # of max polygons. Atm. Every 16*16*256 chunk is a single mesh.
However I'm not sure if thats possible with the rendering pipeline Acknex uses.
i've built Kubus first in C# as well, based on OpenGL call lists. It was pretty fast as well, i've also built some kind of automatic view range adjuster that rendered at 60 fps with maximum view range. maybe i'll do something similar in lite-c but i don't know yet...
A Fountain (yeah i know this is only the fountain bowl ), still WIP... any suggestions for more details?
Blender Editor Shot:
regards Sebastian
Are you still working on this? First thing I noticed: The bottom ring, the one sitting on the ground, is more detailed than the top ring. I would swap the two textures. Or, alternatively: Make the bottom ring just plain grey stone and add a rim of ornament only to the top ring (maybe a stripe in the middle of the place that you can sit on?)
I don't think I've ever seen ornament on the lowest ring...
If anyone has heard of the website, StumbleUpon, then you understand the idea of what this is supposed to do.
I spent about 2 days writing this program to randomly generate QR codes that in theory I could scan and follow to a random website, or get a random block of text or something.
Unfortunately, not a single one of the codes generated have worked so far (about 200). The idea is good, but I guess it just takes some luck.
@rvL_eXile The ring cup texture simmetry is very visible on its seams for my taste. A three stone pieces texture from bounds to bounds repeated without simmetries would look better I guess. I would also add tiles to the ring inside part to complement the nonporous bowl.
nice work. in case of a multitexturing terrain shader, you could generate blendmaps easily: based on the heightmap pixels you can define slope data for cliffs, and height data for underwater, normal base vegetation, and mountain vegetation textures as a start, and in some areas blend over another detail texture due to a noise map to get a bit better look.
Working on Voxel Terrain in GS. There is a lot of things on the todo list, but the base is now working. It is surprisingly fast (without the video capturing software in the background).
Working on Voxel Terrain in GS. There is a lot of things on the todo list, but the base is now working. It is surprisingly fast (without the video capturing software in the background).
Looks really awesome! What about different textures? I think a multitexturing shader would give the terrain the needed look.
What method do you use to create the mesh for the terrain? Cube marching?
Looks really awesome! What about different textures? I think a multitexturing shader would give the terrain the needed look.
What method do you use to create the mesh for the terrain? Cube marching?
Procedure for my voxel terrain: - Donwloaded a licensfree edgetable. - curse 3d gamestudio for not supporting initialising of multidimensional arrays - load edge table and tri table from file - use as less structs and objects for triangle creation as possible. So I modified the marching cube lookup procedure to make is faster and onliy using Lite-C predefined structs like Vector. Ok the stack is pretty full for some moments. - every chunk has it's own grid, to reduce the array size. I am working on a better method to save the voxel-data (volume textures) - I let Directx calculate the normals, that saves calculation time in the mesh creation itself. - There is a great inbuild vertex welding function in directx wich works very fast and very reliable (not in the video) - Using the Dynamic model to (re)-construct the mesh. - At the moment I am using a simple triplanar texturing shader. Later when I am solved the other problems, I will write a more advanced shader, like the thing in the c4 engine.
If you use a combination of normals and height (z-pos) you could achive a nice terrain editor that automatically sets textures like rocks, grass or sand
I would also like to see your Terrain Engine in TUST. For the connected chunks you can take a look at Kubus. It has a complete voxel chunk system like minecraft and supports huge voxel fields... Also i got some good ideas of implementing multitexturing into it.
PS: You can enable welding and normal calculations in DynamicModels Docs on how
@Rackscha Yes I will release it for free, but give me some time to solve the iusses with it. I worked on it for a week. It is not too complex code, the tricky part is to reduce the memory usage.
@MasterQ32 Ha, didn't saw that. Does the welding has a tolerance?
@Bone For the ease of use I use the function in a DLL for now, but I think you can just add it to the d3d9.h header file. Don't forget to implement the used structs too. function: msdn - link
I am working on a resource manager system. It will able to pack files into a resource file. It will have an interface to handle the files inside a lite-c program. You can store/retrieve/delete/view text files, image files, model files and so on... In my plans, it can encrypt/decrypt and zip/unzip the files.
I accept happily remarks and/or proposals for the system. I don't know at this moment should i attach a print screen image, on demand i can put here.
Right at the moment I am working on the event editor of the gui for my entry contest. All options which are visible, are already ready to use. As result the developer gets a ready to use script.
I will add some more Options until the deadline of the summer contest. Heroes, Enemies, Items, Skills and Equipment creation Tools are also ready to use. It's really a great oppurtunity to learn C#. =D
I've just implemented SSAA and improved the normalmapping of my rendering pipeline:
Also I implemented Y'CbCr colors with chroma subsampling for the diffuse textures in the gBuffer. This makes it possible to use just 2 channels for the colors and now I don't need a 4th rendertarget for that
Also I implemented Y'CbCr colors with chroma subsampling for the diffuse textures in the gBuffer. This makes it possible to use just 2 channels for the colors and now I don't need a 4th rendertarget for that
That's a nice idea. I wonder, though, if you save your diffuse textures that way (i.e. save the work to convert them during gbuffer rendering)
Like RGB colors Y'CbCr uses 3 channels. However, there are 2 color channels and one luminance channel.
The trick is, that you have to save the luminance (Y') in one channel and in a second channel you save Cb and Cr alternately, like that: (one [xx] = one pixel) [Cb][Cr][Cb][Cr] [Cr][Cb][Cr][Cb] [Cb][Cr][Cb][Cr] [Cr][Cb][Cr][Cb]
If you read them again either Cb or Cr is missing (of course) but you can just take the missing information from a neighbour pixel because they are likely to have the same or a similar color.
So actually you loose 50% of the color information (hue & saturation) but you still have 100% of the luminance (the more important part). And although there are one third of the information missing, you don't really see any difference.
edit: I hope you understand what I'm trying to say
I do, I study that shit What I meant was if you save your textures using YUV as channels instead of rgb so you don't have to convert them when rendering but rather just downsample the chroma channels
well, the shader doesn't work for all images (the more fine details the worse / unrealistic it'll look)
@rayp: should be possible but I think it won't look very good in motion (also, the performance isn't that good so there would be a significant fps drop)
finally converted my shade-c based shaders to SM3, after sucking a lot with under-water fog thingies... approximately looks and works as in SM2.0, but now I have more possibilities. the first one who finds the only object with a SM2.0 shader wins a free copy of MapBuilder (there are some incompatibilities, so I don't know whether I will release it or not, or if yes, how...)
@3run: IceX3 was used for this project. If the project will be finished in a few weeks, I hopefully will have more time and can finish/release the community version and write a manual.
[EDIT]: @Kartoffel: just saw you will release the sky shader to the community. That would be awesome
txesmi@ holly cow (not meant to offend anyone ), that looks sweet man. what kind of a game is it going to be? I imagine darkness, blood and body gibs lying around and monsters chasing a player which is running around this levels, and trying to survive with a flash light and a plasma gun (or some other shooting/killing shit) in his hands.. dynamic shadows from all light sources (from the flashlight and some wingiling lights on the celling or, so, so they make shadows move around, and bring some pants shitting atmosphere in the game
@txesmi: What happened to your top-down knight game?
As some of you may know I was on vacation for the last 3,5 weeks and as I could not work on Superku I programmed the basics from scratch for my next project in only a few days (on the flights and a few other nights). It works surprisingly good so far and it's a fast-paced RTS with a twist (hopefully) featuring 3 races, a basic campaign and multiplayer:
Btw. I'm pleasantly surprised by the speed of my beloved acknex, in the video and the screenshot you can see 300 animated models each one with 3 additional transparent sprites/ models (the smooth fog of war currently uses transparent models, too, which could be sped up a lot I think) moving with pathfinding.
@Superku: cool! are they bone animated characters? I see no shadows, but they will cost probably significant performance. of course it is heavily scene size and complexity dependent, similarly to pathfinder performance. collisions in my system eat a lot of power currently, I also thought of another solution...
@Superku: Wow, Felix, it looks awesome. And it runs so fast and smooth. Really impressive.
Are you going to make a Superku-RTS? With thousands of 'Superküen' (in german: 'mit riesigen Superku-Herden' ) 'Command & Conquer Superku' or 'Superku Total War'?!
Surprisingly this one is not going to be a Superku/ cow game but a somehow trashy battle for earth. Yes, I'm using bone animations and because of your post, Sivan, I've implemented my own type of simple shadows today (no selfshadowing, 1024² shadow map resolution). With again 300 units the framerate goes down to about 135-145fps, close up/ in regular view distance it's around 200fps so still pretty good.
(btw. just noticed that the terrain and unit shadows don't match, will get fixed)
@Superku: agree, a simple shadowmapping is quite enough, if you limit camera angle and distance, low ambient and high diffuse result in the feeling of self shadows, pssm would become too slow with many entities. I've never had such a high fps even if I put only a terrain on screen... probably pro edition features gives a nice boost. what kind of pathfinding do you use, and what is the map size in character-size units?
just before going to one week holiday, I converted shade-c variance shadowmapping into exponential shadowmapping. linear depth, filtering in log space. it seems to be a better choice. fast but not perfect yet, first shot:
little update: I think ESM ready with tweakable parameters, including 2 selectable blur type. it will replace VSM released in MB 2.2 and 2.21. throw them out
it is a bit faster and nicer than the original shade-c vsm. normally I get at least 40 fps, but it is level dependent, and there are a lot of things to tweak performance in the corresponding shader package too. I get about 30 fps with my RTS pathfinder having 490 units moving with vertex animation in a terrain only 500x500 tile sized scene. previously it was about 25 fps. the simple blur takes only the actual pixel itself, and the 4 adjacent ones (but sampling distance can be adjusted). I made a poisson blur too, taking 12 pixels in total. probably will make a gaussian too (sample 2x5 or similar). I used this little blog for my implementation: http://www.olhovsky.com/2011/07/exponential-shadow-map-filtering-in-hlsl/ (Boh_Havok also read it, if you check some outcommented lines in shade-c evo shaders), and read a short nvidia presentation, and some others I have not understood...
Sorry for the late response and thanks again! It's basically the flood fill approach but I'm currently having a few troubles with the unit-unit collision detection, especially when there are many units they start to penetrate each other or even stockpile. However, my three approaches so far which stop that lead to a chunky unit movement, I guess this will take a while before it works smoothly and solid.
@sivan: Why did you go for a 1366^2 shadow map resolution (camera size/ texture size?)? Seems a little inappropriate to me.
Boxes and crates are fantastic, aren't they? I find it funny to see them in pretty much every indie or AAA title (and to observe their appearances, their usage and so on).
(Don't worry, I will not post any more Superku screenshots here, but I've just implemented those, too!)
@superku: there is an index which lists the time in games passed until you see the first crate/box. That's why valve added their crated right after the train opening in HL2, just for a high(in this case low) score
Haha that's pretty funny but I sadly could not find it via google. I remember an article that basically said "When you see a crate in a video game you know the level designer has run out of ideas.", it used a picture of CS 1.6, thus basically 10+ years ago and from then on until today I liked seeing them in a game.
Finally I got a step further to solve my vertex welding problem in my voxel terrain thing. The artifacts you see are the chunk borders and an annoying rounding error I did concerning positive and negative coordiantes. I implemented a perlin noise generator to build random terrain. I think I will kick the automatic normal calculation and calculate them myself to avoid the black borders and sharp edges. Keep in mind that the terrain you see there is huge. It doesn't look like it, because the perlin frequency is too high.
Kartoffel@ maybe you could change the background depending on the genre of the music? F.e. if you listen to trance music, make some space or starfield at the background, so it will looks like you are moving throw it?
@Kartoffel: what I always expect from a music player is to have an option to display a simple frequency curve or frequency values corresponding to a parametric equalizer sliders, if you have this feature... but it is rather a musician habit... if something sounds strange I can see whether the quality is wrong or the actual song
these pics are specially for 3run (I have to maximize fps to 40 because of my crashing ATI card):
esm with 512x512 shadowmap, area covered is 1500x1500 quants (HD screenshot):
esm with 1024x1024 shadowmap, area covered is 1500x1500 quants (HD screenshot):
esm with 1024x1024 shadowmap, area covered is 2500x2500 quants, with 480 pathfinding moving vertex animated characters without self shadows (1024x768 screenshot):
sivan@ Thank you man, you screens made me speechless, this is just a pure awesomeness One question about the lightning, does it use perpixel computation (for all models + level blocks)?
Sivan, looking at your last screenshot and your new a8 pro, the very first thing you should do, is to switch to skeletal animations with Gpu skinning, as that's gonna be a lot faster.
@3run: yes, I found a little snippet that apparently sets fine per pixel dynamic lights, but shadows respond only to Sun. moreover I swithed to shader model 3.0 to enable supporting more lights than before. soon ready to publish.
@Slin: yes, this is why I mainly expect to get my pro key , and terrain lod.
Edit: and thanks for an earlier suggestion to check exponential shadowmapping!
improved the look and added support for windows-media-player playlists (some of them don't work because there are characters in the song title which 3DGS can't handle -.-)
looks realy great, kartoffel ! i also think, there shouldn´t be any details in the background. maybe the title in the bottom left, the tracktime in the middle and some buttons for play, stop, pause, loop or something else
Few days ago i remembered my old Digging/DK Imp AI demo i did a year ago.
Instead of just digging, the worker can now claim floors and walls by their own. And instead on stopping to work when a task was unreachable(game did a system halt) it now postpones those tasks and retries them later.
I updated my BlockShader. I hadded perpixellightning and the ability to transfer ambient and selection "light" by using the second UV(thanks to MasterQ32 for the hint).
As i have multiple blocks within the same mesh sometimes, it required a material/entity independed per vertex solution.
This results in the following: I can display claimed blocks/surfaces brighter than nonclaimed The selection is painted directly on the mesh, instead of using a big sprite entity above the level with a Texture that has a bluepixel when selected.
Oh, and whe have a light floating at the cursors position
I'm extending my PSSM demo shared a couple of months ago by exponential shadowmapping, which enables using only 1 or max 2 splits (using the default or Slin's custom view splitting function) over a large area beside good quality and performance. and features a fog fade out effect on shadows too, plus a debug mode to view coloured shadows for each split. I prefer the one using own object shaders (coming soon in next MapBuilder update), but it is more easy to use because of the stencil buffer thingie (what I don't like because don't really understand ), but hopefully someone would use it... a better shade-c water rip will be also included (one .c and two .fx files only), now compatible with SM3.0 object shaders too. and a usual boring screenshot, nearly the same as before, but a bit better:
@sivan: his post was related to Rackscha's post, he just wanted to show that he created something similar... It would be senseless to post this in another thread.
Recently I've created a cover system for NPC, it's works flawlessly, and uses tpathfinding plugin. It's static, so the cover points have to be placed by hand via WED, I'm working on a dynamic one as well.
Worked on the HUD, which is on the models now instead as panels. I know they look horrible and i need a 2D Designer whatever. It will definitely happen. My goal was to test if it really works and it did.
@3run: normally a designer made cover system is okay, it is applied in many commercial games (in older AAAAAaaaaaaaaaa.... games too). but with some hundreds of c_tracing you can check the environment in uniform distances (I have an idea in my head to implement such a map decomposition system, that examines the map, creates navmesh like areas to have less data, and feeds it to the AI and pathfinder system).
This exercise starts with that maze generator and it depends on it. I built a 3d level from its result with floor, wall, ceiling and pillar entities.
You may recall that I added doors to the result of the maze generator. I wrote a system that uses those doors witch open and close automatically as portals for geometry culling, so they are automatically shown or hidden by groups.
All those geometry construction pieces share their texture in order to merge them in a single entity by group.
I built those pieces with a secondary uv set with no overlapping in order to render the lightmap. In the merge process, I sorted them as you can see in the second image: a 2048x1024px lighmap for one single group.
I render the lighmap transforming the resulting sorted uv set to clip space instead of transforming the object space vertex coordinates. I used a previously rendered depth cubemap per light to calculate the shadows.
It takes its time to finish but the result is as fast as wind because there are very few entities and no shadow calculations on the flow.
I got some troubles rendering cubemaps because sometimes misteriously are not rendered properly. For now, I'm looking for a workaround...
Wow alibaba and txesmi, impressive work. Particularly like your use of lights alibaba and txesmi, always marvelous work. I found Vadim647's old bmap generator and am playing with it off and on, thinking about the dungeon generator but the shader work is amazing in your examples.
yeah, the shadows and ambient occlusion looks really smooth, great job! and I agree with superku, some antialiasing like mlaa/smaa or just ssaa wouldn't hurt
alibaba@ looks pretty nice, I liked how those warms were moving around the player) Where can I get the example of that flock behavior script? Was it released?
I couldn't even imagine this to go this far and this beautiful!! I'm impressed, it's just the best random level I've ever seen.. Can't even say anything shadows, they left me speechless.. Man, you are awesome
@txesmi - I see that you are running at 477 fps. HOW!! lol I'm using that algorithm, and I have all one wall model I use. The wall has 7800 polys. Is that the reason my fps is 10? My only work around is adding distance fog, and setting the walls invisible after the fog, but then my fps is max 13.
EDIT: The languages are mixes, because i had no time to record every sound i needed, so instead i used some old files as palceholders (mostly all english voices)
Iglarion, I really like the look and general quality and while the bloom might even be kinda physically plausible, it usually looks better to filter out everything except some really bright parts, while your shader looks like it is just taking the whole image and adds it blurred to the original one, making everything brighter and blurry, which I don´t like. A bloom just causing some extra highlights on the other hand might look good.
If I'm not mistaken the difference in all three screen shots is just the intensity of the "bloom". I like the 3rd one the best, too. The new grass is very nice!
Created the Face of the Building. Using 3x 2048 / 1x 1024 Tex for the whole Model. Polycount about 26k tris (not optimized yet). Details still missing!
I had a little time again for MapBuilder to make some new example levels for next update, featuring exponential shadowmapping (both for shade-c based, and pssm). I prefer the shade-c based solution because it has self shadow support on alpha textures, and a bit faster, thus some thousands of nice grass models can be placed into a scene. I decided not to wait for engine update, so hopefully in next week it can be out, and will be followed by some usual stuff like little contribution demos, and after bug reports the game script package will be also updated.
I prefer the shade-c based solution because it has self shadow support on alpha textures, and a bit faster, thus some thousands of nice grass models can be placed into a scene.
Once i placed 300 grass models, killed my framerate with shade-c. Guess i did something wrong, when reading your post.
MapBuilder currently utilizes only the (modified) water effect, and shadowmapping matrix calculations from shade-c (shaders are totally rewritten), everything else is removed. the new exponential shadowmapping is simply better and faster than the old variance shadowmapping.
in fullscreen I can get now over 30 fps, positioning the camera as seen above (maybe triple buffering, or using shader model 3.0, or loading the level from wmb - except grass - boosted it, I don't know), using 10000 grass models animated (using only ambient, but the shader enables using diffuse, specular and emissive too), with self shadows, without shadow cast by grass. I'm now satisfied with it later I will make a demo of this shadowmapping separated from MapBuilder.
I think I wrote the most basic bloom shader ever I don't really know how they work, this is a very experimental thing, but it works fast (no fps drop). it applies no blurring, samples only one pixel at a time, and there are simply 3 parameters to tweak: lower limit (to decrease dark contrast), luminance (produces contrast), and +/- brightness correction. atm it's okay, for me at least maybe it is not a bloom shader at all...
@Kartoffel: great quality, and good luck for the performance tweaks! first of all, you could make a reaper game, to cut off the grass to improve performance
just made a new seeding panel to place entities in large batches, taking into account terrain height, -slope, and game area, using 2 different methods: random placement or full coverage.
@oliver, yes some random pan and tilt. all entity vertical positions can be updated by a button, if terrain deforms afterwards.
just made some performance tests with 480 units pathfinding in 20 groups, utilizing pro features: terrain cache (lod currently results in ugly terrain self shadows), bone animation, and sprite instancing. terrain resolution is also lowered now. no magic speed, just acceptable results, now I can avoid using static lightmaps. and with some further character and group code optimization it will be a bit better.
high settings for future pc-s (I get around 17-23 fps depending on camera view):
lower settings for current pc-s (I get 25-35 fps depending on camera view):
I really wish I get as good as you guy. Working hard with my RST/FPS game. Testing my A-10 flight.
Posted By: Anonymous
Re: What are you working on? - 11/05/1319:11
@WrekWIRED Looks nice. Keep posting and working. And keep learning, I'd say that the really great looking stuff in this thread comes from people would spend the time to search and learn outside of the 3dgs area. But even great storms start with small winds so just keep moving towards your goals.
@sivan Every month something new and better. How long has the dev cycle been now? It's always looking more awesome.
okay, what it does is: - producing blur depending on a calculated brightness quotient, and a blur strength factor - then doing brightening/darkening dependig on a treshold value and a brightening factor (applying contrast) - and then adding some overall brightness/darkness compensation depending on a corresponding factor
after debugging bmap operations and code reorganization of MapBuilder, a new customizable gui is the target. it will be followed by a new entity modification system, and another camera option. hopefully later the panels/menus can be replaced by something totally different... a tiny first step:
Looks pretty good to me. Only the grass looks a little out of place there. I once got the tip from Iglarion to fade the grass out on the bottom of the sprite/ model, i.e. use an alpha gradient where your grass currently gets darker (don't know if that will work for you as you don't seem to use alpha transparency right now?). Additionally, the ground texture needs to look similar to the grass models(' textures) or vice versa. What you could optionally do is to change the blend map of the terrain (if any) or use vertex colors/ vertex information to blend and change the terrain texture slightly wherever grass is (or buildings and so on).
thanks. I use alphatesting in this case to be a bit faster, and because they need to be set to OVERLAY when a bit under water, because of the refraction view of the water shader. once I will rework my grassing methods, for example I want to modify their ambient due to terrain lightmap colour, and a procedurall grass mesh generation is also a plan. in this particular case, probably the grass texture should be modified a bit, it is too contrastous. by setting material colours I got very good results easily, but it had some disadvanageous effect when using a day-night cycle. but currently the gui rework is the first, it goes well
reworked the bloom for my small pp-shader pipeline and implemented smaa: (the first bloom layer is now 5-directional)
(still no idea for a small game... anybody got one?)
Posted By: Anonymous
Re: What are you working on? - 11/18/1309:10
Love the bloom cubes they make me think of Tron. So for a game idea what about a simple sidescroller. I like that the Tron world was modifiable at will by the programmer and made of bits that could be disassembled and reassembled into new things. So how about a game where the player can brake things down to bits and reshape the world.
Giff to be removed in a week - took hours to make this crappy art lol
new after-placement object modification added to MapBuilder:
ctrl+click to select a model or map-entity, then possible to move it by gizmo, and fine move by sliders/buttons as before. maybe I will add rotation/scale, but the sliders just work...
moreover, the gui panels have been made to be movable, and their position is saved.
@Malice: It's a really nice idea but I'm afraid it's something too hard / complex for me.
@sivan: nice improvements
---
I just made a little antialiasing test. I used 9x SSAA and a downsample shader which uses the rgb subpixels for more accuracy (something similar to windows clear-type):
meanwhile my gizmo works fine with move, rotate, scale, sliders only for fine tuning. the whole thing will be more useable, if once I get to the end of my to do list (what is always extended...).
Posted By: Anonymous
Re: What are you working on? - 11/21/1303:38
Quote:
@Malice: It's a really nice idea but I'm afraid it's something too hard / complex for me.
Surely you are being kind. It is not a hard idea at all. But it is a lot of work and you did ask for a little game. Sorry I just figured short and little could be equal. I'll think of something easier.
@sivan - awesome news. That was something I had trouble with on the last test release.
@Malice - you are not alone with this problem, so I had to solve it. the other unliked thing is the camera that I change in the following days. "I serve the working people" (a former military slogen that was used during the communism/socialism here)
Exciting ladder climbing to reach the incredible treasure that surely awaits in that golden chest?
(Okay, so the only real new stuff is the animation here. Sorry. Just didn't have much time to work on this, so I'm slowly going back and thus: small steps)
I've checked your indieDB page a few days ago, nice to see some progress (on the editor, too). Have you worked on (screen space) reflections for the ice levels and can you as a player test the level/ dungeon while/ after editing? It was a pain back then to design it and then having it getting deactivated again and again. Additionally, dungeons should only get published if the editor manages to complete it successfully. How is the status in that section?
Nothing has changed on the reflections so far - I just got swamped in work in real life, so I didn't find the time to do that. Especially since I have a lot of reading to do for that. I do want to try out more "8-bitty" reflections first, however. If they look nice, I may just stick with them. I hope that doesn't come across like I don't value feedback given - I do a lot!
When building your dungeon, you can set it's state to "TEST". At that point, you (but only you) can enter it, so from there, you can see what needs to be changed.
We're actually changing that system right now (well - the guy who's working on that right now has some university stuff going on that forbids him from working on it, but as soon as he has time, he'll come back to it), and we're trying a few things with that - we do have ideas that should hopefully make it much more smooth without increasing the server load too much!
Quote:
Additionally, dungeons should only get published if the editor manages to complete it successfully
There's a "VERIFY" state. You can also set it to that - if you manage to beat the dungeon (without using too powerful orbs), it's set to verified and only then appears to other players. Exception: If you wish, you can play a "Verification journey", where you verify someone else's (unverified) dungeon. In turn, you get great loot for you taking the risk of trying out a dungeon that may be impossible. This allows players who can't be bothered to still have active dungeons in the game.
looks smooth, do you use c_trace? I use always customized terrain normals because the default ones looks not really good. maybe they are badly attached to the limited list of possible normal vectors storable in hmp.
thanks! ---- what I'm working on: I thought about creating a simple multiplayer fps (like ut or quake) when I have more time. ...the hardest thing is going to be the multiplayer, i guess. I've never written any mp-games/applications before
Well... since I don't know anything about how multiplayer works it would be nice to have some resources or information on this topic, so anything helpful would be great Maybe you know some good tuts or similar stuff.
Ok, i'll try to find it, or reprogramm it. Have some ideasd for more compressed data anyway(if someone remembers, i posted some code for int to signed byte and vice versa )
Could you please provide a link to this encryption? I tried it myself, to simulate a 4 byte integer I had to use 3 single bytes, which just saves 1 bytes which is not that awesome^^
@3run: How did you manage to let it only rain outside?
I use region_check for the particles, so I place a region called let's say "no_rain" above the celling, so the rain won't pass throw it. Plus, I use those 'mist' particles, at the place, where I remove rain drop particles, so it looks nicer, here few screens:
Quote:
The same technique with regions was used in the template, provided by MasterQ32 some time ago.
A prescanned file containing the coordinates of obstacles are the same like regions ... You just dont have to save them in the world. Wont make a big differene at all.
You are using the c_traces to find the corners of the regions and save them to a file. A dynamic way to build them, you dont even have to open WED and place all these regions which would be very annoying for me.
@Puppeteer: looks good, I want to do something very similar in near future. having a real time debug mode is a great help in finding even minor bugs in pathfinding or ai system.
Implemented a texture LoD baker into my level editor. It generates a diffuse LoD texture from a blendmap and 4 detail maps for LoD terrain tiles far away from camera.
WASD + mouse: camera movement Space: change terrain display mode c: generate a new terrain x: new random pathfind f: make a screenshot
light green node: pathfinding start blue node: pathfinding target dark green nodes: points accessed for jump point generation dark red nodes: unprocessed A* nodes light red nodes: processed A* nodes cyan nodes: resulting path
No, those are a 33x33 terrains, stitched together. My next steps will be improved tree rendering, (i got some nice "instancing" ideads ) automatic multitexturing, maybe depending on slope and terrain type
Yesterday I've implemented also a new instancing idea for vegetation in my editor. The previous one I used (baking models together) was really fast, but increased nexus too much, because every single terrain tile had it's own, unique baked entity.
good to hear you find out great workarounds really missing from the engine features.
if there is an interest I could add to my editor auto multitexturing based on height and slope data, which are available at the moment so it would be fast. next edition is almost ready I just have to add a new simple free camera mode, and to remove some unused/unnecessary data related to RTS AI.
...what I have ever wanted to check is the scripts of the extra terrain toolset of unity 3, I don't know it is still okay with unity 4. it contained some random terrain heightmap generation and nice simulation of different kind of erosions afterwards.
yes, it is not difficult to auto generate. just set up a few height dependent textures with min-max intervals: underwater, seashore, grass, high mountain grass, and a slope dependent cliff texture that overrides the previous ones. or assign to the brush height or slope limit parameters, and pixels are modified only if values are within boundaries.
The shader based on slope data looks really nice. The terrain generation looks not really realistic. Looks like Perlin Noise? Maybe you could try an Erosion Algorithm?
Yeah, it's just perlin noise right now. I have to look into more advanced techniques like erosion and so on. But i think i got the texturing pretty good. It doesn't look like it repeats every two meters, but in fact, the texture does
Inspired by the idea of an erosion terrain generator, I'm working on a Voronoi heightmap generation script. Works very good so far. Needs some small improvements, then I post it in Lite-C Contribution forum.
Work on integrating new Shade-C Evo in IceX3 Leveleditor. Most things are working now. Hope I can release soon the anounced version of IceX3. But it's still a lot to do.
I will post the fixes, but there not so many Most things I've done are workarounds. I've added fog to shade-c evo, but for that I've had to kick out the emissive mask (never used this ever).
And the wrong fog color I reported as bug, is now working, but just because I adjusted the color by eye.
Sky is still blurred with DOF, but I think it looks nice in combination with fog (nice fadeout of level). But only if the sky has no clouds
And shadow has no alpha clipping yet (EDIT: I've solved it, now it's working )
great news. this evo pack is much more complex than the old ones, probably it requires a lot of time to really get familiar with it... I see you made a compatible terrain multitexturing shader, but I'm afraid a compatible water shader would be a much more difficult thing...
I already spend a lot of time months ago with the first version of shade-c evo. The API is very easy to use, it took me 5 minutes to convert my multitex shader to evo format.
A water shader without transparency would be also no problem. But if the water shall be transparent it'S quite complicated. I didn't test it yet, but I guess the current 1-layer transparency is not enough. Hope the final version comes with more transparency layers.
yes I see it's well organized. what I don't know how fast it works in case of large open terrains with many static and many dynamic/animated entities. I could take out and modify from old shade-c packages what I need to get a fast system, but there are some limitations... water: I think one layer transparency can do the job as a refraction view, and its coverage over mirror view is only modified by the water depth.
what I don't know how fast it works in case of large open terrains with many static and many dynamic/animated entities. I could take out and modify from old shade-c packages what I need to get a fast system, but there are some limitations...
This is what I will test upcoming week. But there's missing some LoD techniques in EVO that need to be implemented.
final testing goes well of MapBuilder 2.3, I just have to check all the example maps in all shadow modes and resave them to be updated, then ready. maybe I'll have time to add game examples too within the editor package for easier usage...
Indeed, but why is the entity count so high, because of grass models? Have you thought about merging them somehow together in small clusters?
This would only work for creating clusters dynamically every time the level is loaded. Otherwise I would not recommend clustering for objects like grass, because this increases the nexus dramatically. (Precalculated) clustering works good for (2D) trees only (and other big 2D objects).
yes, there are about 15000 grass models in this level, and fast enough with the custom shadowmapping based on the old shade-c one, but much slower with pssm (depending on number of views used). it's main drawback is slow creation, because currently cannot be handled by a wmb file (but maybe creating several wmb files can make it faster). I asked Jcl to solve it (max entity count is currently about 3000 in a wmb).
grass merging is not a plan, I thought of dynamically created grass made of dx triangles whose advantage is better terrain surface alignment, but yes, the nexus size can be a problem what maybe can be avoided by dynamc memory allocation which unfortunately makes things slower again, so I don't know...
Actually, batching grass together to clusters of maybe 5k to 10k tris is a pretty good idea. The only "issue" is that it needs to be aligned to the ground. So you will need many different meshes to cover everything. But for terrain if you´ve got a heightmap you could get away with just one mesh by doing the alignment in the vertex shader of the grass sampling the heightmap. Also in case you did not do it yet deactivate collision mesh generation while creating passable entities by setting the global variable collision_mode to 0 and back to 1 when you are done.
Edit: You may want to consider even bigger grass meshes because the few additional polygons you save from more efficient culling with smaller ones are most probably cheaper than the overhead of more entities. Or using instanced rendering might be quite simple to implement in Lite-C, the only harder part will be to provide the transformations to the shader, as you would then probably just create a custom grass renderer.
The only "issue" is that it needs to be aligned to the ground. So you will need many different meshes to cover everything. But for terrain if you´ve got a heightmap you could get away with just one mesh by doing the alignment in the vertex shader of the grass sampling the heightmap.
I've tried exactly this some years ago and remembered it now because of your comment. I did not get this to work back then because the vertex shader would not except tex2D commands, the trick is to use tex2Dlod (supported from VS3.0 and above) instead. Works fine and is amazingly fast:
You will still need different meshes if you don't want gras all over the whole terrain. A solution could be a seedmap, but I didn't try that yet.
If this is only for relatively small areas you could just move the grass under the terrain in the vertex shader. I gave this a try, see here:
It does not look very good but I think the concept is convincingly simple and fast.
Quote:
EDIT: Superku: do you have the shader code for this? Wanna try this in IceX3
What I did is the following:
Code:
float2 texpos = float2((worldPos.x+1024.0)/2048.0,1-(worldPos.z+1024.0)/2048.0);
float4 info = tex2Dlod(heightsampler, texpos.xyxy);
float height = info.g; // terrain height on green channel
float seed = info.r*2; // "seed map" on red channel, grey = normal length
Then multiply inPos.y by "seed" (the grass polygons should start at z = 0 in MED), multiplay inPos with matWorld and increase Out.Pos.y by some value (the terrain height) and height before you multiply it with matViewProj.
I already tried almost every approach to generate fast, massive grass for use in a real commercial(!) game, but no approach was satifiing. The approach mentioned above with placing grass via heightmap and seedmap in vertex shader is something I didn't try yet. I will make some tests upcoming weeks.
Edit: We added a fast path that covers exactly these kinds of things. Same code, "little" bit different FPS (it's capped at 1000):
Edit2: Added another fast path. 1 million grass entities, my CPU is bored as fuck... GPU on the other hand is currently crying about the fragment operations.
(to be fair, the screenshot is NOT staged, but the FPS changes quite heavily depending on how good it is with keeping up with all the fragments. But then again, I'm on a shitty HD 6770M, so with an actual GPU this isn't impossible to achieve in a stable way)
Nothing against the hard work of Rayne creators. But I don't think it will be more powerful than 3DGS if no state-of-the-art WYSIWYG editor and easy-to-use shader framework will be implemented. What I've seen so far was only a nice and fast engine for hardcore programmers, not usable for commercial games yet, but I know, it's still in development.
What I've seen so far was only a nice and fast engine for hardcore programmers, not usable for commercial games yet, but I know, it's still in development.
Since we are already this far down the rabbit hole of off-topicness, please don't mind me asking this: Why do you think that? When creating this post, the biggest real "fuck you" in direction of Gamestudio I could come up with was attaching the actual source used to create the example above. Granted, this is not a real world example, but comparing it with the trickery that both Felix's did before me... Do you really want to tell me that I came up with the hardcore programmer example? Because I, and that may very well be just my opinion, think that it was pretty straight forward. It's written in C++ and uses another API than gamestudio, granted, but so does Unity and it isn't a hardcore programmer engine either. I mean, you don't need to know what happens under the hood in the example: It doesn't matter how the instancing data is created internally, it doesn't matter how the resource loader works and what it does to deduplicate your data. The only real strange thing I can find is the used random number generator, which, honestly, you could've just replace with rand() % 200 and be done with it. Our rng generator just gives you some nice extra functions on top of that (used to generate random floats within a range) and the guarantee that the produced numbers are uniformly distributed. Given that rand() is considered harmful and usually implemented using a cheap LCG, this might not be the worst thing in the world.
So yeah, sorry for the rant, but I'm seriously curious why you think that's more hardcore programmer oriented than Gamestudio. Because for what it's worth, we put a lot of effort into designing the API and the default stuff so that it IS easy to use. Our shader framework may not be on par with Shade-C evo, but load a model with normal maps and/or spec maps and Rayne will automatically generate a shader for you at runtime that supports that. I'll give you the WYSIWYG editor, that is planned, but not there yet. One step at a time, we are just two people...
Besides, who in their right mind creates commercial games with an alpha engine anyways?
Edit: Throw in the following two lines and you got LOD and automatic removal of distant grass with the instancing:
Code:
node->SetPivot(_camera); // Use _camera to calculate which LOD stage to take
node->SetLimit(10000); // Max 10k visible around the pivot
Do that with Gamestudio in less lines of code, I dare you
My quote wasn't about the usability of the code example above, it was more about the current development state. I know it's alpha status yet. Just saying the current (alpha) state is for hardcore programmers, because of the missing script libraries yet.
I think he meant something like Shade-C EVO there. And yes, we don't have Shade-C EVO, but it's not like we don't have an easy to use shader framework. If you don't want to, you don't have to write a single line of GLSL or select any shader at all, Rayne is quite capable of generating its shaders at runtime. Truth be told, we only have one shader which is shared by every object in the scene, the rest is done "automagically" (and again, if you don't like it, rip it out and do whatever you want!). The concept is pretty similar to what AAA engines like CryEngine do: Providing one ubershader template and letting the engine create the rest at runtime because, let's be honest here for a second, there is work that humans should do and then there is work that computers should do.
Enough off topicness though, if there are any questions about Rayne, put them in the Rayne thread! I was merely posting here to just nudge you guys a bit, not to completely derail the topic.
Truth be told, we only have one shader which is shared by every object in the scene, the rest is done "automagically" (and again, if you don't like it, rip it out and do whatever you want!)
I think automatic or forced movement would be at least a nice option to have for a more difficult experience. If you don't know them already you should check out the abstract platform levels of Super Mario Sunshine, they are straight forward too and should give you some ideas, I loved them back then.
by automatic movement, do you mean that it just keeps moving forward or something else? it sounds like an interesting concept and can certainly add some challenge to it. however it would probably mean that i would have to redesign alot of the levels.
I cannot talk for Quad but I thought it would be interesting to give the funkybox a constant minimal speed so the player can only control the direction and jumping and attacking. Then the player would have to change the direction multiple times per second to stay more or less still on a small platform.
Btw. don't be afraid of redesigning (and deleting) things, this ... "concept" helped my current game a lot, it only got better with each level or even gameplay redesign.
Yes, and maybe the speed could increase over time slowly but steadily (not sure about this one, though). Crystals/ diamonds could then be used to slow the funkybox down a little (every or every 5 diamonds).
interesting idea, however i can imagine the idea about increasing the speed can be a little risky. there would have to be a limit otherwise i can imagine you can just jump straight trough entire levels. of course that could also be an interesting aspect of it since you could make places where you would have to gain a certain amount of speed to complete it.
Yes, that's true, then scrap that idea. However, the diamonds IMO should have some impact on the game and not only give a high score, maybe the funkybox could have better (more direct) handling the more diamonds the player collects but the funkybox consumes one diamond every second or so (collecting a diamond would give the player 5 of those then).
well right now the diamonds acts as a way to get more extralives. the red bar in the bottom fills up when collecting diamonds and when the bar is full you get an extra life. but then they still does not have that big an impact.
i guess if you had a gamemode where you slide all over the place then you could make that the bar, instead of giving extralives, it would allow you to brake as long as its not 0.
I finished a realtime customizable bevelled plate for my game. You can control the size in x and y direction, plus height and the length of the bevel slope:
Nice one! But I think the texture scaling for the bricks is too low (the bricks are too big, I think). Is the room beneath the bathroom shattered with the things from the bathroom?
Hey Christian, thanks for your Feedback! youre right, the Bricks are really a bit to large. Thanks i'll change it. Actually im doing the "outside" details. The Interior has no Details yet, so the debris from the Bathroom is still missing but this will be changed if i start to create the Details for the Interior.
Yesterday ive beveled the Edges of the buildings to get an smoother look / created an small shelter for the back of the building / retextured the rain gutter / created an chimney at the middle top of the roof and created a pile of debris in front of the building (beneath the large hole).
Next Steps will be: -creating some Pipes for the Rain Gutter -unwrapping + Texturing the Shelter and Pile of Debris -crating some Ivy -editing the Brick Texture
I just discovered replacing D3DXMatrixOrthoLH with D3DXMatrixOrthoOffCenterLH is a great thing in my shade-c based exponential shadowmapping. after doing some further improvements in view and shader calculations, the result is something really much better: - better area coverage - better edge smoothing quality - less flickering - same speed now I'm ok for today
it's not pssm only one view + a blur view. basically no need for more splits, it's now really cool even for a 3rd person game, what I want to do in the following weeks as a navmesh pathfinder testbed. I do not plan to deal with pssm anymore, too slow for me.
disadvantage: you have to use object shaders for those entities which should have self-shadows, what includes the corresponding calculations (but fortunately I made a full set for terrain 1-skin, terrain multitexture, sprites, models, grass models, trees, map-entities), not like in case of standard pssm when using stencil buffer that requires only setting the SHADOW flag. but "sadly" it is also an advantage: you cannot get engine bugs and if your model only has to cast shadow on terrain, no need for special model shader assigned to its material.
and I use the CAST flag in the opposite way to recieve only shadows, this way you can use static lightmap on terrain while not losing e.g. shadows casted by characters on buildings. and my editor can create static terrain lightmaps (slowly as hell).
@3run: next week it should be out, just I have to do some final tests and re-tweak shadow setting all of the example maps, which is boring... and the homepage update is also boring... and uploading is also boring... I'll get very bored next week
Vertices: 33.677 All Modelpieces merged together, this will not be happend, if the Model is finished! The Building itself has around 24k vertices / the interior about 8800. Enough Space to add some more
Wir -bastiuscha und ascalon- arbeiten gerade an unserem Multiplayerspiel weiter. Ein Level bzw Spielart ist schon fertig gestellt. Näheres zu unserem Projekt findet ihr in einem anderen Tread "Bullet Madness". Dort habe ich eine kleine Spielbeschreibung und Screenshots gepostet.
we -bastiuscha and ascalon- are working on our multiplayergame. One level is already finished. You can find more information in another treat called "bullet madness". there you can find further information about our game and some screenshots.
Actually it makes a lot of sense to place the chimney in the middle of the house. This way you can build a big tile stove which heats all rooms.
Okay, ive got only one Problem... Theres no room beneath the chimney, there only the corridors. Does anyone should recognize, if i dont place the stove directly beneath the chimney, just a bit more left for example?
Heres one Screenshot, (Chimney = red / Rooms = green / Location for the Stove = blue)
It has been a while since I've posted anything here I've remade from scratch one of my older projects called 'LAST STAND'. Here is a video: Last Stand WIP01
It's not going to be published as an Open Source project (sorry). It's probably going to be sold (as a template or so), when it will be finished.
I've already commented on your youtube video but I think you should not (primarily) try to sell it, the 3DGS user base is pretty low for that, but create your own game with it. If you add some enemies you are fine, just throw a few levels in there and you're done!
Btw. I've once (or multiple times) worked on a shooter myself but I could not get the shooting mechanics right. However, I implemented Superku style movement but as a FPS. You could run on walls (around corners, too), do double (and rocket) jumps, and so on, and to my surprise the movement in Titanfall looks identical to my old game. This may or may not give you some idea(s) for your own game.
the_clown@ I simply googled for them, then I've adapted them for Acknex one by one.
Superku@ I would love to see that movement in action, if you could show a video or so, that might look interesting I think that I'll keep player's movement as simple as possible (running/walking/jumping), it's just an fps at the end. I already have few ideas for enemies (this is the main goal for me, to make different behaviors for them).
rayp@ dude, CED rocks! I use it quiet a lot, but most of the complex sprites were created by hand.. Cause I've found some troubles in CED (with adjusting the height, width of the sprites).
3run, nice work, but what is CED? you could utilize special walking when the player takes up some cool equipment. magnetic boots from Star Trek, or some crazy stuff from Iron Man...
Did you do them by yourself? I like that the player can kick the barrels
I've adapted sprites myself, to make them work with Acknex (name+5.pcx type of sprites etc). You'll be able to kick enemies as well
Originally Posted By: sivan
what is CED?
It a tool created by rayp for me, used to create cycle of images (animated sprites). I'm not going to add any kind of special movements, but I'm thinking about some pickable modules, for God mode (10 sec) etc. We'll see how it goes.
Added "paint on texture" to the vegetation placement in IceX3 Editor. Now you can select a specific texture and the vegetation will only created on those parts of the terrain, which shown the selected texture.
Added "paint on texture" to the vegetation placement in IceX3 Editor. Now you can select a specific texture and the vegetation will only created on those parts of the terrain, which shown the selected texture.
Finishing a 2D perturbation fire shader that I left off on in 2005.
vs/ps1.1 required for demo scene.
Since this publish folder I've added pulsating coals and burnt wood. Right now I'm working on adding particle effect embers and heat haze. And dealing with the smoke.
I figure I'll work until I hit the max instructions for ps1.1 and then move it up to higher shader models and bring in depth and procedural noise.
Thanks oliver2s. I have a fixed math point now, and will seem to vary based on users computer. What do you suppose would be the best engine variable to use to keep it somewhat consistent across processors? TIA.
I'm currently working on a project. It's a third person stealth / heavy combat type of game.
Here are some pics of the beggining, Everything is still pretty wip and all the textures are still placeholders. But please leave your comments they are very appreciated.
I'm currently looking for people that are willing to join the project. Right now the most needed is a programmer in C#
Implemented a new approach for grass rendering into IceX3 editor. It's based on the shader method Superku mentioned some pages ago in this thread. I also added smooth fading out, better shading and color correction. Works incredibly fast.
The shader works like this: - the grass is placed by vertex shader from a heightmap of the terrain - the shading of the grass ignores self shadow gouraud shading. Instead the exact shading as the terrain is used, based on a normal map which was generated out of the terrains normals. - the grass color is lerped into the color of the terrain texture beneath the grass. For this a diffuse texture is used, which was generated out of the terrains texture.
Problem is that the grass have just 2 quad polygons (crossed). With standard gouraud shading it doesn't look really good. You can see the same shading problem on the trees in the background (which are also crosses 2D sprites). I guess I have to write a special self shadow shader for stuff like this.
That´s why I proposed to interpolate the normals, so you may just use 30% of the grass normal and 70% of the ground normal or something like that. If the grass is randomly rotated it should look quite good.
I've implemented your method. The grass normals are used by 1/3 and the terrain normals by 2/3. In the far distance only terrain normals are used, because otherwise the smooth fade would not work properly.
@oliver2s: Nachträglich auch von mir alles Gute zum Geburtstag.
Deine Screenshots erinnern mich irgendwie an das neue Rome Total War. Was mich jedoch stört, das sind die Bäume im Hintergrund. Die sind viel zu blau für diese kurze Entfernung. Das sieht unnatürlich aus, vor allem auch weil die Hügel und Felsen in der Ferne NICHT blau sind.
Danke. Mit der Entfernung des Blau hast du recht, das wird noch korrigiert. Dass Bäume blauer erscheinen als das Terrain/Hügel liegt wohl am Bug im DOF Shader, das HDR verstärkt wohl noch den Effekt. Denn Bäume und Terrain haben den gleichen Fog-Shader.
I don't know much about (modern) modeling and texturing - thus a 2048² texture for a garage door seems a little excessive to me - but I think the house (+ garage) look really good, well done!
Decals will be additional for the Finished Model, cuz some Engine cant even handle decals really good (Shadow Problems). Inside/ outside the Garage are already some Decals but not really nice made yet...
Wrote a color scheme picker tool in Java (Swing). You can load a template XML with all predefined colors, the GUI is generated and then you can pick each color separately. Exports to human-readable XML.
I really like the lighting on the second show, great stuff! Is there a pink light in the first shot, that I don't see? Because there a pink fragments in the lionhead.
How do you blur the scene for DOF? Box-, Gaussianblur or something else?
[EDIT]
Originally Posted By: Kartoffel
@HeelX: great tool, can be really useful
I was too tired to do that with Photopaint and I also wanted to begin with XML code for my project, I thought that was the best way to start with it
I was very suprised that colorpicking in JAVA is working actually very well, just a few lines of code to do that.
I'm currently working on some improvements for my rendering pipeline. (improved the lighting, material shading, depth of field, bloom and SSAA):
The shader of the lion's head looks very good.
But I'm a little bit confused about the DOF shader. It looks more like a model with a very low-res texture. The edges (border) of the model is too sharp. I think this is the reason why the ball with the white texture looks much sharper than the basin/bowl under the ball.
I need some critics for the following forest scene: please make critics about the vegetation movement in the wind. Do you think it's to much or is it ok?
The animation of the LOD trees looks really bad. Also I think the grass is moving too fast compared to the trees. So you should probably make the grass move a bit less and the trees a bit more or something like that... I really like the trees in the foreground, are they yours, or can I buy them somewhere?
Even with a 3 years old laptop, why is it that slow? What are your laptops specs? Are there that many trees outside the visible area? Do you blend out far away grass? Does the recording kill the fps, or is it > 30 fps without shadows? Just wondering as it should be faster...
@Slin: the LOD movement is indeed bad at the moment. It's hard to make them move better, because they have only 8 vertices and they are clustered. The trees and grass models you be bought on Dexsoft Games. My graphics card are a Mobility Radeon HD 5450. The last AAA game I played with this was Fable 3 and I could only play if I set all graphical effects to Low. The grass is only visible in a range of 100 metres. It's the shadow that kills the fps and even all other Shade-C effects I've activated (SSAO medium, Antialiasing, DoF, HDR).
@Harry: the movement is basically really like a metronome, it's based on the 3DGS default tree shader. I have to implement some random wind speed and direction variable.
@oliver: in a GPU Gems there is an article about vegetation vertex animation. maybe in part 3, which can be read on NVidia site (but no source code as I remember). basically your scene looks very good for me, only some tweaking needed already mentioned above.
Edit: the gpu gems stuff is fine for small oscillations of leaves, and looks fine when combining with the original 3dgs wind oscillation. just implemented (not all features of course).
Here is one screen of my current game that shows a path and two palmtrees. The left one is the previous version without shadowmap. The second one is with shadowmap
@Sebastian: the poly count seems to be ok. The texture could be in higher resolution. And maybe you could post a screenshot with normal mapping and shadows applied?
@Benni: the new palm tree looks definitely better. The border between the grass and the path looks too hard, there's no transition.
HEy Oliver, thanks for your Feedback. The Texture resolution is 2530*889 px i know, not really a power of 2 Just added the ivy directly from GIMP to see some results in Blender, to get a better feeling for the leaf "density". The lack of Detail in the Texture should be come from the mipmapping / bad alpha clipping in Blender itself.
If ive made some more Models for the whole environment, i would create an "ingame" scene with Normal mapping / shadows and all the fanciest shader stuff
Nice shots! But it looks to me as if the normalmapping on the first and the last shot on the bricks is not correct. I believe on the first image, the handedness is wrong, the normal is pointing into the wrong direction. Do you evaluate the handedness due to inverted UV coordinates?
shader stuff reorganized, now considering support for a navmesh pathfinder (Peragro) to have fun with some enemies easily... a random terrain generator, triplanar texturing for terrain and rocks, and physX support also would be fine... someone could support me with work or a time machine... and another boring screenshot with seeded vegetation, manually placed buildings, rocks, terrain deformation. only a few minutes needed to do. some more free building and better vegetation models also would be good to have...
thanks. yesterday I tested CryEngine3, and unfortunately it is far from that quality beside similar performance... probably shade-c evo can do more after some bugfixing... it would be more fun to spend more time with game making than visual stuff and tool development...
yes, the main characters require some more advanced stuff, normal maps, specular maps... a bit later I had to do it if I make a little 3rd person game.
I need your opinion again, fellows. I've implemented triplanar texture mapping. But somehow I think it looks worse than the standard texture mapping, especially for rocks/cliffs. What do think?
If all your hills are looking like this, then i wouldn't bother with triplanar texturing though. For a terrain it only makes sense for steep angles, like cliffs and the like ... or if you are using voxel terrains
standard looks a lot better, but this is mostly because you use one texture for triplanar mapping. Of course the top of a rock looks different than its sides, so you should use another texture for the xz and yz plane (z pointing upwards), which has mostly vertical wrinkles.
As i was working on a terrainshader last year, here's some more ideas that might help in improving the overall look.
Adding noise and optional heightmaps to your splatmap can really improve the look of your final texture. Especially rocks benefit from this technique:
original
noise on splatmap fetch and heightmap for better blending (gives a more 3d-ish look and creates "small rocks")
since we already have the heightmap for our rocktexture, we can also create ambient occlusion in texture space very cheaply so the rocks aren't peter-panning anymore. (ambient occlusion has been overdone in this image to show the effect. The final effect should be a lot more subtle)
Another thing to note is that there isn't a holy-grail-blendmode for mixing the final texture. In fact, it makes sense to create different blendmodes for different types of materials. For example mud mixes with it's surroundings, while sand and snow lays on top of things. Generally speaking, start thinking in layers and have a look outside to see how different materials react with each other. In the end it all comes down to the question: "Will it blend?"
original snow, using "default" blending
fixed snow, using replace instead of mix
final snow with noise on splatmap lookup
Creating a nice terrain shader really is a neverending story. Almost as bad as creating "perfect" shadows. I'm not saying you NEED to implement any of this, just thought i'd share some of the stuff i did while working on my masterthesis in 2012/2013
[edit] Re-Reading this, i got pretty off-topic and this hasn't much to do with your original problem...oh well
[edit2] Also, to create a more "random" and less tile-ish look for your diffuse texture, you can use a cheap trick (if you aren't already doing that anyway). Fetch your texture with original scale/tilingfactor. Then fetch them a second time, this time without/less tiling/scaling. Now create your final diffuse output using this formula:
What do you mean by that? I mean it still looks pretty much like a linear interpolation, but with the blend map having a steeper transition?
I think simply the snow tilemap gets always max weight in blending calculation, even if other tilemaps have a non-zero blending value. it seems to be a good idea to put it into the terrain shader to make blendmap creation easier...
@pegamode Really nice models, these last ones you posted. I like your game's art style.(Let alone someone liking it, having a distinctive art style is a big deal )
I was fed up with editor and shader stuff, so now dealing with my RTS babies again. reworked the main character update loop, and started to make some simple debug tools showing actual state, order, path, selection, AI processing time in ms etc. very far from ready.
How does it look/ work when you try to send your troops through a 3-4 tile choke or when one half of the squad moves against an obstacle?
just debugged this part yesterday groups now go through without problem even if arriving at the same time to the 3 tile wide, 7 tile long gate. the screen shows an A* seqarch of the main group leader (non examined black, closed list red, open list blue):
the full system is basically very complex, tasks are spread over a few frames resulting in a balanced load, and related both on pathfinder and movement parts (I had to draw by hand several flow-charts to get it work). I use a hierarchical pathfinder consisting of 2 parts:
1) the group searches a path only among clusters (actually 8x8 tiles for terrain, but varies after terrain decomposition, and in case of buildings. in Company of Heroes they use 10x10, it is very game dependent).
2) the tile pathfinder is simple: if the straight move path check fails, the 1st row uses an A* like thing towards next cluster middle. others only follow the one is assigned to them as sub-leader. on the picture you can see the search area, it is not big. in case of leaders if pathfinding fails, they use another search ignoring other units to avoid to be got stuck. it leads of course to collisions. (this basic group formation and target finding system will be totally rewritten to be more effective and to show visually better group cohesion)
thus in the movement part as a quite new feature, I implemented a push/bounce collision system, which works surprisingly performcance effectively, utilizing simply c_move (I was unable to do it with PhysX). collision length is measured for each unit, and handled in multiple stages. they can repath after a certain frames, or if everything goes wrong finally they can ignore other unit collisions, but it is a very rare situation so nearly never visible. and there are some counters measuring if a unit is got stuck, then they are not allowed to search a new path in every frames to avoid high CPU load.
later I can make a video, but there are some issues with accessing 2nd building floor (that was okay in past so back to backups again...). I have to work on more visual debugging tools as they make the work much easier and faster. and afterwards, I'll rewrite the army/group/unit/entity structs to be flexible and really game ready. it will be a heavy job, but I have a nice road-map with exact plans, like potential map creation for AI decision making etc.
Still not convinced by a c_move (movement) approach but otherwise, very cool! I'd be interested in a video of a group of units moving through the gate (and accessing the 2nd floor would be cool, too, if you manage to fix it)!
EDIT: @rvL_eXile: The greenhouse looks pretty cool. Only some broken windows look a little weird, I don't think that glass breaks like for example in the first shot the window with the oval shaped hole. I could be mistaken, though, and it may depend on the glass.
Still not convinced by a c_move (movement) approach but otherwise, very cool! I'd be interested in a video of a group of units moving through the gate (and accessing the 2nd floor would be cool, too, if you manage to fix it)!
thanks okay I'll make a short video maybe next week. the floor bug is just fixed a few minutes ago, beside some other wrong helper functions...
I decided beside c_move after checked again some Total War games, where there is apparently no exact unit pathfinding, only a group level potential field, and a formation keeping potential field for units, so units simply push away other units when groups are going through each other. this is why you can see many pathfinding problems, which is lowered at gates and towers/walls, where they use probably a different, maybe a hardcoded movement. I can use this approach until I find it effective (very short collisions).
I made a debug shader using pathfinder data - the terrain is fine, but object shaders show somehow shifted values, despite they use the same function and sampler... hopefully nearly realtime debugging will work in this way...
@Sebastion: were can we get all these models if they're finished? Or are they only for internal use?
Dexsoft but, theres just a little bit more to do: EXTERIOR: BuildingA Trashcan Greenhouse Glas / Door Greenhouse Plants Fence Small brown Concretewall Gate Building B / C / D / E Garbage Container Postbox Streetlamp Trafficlight Road Blockades Small Plants (Bushes leafs etc) seperated Plattform for Trees Hydrant Kiosk Signs Exterior Lights (Small, placed at the Front) Street Set ( Destroyed / undestroyed) Bench
I forgot to mention that your vehicle editor crashes if put in a long path folder, as normally any other gamestudio project. but works fine if put in a proper place. (I tested only shortly)
@Realspawn: looks great. Isn't it possible(yet) to add force to the ball when the platform moves into a direction while the ball collides? E.g. you move left, and when the ball collides with the platform it adds some force to it to the left.(or rotates the speedvector)
As it seems to be quite hard to hit the last one sometimes?
thx i am working on this game idea in my spare times and i will try to add it but slowly i need all the basic stuff to work and i am pretty proud of what i have so far
i got a lot of stuff i want to add but sometimes i spend hours to solve one thing so i take my time as i do it all in the name of fun all the level stuff is basic and will be replaced for better looking stuf when all works
Rackscha, I hope you feel appropriately embarrassed for your pun. But you could make up for it if you, from now on, make it a point to always comment with a terrible pun on every image posted here!
Anyway, here's an ugly screenshot that highlights the fact that we should make some of Dungeon Deities's screens a little prettier. In any case, Dungeon Deities now forbids the usage of certain orbs when verifying a dungeon if they are too powerful or allow the player to beat certain puzzles or rooms that cannot be beat normally.
That ugly lock is a 3D-model and shows some slight animation to make up for the pain of not being able to use your favorite orb.
At work we have cups or something like that in each office here in our developer-section. Flat/Bad-Joke or saying a line froma movie: 1 cent It sums up^^
nothing special, I'm just happy with my corrected normal map shader that simply works with amb/diff/spec/em and dynamic lights. I needed more time to remove all little issues than expected... (only the tower has it, other walls have the same texture without normal map)
skin shading, first attempt... ignore that ugly hair ;(
the light-positions and -colors are exactly the same, the only difference is the way the lighting is calculated. the right image uses the skin shading (notice the smoother lighting and more natural color)
more skin rendering: improved overall appearance, implemented specular lighting and added even more crappy approximations (screen-space sub-surface scattering and stuff like that)
(but again, everything except the skin looks ugly as f*ck (hair, eyes, etc.), sorry for that)
edit: specmaps were too high... here's a newer (better) screenshot with comparison between default shading, and skin shading:
thanks. the scene produces an acceptable fps depending on quantity of visible characters, collisions, and path recalculations, approx. between 23-35 (the average is 27-28). the new group "class" will hopefully result in better group cohesion and performance...
@Kartoffel: Very cool, looks pretty good. (I'm a bit stunned by the stare-into-the-future stuff, this is really great work). But i think this is the best skin shading gamestudio has ever seed
I have not checked it in shade-c evo, and don't know what is volumetric fog exactly, just had an idea today morning during having a shower to create a fog that fades out upwards, and also fades out towards the camera a bit, and add it to distance fog. I just added a 4th parameter as it is much better if not linear as distance fog.
Another tip which made my fog look even better: use power/square formular to calculate fog, so it doesn't fade in linear but instead smooth fade in which becomes stronger in far distance.
thanks, not secret, and quite simple, it will be out with next MapBuilder update with a control panel, and save/load. unfortunately vertex shader based so some artifacts can be visible on very low poly tall objects. but okay if used with care...
The AI looks pretty solid and intelligent. You should add some friend/foe recognition system so you can put them also into multiplayer matches that are team deathmatch or normal deathmatch. I assume they only seek players right now, right?
I and another mysterious person (it's Turrican) are working on some super secret game project.
This weekend we were creating a new menu, based on this idea we had:
...and that's what came out:
All sub-menus are placed in 3D space and when selecting a sub-menu the camera moves to the right position and angle. We'll try to make an ingame video on the next weekend so you can see everything with all it's fancy animations in motion.
Working on multiplayer mode for my Vehicle Simulation API. Here I made a small video, please don't mind the quality and laggy performence, it comes from the crappy recording programm is used. The multiplayer itself does not lag.
^^ That, Is, Epic, alibaba Really enjoy the Intensity!
I don't have a video, but I wanted to show a little bit of what we've been working on in the last 9 Days, on and off. There are a few place holders throughout it for now if it looks a bit off. 80% (more or less) was handcrafted with help from CGtextures.com
(that's not part of his right arm, its a pillar ^_^)
@DevoN Looking very promising. But for my taste ure showing too much of players arm. Maybe reduce the camera.arc a bit. But thats only my personal point of view / taste. Looking forward to an motion picture. Another thing, i would add some lights near the already placed with very low values 2 get rid of the exact circle - shapes. For a really pro looking flashlight / lighter ( not saying yours is bad, did not see it in motion ) i can hardly advice Shade-C[evo] btw. Could provide an example for that, if needed by someone.
@Rayp: Thank you Perhaps I will reduce the amoount of arm showing a little Would like everyone to enjoy the 'taste' of this game That is a great idea for enhancing the lighter. Will implement that for sure, Thanks for the kind words.
inspired by Unreal Engine 4 stationary light shadows (probably Unity also features something similar), transition of close distance dynamic shadows into static ligthmap shadows: its usage is simple in MapBuilder: set entities to have lightmap shadows, create lightmap, then set them back to cast dynamic shadows (not compatible with pssm)
yes, testing it, ue4 is a huge engine with a lot of great features I don't really need currently, but seems to be a good future option... and maybe the source can give good ideas.
another little new feature, idea is also from ue4, a fake sky lighting: entities have a 2nd fog and this colour is blended also to the sky cube (fading distances are adjustable), thus a smooth terrain-sky transience zone is the result without performance cost. purple fog, pink sky light:
In case you haven´t yet: think about using two ambient terms instead of one for your models and blend between them based on the normals y value. This way you can fake some kind of sky light and indirect illumination from the ground and it is at no real performance cost.
I thought of using 2 ambients, but in a different way, one for the Sun side of the sky cube, and another for the darker side, which could be cool when Sun is at low level. but this normal dependent variation also sounds fine!
In case you haven´t yet: think about using two ambient terms instead of one for your models and blend between them based on the normals y value. This way you can fake some kind of sky light and indirect illumination from the ground and it is at no real performance cost.
One could also look up the ambient color from a blurred version of the skycube itself, based on normals, that way it always fits the visible sky and more or less the ground, given the skycube has one.
before releasing next MapBuilder update, for better testing possibilities I extended the test player movement with gravity, jump, and crawl beside existing idle, shoot, walk and run, and added some other checks preventing under-water movement and handling bad polygonal map-entity collisions (maybe engine bug?):
I found those known troubles trying to render a sport center with its narrow and contrasted ground lines:
I tricked the color mask first four mipmaps and incresased the lines thickness on each mipmap step. I used tex2Dlod with a fraction of the distance to the camera in the X axis: a very especific set up for a very specific game camera.
ups, did not try it. My previous experience did not really help me in these cases.
Anyway I gived a try and it does what I supposed. The lines are so narrow...
I finished the settings and I'm pretty happy with the result. It is very cheap because I compute the lod stage at the vertex shader and use it into a straight single mask sample. I added a camera height divider to the algorithm and it works now for every point of view around the playground.
Auffällig schlecht sieht sieht das kaputte Shading der Gebäude aus. In Blender zum Beispiel gibt es sowas wie den "Edge Split" modifier, wenn du den einmal auf deinen Gebäuden anwenden würdest, würde das alles gleich viel besser aussehen.
Auffällig schlecht sieht sieht das kaputte Shading der Gebäude aus. In Blender zum Beispiel gibt es sowas wie den "Edge Split" modifier, wenn du den einmal auf deinen Gebäuden anwenden würdest, würde das alles gleich viel besser aussehen.
So wie ich das sehe komm ich wohl um Blender nicht vorbei.
Allerdings kennst du dich ja gut mit den Möglichkeiten von Blender aus. Wärst du bereit mir zu helfen? Ich würde dir mal ein Hausmodell zukommen lassen. Dann würde mich interessieren wie das dann durch die Änderung mit dem Blender aussiehtt.´
@grinseengel... it isnt that complicated... Just import your Mesh into Blender, add an "edge split" Modifier to your Mesh, apply it and export the Model... Done...
hi, I'm just interested in the MDL related pipeline of latest Blender, can it import/export directly or what is the proper/recommended procedure? thanks. p.s. sorry for being a bit off-topic
Directly import MDL into Blender? Never heard about that... If i wanna import MDL Files into Blender (what isnt possible in my opinion), i import it into milkshape, export the MDL file as .obj, import into Blender, do what ever i want to do, export as FBX.
inspired by Emre's great water shader, I added a secondary wave tweaked by a single new parameter to my water shader to get non-repetitive waves, resulting in both lake or river like ripples achievable. I know a video would be better, maybe later:
I just hope once I can make a more life-like effect... but it does the job for non-sea-battle games...
hmm... indeed a plastic like setting, what is less visible in movement. but decreasing Sun reflection and separating speed and scale seem to help a bit... no place for more sliders
-- This is the "main function"
fps_max = 60
level_load("")
local entity = ent_create("_cube.mdl", vector(400, 0, 0));
function update()
-- Update game here...
entity.x = 300 + 100 * math.sin(0.1 * total_ticks)
end
it's just a star shaped bloom shader >.> ( you're clearly playing too much shooters xP ) edit: it's just the suns brightness I guess.. it should be waay higher but I didn't want you to go blind with a little bit more realistic brightness, staring directly into the sun would look like this
but thanks! (even if the water shader could use a lot of work and optimizations ._.)
reworked my water shader, and modified the editor panel accordingly. hopefully it can produce less plastic-like results, and if using a higher poly water terrain (65x65) vertex waves have real effect (but it could be improved further). shore fade-out is also smooth now.
new reorganized gui:
stronger ripples (normal map based) and vertex waves:
blending in a foam texture (can be set to have wave height dependent intensity):
just added a little terrain editor feature to get straight horizon in case of sea/ocean levels, enabling one button setting of minimum (and max) terrain height, because of the water shader transparency behaviour (water becomes transparent if nothing is under, so ground minimum should be set just under max water depth). now time to upload update.
Afaik that problem related with DeptView. You are A8 user right? You can clear the view's background. Set NOSKY flag for depth view and clear it's background.
I don't really need currently a water shader that looks very good from close, I need something just fine for a strategy game. your water shader is very nice from a close view, but quite repetitive from a higher camera position, and experienced also some artifacts when adjusted fog and some other parameters, maybe clip_far, I don't remember. moreover, it would require a lot of time, and more shader knowledge, to implement a really great one, and I'm not sure it does worth it...
Working on an interactive perlin noise shader (it's ready only interactivity is missing), voronoi shader, rigid noise shader for fast procedural heightmap generation.
yes I did. I used the vertex shader for a fast terrain preview. I used directly the noise shader algorithm to set the height. When you shift the positions inbetween 0 to 1 the result matches the preview on the bitmap and the result is more accurate then a texture lookup.
Oh nice. But i think you should do it on the CPU because DX9 has no support for transform feedback (reading the vertex buffer back to the CPU) to generate collision heightmaps.
Yes, you are right. I planed it to be used only as a preview technique. The calculation of the normals in my shader is very expensive and eats the framerate for breakfast.
The problem with cpu deformation is, that I need to use cpu bmap lookup. There I loose the speed I gained by generating the heightmaps with a shader.
Perhaps the terrain preview will only be semi realtime, when you release the left mousebutton to set the value, not during every frame during value changes.
i think for preview it's perfectly okay to use the shader directly, but for cpu feedback, texture lookups are pretty fast (only copying the texture back is expensive)
Here my heightmap shader with realtime deformation of a preview terrain. The normals are a bit off, because I calculated them inside the pixelshader with the help of the heightmap. This runs with maximum fps. It's textured through a simple triplanar texturing shader.
@jenGs: that's really cool. Looks really fast to me. This realtime preview is much better than the one I implemented in my terrain editor. Is the voronoi algorithm implemented in the version showed in the video?
Yes it is implemented, but there is no gui yet. No way to create the modules interactively yet. I have implemented Voronoi, Perlin, Ridged and Billow modules. Now I am working on combiner modules like add, substract, divide, heighest, lowest ...
Wow, looks awesome! Really nice terrain effects, looking forward for combined versions. Just some idea: Combine the modules via bmap_process and high resolution floating point textures. This will prevent most of the information loss...
I was a bit bored by texture coverage creation for terrains, so made an auto multi-texturing shader, using a low and high level grass, cliff for high slope, shore sand, an underwater texture, and 2 others which can be custom painted using texture alpha (e.g. for roads). the fun thing is that it follows terrain deformation.
has been a long time since I posted something. I just wanted to let you know that I stopped working with 3D Gamestudio for some time. Nevertheless I'm working on a big thing right now.
I just finished working on my very own Game-Engine and Level-Editor.
Im working on a horror game. It must be coming along because I've lost count of how many times I've scared the crap out of myself. Mind the resolution. I run it in 600x400 because my computer suuuuucks.
Greasy grass
Before you find the book of the damned
After..
House on haunted hill homage
The blood on her - is not her blood
It's not quite a slender rip off. It does have linear game play, but as you progress, it gets more and more random. No flash light. You have a lighter. You have to keep flicking it until it lights, and if you turn too fast it goes out. There is a blinking system in effect. It is unforgiving.
Quick answer 1: Yes, naked. Whole game. Quick answer 2: Because your going to want to look and you'll immediately regret it.
I'm working on sky enhancement for R.A.G.E. Project. Look at these two screenshots and tell me what do you think P.S.: No shaders used,just scripting Nice morning
I thought it was about time to show you what Kartoffel and I have cooking. It’s a classical side scrolling shoot’em up called Rigid Force Alpha, where you steer your advanced space fighter to strike back a whole armada of enemy units.
The soundtrack people are Lauri Turjansalo (music in this video) and Michael Chait.
Looks pretty amazing and cool, good work! A few things: - The beginning (around 0:16) is a little dark, I think. - The collectable green-black-grey things look a little weird IMO, I think I'd add some transparent and subtle but relatively bright greenish glow. - The explosion on bigger enemies (see 1:09) is too lackluster, there should definitely be some chunks and sparks flying away and black smoke. - Same goes for the green blobs, they should explode into small blobs (which then not fade out by alpha but by size). - Your indieDB-link gives me a page not found error.
It looks like a lot of work already went into the project, I'm interested to see how it will turn out!
Great job guys! I'am very enjoyed watching the video. Every part I've seen looks very good and polished, so there is nothing to add. This is yet another game that represents Acknex engine in the best way, and i congratulate you for that. Is there an option for two players, and how many levels will be in game?
Turrican can't work on the project for the next ~2 weeks and the video was recorded at the last minute, sorry for that
The levels / gameplay shown in the video aren't finished. It's a bit chaotic, some things are already polished but others are still wip, subject to change, placeholders (, buggy) and this kind of stuff. However, we thought that we really should have uploaded a video by now, mainly showing the gameplay and levels we already have. Quality and framerate of the video aren't that good but I guess it's good enough to get a first impression from.
The beginning (around 0:16) is a little dark, I think. Yeah, it's quite dark and with the not-that-good video quality it looks pretty blurry but in the actual game it's okay. The HDR-settings aren't done, yet. We might change it but there's still a lot to do
The collectable green-black-grey things look a little weird IMO, I think I'd add some transparent and subtle but relatively bright greenish glow. That's a little embarrassing but to be honest: I have no clue what these things are good for. They were already in the game when I started working on the Project. But it's a good idea, let's see what Turrican thinks when he's back ;P
The explosion on bigger enemies (see 1:09) is too lackluster, there should definitely be some chunks and sparks flying away and black smoke. Same goes for the green blobs, they should explode into small blobs (which then not fade out by alpha but by size). At the moment we've just got 2 different animated sprites for exposions. They're okay for smaller enemies but we'll create more complex animations for others. Our main focus at the moment is the level design and enemy-ai. ( ...and yeah, there couldn't be someting less convenient than a red explosion for these green-blob-things )
again, thanks for the input
Edit @Iglarion: great to see that you like it. We haven't thought about a two-player option, yet, but it's worth considering, thanks. There are 8 levels/worlds with 3 stages each and an infinite-mode planned.
Do we have several weapons as a player what we can switch too? Mostly with these games you can carry only 1 weapon which I think is a shame cause it deprives it of weapon switching tactics and becomes a bit repetive.
I'm not sure what Turrican planned but I think in the final game you can carry multiple (but not all) weapons at the same time. You'll also be able to upgrade weapons (different weapon 'levels' already exist) but we don't have something finished, yet.
@Kartoffel Your game reminds me of "Space Impact: Kappa Base"! You could definitely use some feature from that old N-Gage 2.0 Game, since N-Gage was canceled. Look up for some Gameplay footage.
@Turrican/Kartoffel: great job, especially as an early stage video! the whole thing seems to be very well planned/made, even the music is perfect for this style! is there any story?
Thanks for the great feedback! I think we'll post information about our game more frequently in the future. (And there's a demo planned, we're just not sure about the content)
...now something totally different: I recently got very interested in pixel-art games so I started playing around a bit. and who could have guessed it, the first thing I did was writing shaders:
click image for full res (..yes, I know that the bloom is way to strong, but I noticed that after uploading the screenshot ._.)
joking aside, the first thing I actually did was thinking about how I can correctly render pixel-art in Acknex and implementing it. The theory is qute simple: I set the camera's ISOMETRIC flag and use its view.top/bottom/left/right values to render a specific amount of quants horizontally and vertically. I also changed the camera's angle so that the x-axis is now horizontal and the y-axis is vertical. Furthermore I chose the simplest scale ratio: 1 quant = 1 pixel - the same scale is also applied when loading sprites. After the camera's setup I applied a simple shader which scales the image up to the window's resolution without any filtering.
Now, in the screenshot above you can see my first try on combining pixel art with somewhat 'modern' postprocessing. It features a simple multilayer bloom rendered at the low-res version of the image to maintain the 'pixel-look' (the bloom is going to be less strong in the future, the high intensity was meant for testing purposes), tonemapping which also looks good with non-hdr input, a temporal video grain effect to add some sub-pixel-structure and a slight 1px blur to make the image a little smoother.
Sadly, I can't draw pixel-art, yet, so there's nothing more than these SPHERE_MDLs for now. However, learing more about pixel-art and practising are going to be the next things I'll focus on
I think either the approximation needs work or you need to have bigger pixels, because it kinda just looks like a downscaled version of a screenshot with severe lacking AA
I honestly don't think that that's a problem. Like I've said, the screenshot I've postet isn't really pixel art, it's just a postprocessing and rendering test for pixel art with some lit spheres.
This is what it looks like with an actual pixel art image (which I obviously didn't create):
regarding the pixel size: The downscale version in the forum here has too small pixels, yes but at the actual size it has 4x4 pixels and I think this is enough :S ...
Personally I don't find deliberately pixelated graphics too appealing, esp. now that there's tens of thousands of indie games that utilize that visual approach.
Even though I agree with you, I still enjoy this style. Just like a lot of current games are first person shooters, I don't think that it's a problem if numerous games share a specific style or gameplay. Of course, like you've said, there are a lot of games that use this visual approach so just having a pixel-art-style doesn't necessarily make a game unique anymore. You need something like unique game mechanics / ideas or a really good gameplay to make your game stick out.
Well, I guess it's not a bad choice to try it out at least... Anyway, thanks for the reply.
The shaders I wrote aren't supposed to turn an ingame scene into a pixelated-with-no-antialiasing image. The screenshot didn't look very different before. They were meant to support lowres pixel-art by using postprocessing shaders which were specifically created for this style.
The shaders I wrote aren't supposed to turn an ingame scene into a pixelated-with-no-antialiasing image. The screenshot didn't look very different before. They were meant to support lowres pixel-art by using postprocessing shaders which were specifically created for this style.
, yes I know, no misconception. Maybe bad wording on my part . I quess with pixel art I personally think more about extreme pixelated games. Maybe not the best example but oh well;
oh, okay. I thought you misunderstood my post, sorry But thanks!
( I guess I'm the only one in here who actually enjoys this style >.> )
Edit: Looking at the first screenshot you've posted I actually don't like this extremely low res style either. I like more detailed ones, but still pixelated:
( I guess I'm the only one in here who actually enjoys this style >.> )
Nope, I also really enjoy this Style and actually like the fact that there are thousands of games that utilize that style. I'm also a fan of low poly graphics like on the Nintendo 64 or PlayStation 1.
I like the ground, the detail in it. The horror woman looks good too, except her left arm (for us the right), looks weird (but could be that is intentional).
I upped the gloss factor because it is raining in the woods. I just realized that's not happening so I screwed up some code somewhere. The chicks arms could just be funny looking due to my ability to blame animating in med.
I updated the hair, made her less shiny. Also took a high resolution shot of the tunnel through hell. I gotta stop dicking around tweaking things and actually focus on the game.
I think it would be nice if we could see her eyes, on the 2nd pic she looks like a mad emo. Oh and she is one of the key characters in the game right?, so she deserves some love .
To answer your other question, I personally think while eyes would be the best (most creepy, ghostly). Fully black eyes could maybe also look great though if there is enough light .
The hair is slightly emo, but i can't get it to that flippy thing. I tried to make it look wet, and honestly I think having the eyes covered from the part in the bangs really adds to the whole don't look at her thing. I tried raiden style glowing white eyes, it was far too raiden...ey. Black eyes, look a little alien. I think i'm going to stick with the milky-greyish white. Really screams "im dead". I also just programmed a "page reading" thing, when you find pages you can hold em up to read holding F.
I've just implemented a (hexagonal) DOF effect into our game :
(as always, click to view in full resolution, you can't really see the effect in the downscaled picture)
It produces some artifacts, esp. with forward rendering, but the result is not bad in my opinion. It's going to be an optional effect (some ppl don't like it...)
These are by the way our max settings, I think it's good enough. The only thing that is still missing is shadowmapping for the sun - but in that shot it's not needed.
Indeed it looks fantastic (the whole shot)! I'm normally an enemy of DoF because it is either badly implemented or does not have any right to exist at all in that game (because it limits gameplay and visibility, esp. in 1st person games) but here I would definitely want to keep it activated. Only the red health bar looks pretty cheap and the ring around the ship (forcefield?) looks a little lackluster, too.
@Stansmedia: The bone tunnel looks pretty cool, esp. the first shot, I like it.
@Superku: Some time ago I already wanted to implement DoF but I thought it wouldn't look good. However, I saw it some weeks ago being used in a side-scrolling game and I'm glad that I've changed my opinion. It doesn't look bad and might even help you to focus more on the important part: your ship and the enemies... so it doesn't really distract you like in a lot of first person games which use DoF (like you've said).
The ring around the ship is actually animated but still very simple, yes (maybe we'll make something more detailed in the future). About the gui: This is even older than that 'ring' (...as you can see in this ancient screenshot)
I'm sure it's going to change, but thanks for pointing that out
I've just implemented a (hexagonal) DOF effect into our game :
(as always, click to view in full resolution, you can't really see the effect in the downscaled picture)
The DOF effect looks good. The only thing, I do not like in this scene, are the bright lights in the background. They do not look blurry enough. The border of the lights in the foreground (left side of the screenshot) looks much smoother than these bright lights in the center of the screenshot.
@Harry Potter I'm not sure what do you mean? The hexagonal blur causes hard edges, but that's exactly what it should do...
I think it's not a problem of the DOF shader. It's more a problem of the level design (maybe caused by different resolutions of the light textures/sprites?! Or maybe an additional blur shader is used for the lights number 1, and is not used for the light number 2?!).
The problem is, that the light on picture number 1 looks more blurry than the light on picture number 2. But the light on picture number 2 is far away, and should be as blurry as the lights on picture number 3 and 4.
The problem is, that the light on picture number 1 looks more blurry than the light on picture number 2. But the light on picture number 2 is far away, and should be as blurry as the lights on picture number 3 and 4.
, agreed, though I personally don't mind as it is now.
@Harry Potter The light in your picture#2 is extremely bright compared to the others. (through the bloom and dof you can't even see the actual shape of the light as it is in the texture)
The bokeh-dof and hdr rendering in the game does exactly the same, so I honestly don't see that this is a problem...
Maybe it's just because it's a still image, I dont know, but it seems perfectly normal to me. (you can see the light's brightness better in the actual game)
I see what Harry Potter means, but I don't see anything wrong in the picture - as Kartoffel says, the lights in question are obviously very bright, and I think that the image reads very well in that regard. It'd look wrong if the light was as blurred as the referenced lights that are not as bright.
Maybe it looks better, when the lights in the foreground (lights number 1) are not blurred so much. Because it looks confusing to me when an object (or light) in the foreground is blurrier than an object in the background.
I believe this impression is very much due to the nature of the Bokeh. As far as I can see, the scene is rendered with HDR and a nice bloom shader, resulting in all lights to have blurred bleeding. As the Bokeh filter is applied, the lights in the background get much sharper edges due to their brightness, and therefore may seem less blurred than the lights in the foreground.
But this always happens with a Bokeh filter afaik; Consider this reference picture: http://media.pcgamer.com/files/2011/03/Unreal-Engine-3-Bokeh-depth-of-field.jpg There is a light (or multiple lights) straight above the cigarette with a very sharp edge, much sharper than for example the purple-ish light in the middle of the mostright third of the picture, even though the latter is a tiny bit closer to the camera than the former.
Kartoffel, this looks really great man, we waiting for new video . Stansmedia, this 3 shots looks great (especially the young lady in the bath ), only i think you need corect just a bit color of light. Looks like yellow is to strong in first two pictures, and this looks a bit strange. Just my personal taste, maybe i wrong, but i think there need more tone of "white" light, to make the whole thing looks more natural. But all in all, great work guys!
It was only recently I realised I picked the same one he has lol.
, you can't really ignore to use the lasertrain as an avatar once you see it . Except if you have a dreadnought from Homeworld as an avatar that is ofcourse.
Quote:
If there is anything in his office it's a photoshoped poster of the wolf of wall street.
@rvL_eXile, nice texture. Now better lightning and you'll get quite a beauty .
Very cool exile. Making that for your own project? It looks awfully perfect for a decrepit cabin in the woods haunted by a blood thirsty demon woman...
Soo... Going to be available to purchase for a fair price?
I've been working on a theme song/title screen song for my game. If anybody wants to listen to it (it's a metal song) and has some decent speakers please let me know how the levels are. I've been mixing on headphones because I am too cheap to get some studio monitors. Too much bass? Nasaly? Highs peircing? Please let me know
yeah I just succeeded in creating a relative fast workflow of importing makehuman models into gamestudio.
I know, the models it produces don't look very good, but I am not able to model any humanlike figures, so I am very happy with the results.
makehuman->import into blender with the makehuman blendertool->apply a bvh motion file with makeWalk->export as collada->import into med->assign textures->ready
I don't know why the textures arn't imported into blender, but the UVs are, so no problem in med.
I just improved my skin shading. The code is still dirty as hell because I focused on making the skin look good, not the code. Also my gpu heats up quite fast when testing it, but that's again caused by the dirty and careless coding. However, I'm pretty happy with the result, it looks now muuch better, especially on high lighting angles:
Click to reveal.. (propably-not-very-safe-for-work)
haha that bathtube scene. I must say, I am normally not really a fan of most nowadays horror games (I don't like the cheesy quick scares), but definitely going to try that one out Stansmedia.
Thank you. Here's another shot showing the actual difference between skin rendering and normal lighting (in an extreme case):
Note that the skin texture is optimized for this kind of rendering, so you could get a better look with normal lighting. But you still wouldn't achieve a 'fleshy'/realistic look like on the right side.
Same here, though with more ugly clouds. Stars and clouds are by MasterQ32. I added a colourable sky and the aurora. Not that bad for my first shader experiments. =P
Added better reflective mapping and also emissive mapping:
@FEL: I'm trying to implement a simple shader framework without any bloat. Just some materials and a simple post processing chain, both usable without each other.
EDIT: Added tonemapping and automatic brightness adjustment:
EDIT2: Added bloom and improved everything: EDIT3: Added again more improved bloom and everything:
It is Tesselation (Look at the sides, the surface is not flat)
no, this is also possible with pom with a little math. Crysis 3 uses this technique because it is good enough while it's still much faster than tesselation.
"The more you know... Nice technique, but only has "into" the surface, not "out" of the surface offsets" And thats the quality, Cryengine can do Out and Into in same Frame. Now get back to the Story, DirectX11 for A8!!!!
"The more you know... Nice technique, but only has "into" the surface, not "out" of the surface offsets" And thats the quality, Cryengine can do Out and Into in same Frame. Now get back to the Story, DirectX11 for A8!!!!
Exactly. Well, actually it is still just inwards, but they're extruding the mesh along it's normals before performing the pom.
Of course, it's not as good as tesselation when used for very big deformations, but it is waay faster for small Details.
Why shouldn't it support the c_* functions? Those are implement in gamestudio and thus available.
I'm only replacing the render part with material events and two hooks for clearing and presenting the backbuffer.
Speed and performance must be testet but should be equal to the speed of gamestudio - DirectX 9 Render Time + DirectX 11 Render Time. Only thing that is slower is buffer updating (animation) as this is currently done twice the time (original buffer and then my DirectX 11 buffer)
Oh okay, I thought you started writing your own renderer from scratch.
No, i don't want to change the API of gamestudio as much as it is possible. Right now i have the problem that DirectX 9 has a weird mesh system (subsets with IDs, i don't know how to get the texture from the subset id...)
I am working on an obstacle course combat racing game. No vehicles, you play as monsters. Each monster has a different platforming special ability (flight, wall running, burrowing, etc). There are no item pick ups apart from cash and energy refills during races, because I didn't want to include the usual arsenal of pick-up weapon items that every character can use. Instead, every monster has his own unique set of offensive and defensive abilities, but you can only go into a race with 2 offensive, 1 defensive and 1 passive ability, so you have to pick before starting a race or tournament. The game has a Cel-Shaded art style and the character designs are similar to Digimon. I don't have a final name yet but for now, the working name is Project Monkey. I might eventually call the game "Oya, go!" Here's the steam concept page with screenshots, concept art and an early trailer http://steamcommunity.com/sharedfiles/filedetails/?id=281213043
Started playing around with pixel art again and fine tuned the hdr rendering and other shading techniques
The scene in the screenshot was quickly thrown together, so it's not even remotely 'good', I know that. But it's main purpose is to test the hdr effect and that works pretty nice.
(click for full resolution)
btw. the flames are actually animated (using a turbulence shader) which looks pretty good The next thing will be image-based lights to get flexible lighting.
shade-c is using a simpler bloom blur (shade-c uses 1 layer, I'm using 4 because the low resolution allows this, while being still really fast and with low memory usage) and most of all, no high dynamic range rendering
thanks! but the theory behind it is actually really simple, it's just a lot of fine adjusting to get the maximum quality out of it (= in this case blur radius and hdr quality) without any visible artifacts and using color modifications until it looks nice to you
Some demo will be there for sure, maybe doing an upload later or so. But first i'll check out wmb map support so i have this built in before releasing, for you to play around
@FEL: this would require rather big changes, sorry
but bloom is not everything, and I must admit that I usually over-use this effect and shouldn't care that much about it. It's better to focus on gampley
Okay, i'll do one. But you will be disappointed i think. It just looks exactly like gamestudio, level blocks excluded (they render with 0.5 fps, why so ever) And well, lots of stuff is missing (no visible debug panel or input console), only a default shader and nothing special at all
I'm SO sorry, but AckNext won't support any kind of level blocks. Gamestudio does some ultra advanced hackery internally and i coundn't figure out how the meshes are built and how they are textured and so on. But you can convert your levels to mdl7 and be happy:
uber pro tip: Having good displacement in a game needs a shitload of work to keep the quality high and consistent, which is hard for indie devs, even experienced ones. Also Acknex isn't famouse for it's fast rendering.
That's what I meant but I was too lazy to write everything down.
And by the way, pom with silhoulette correction is still faster, esp. for high frequency detail ;(
Edit: ...not trying to be offensive or anything. Oh and my grumpy attitude doesn't mean that I wouldn't be pleased to to see tesselation working in A8.
Yeah i know the $$m Files. But i think it's better NOT to support them as they are depracated. So only support for "modelled" WMB levels, should be enough to build great games... Later i'm going to improve the shader/material system so gamestudio materials will "just" apply onto AckNext rendering
EDIT: First custom shader (hosted in the original gamestudio.fx file):
Well, I render the scene on a low resolution with high dynamic range rendering. The lights are sprites aswell that are rendered before the level. These lights are written into a seperate image so that I can access the per-pixel light value when rendering the actual level.
After that apply a multi-layer bloom effect and in a final shader I scale everything to the actual resolution and add some nice effects (tonemapping, noise, etc.)
Well sadly, IndieDB/ModDB takes goddamn ages to authorize new posts (they can't be viewed without authorization) and thus no one really sees our game on these sites ._.
It was better some time ago where we had ~1000 views the day the news were added but by now the numbers of views have dropped a lot.
bmap-fonts with custom per-character width ...unfortunately, every character's width needs to be hardcoded, but after that it works perfectly.
edit: Seems like generating the text is reeaally slow. This only needs to be done when the text changes but in the example this just barely works within a single frame (at 60 fps). So if there's a lot of changing text, it could cause fps drops
=> damn bmap_blitpart, looks like I'll have to write a shader for that.
Well, I don't know what this function exactly does, so I don't know under what circumstances I have to use it and everything... sorry (It's probably not even going to work for my case, although it would be pretty cool because I could color the text much easier and also because of the performance, of course)
Renderziel 0 kann nur dann gesetzt werden, wenn der View kein eigenes Renderziel hat (view.bmap) und läßt sich durch bmap_rendertarget(NULL,0,0) zur Wiederaufnahme des normalen Bildschirmrenderings abschalten.
What are you doing with your view? Any pp shader stuff running on it?
Err I'm not even using views atm. I just coded some functions which makes an bmap out of a string using bmap_blitpart with a bmap font that also takes care of the character width.
...and bmap_blitpart is annoyingly slow so I'm looking for another method to write the characters on that bmap.
I have written a shader to do exactly what you want. BUT, your font and the target bitmap has to have dimensions that are a power of two. I process the Bitmaps with bmap_process. Until now it only supports letters with equal width. But you could input the different sizes with an additional bitmap that provides the information.
but what I actually want to do is writing the characters ONCE onto the bmap, not every single frame... But since draw_quad is quite fast I might do it this way.
edit:
Quote:
@Kartoffel
I have written a shader to do exactly what you want. BUT, your font and the target bitmap has to have dimensions that are a power of two. I process the Bitmaps with bmap_process. Until now it only supports letters with equal width. But you could input the different sizes with an additional bitmap that provides the information.
nice! I actually know how to transform the texture coordinates so that image sizes which are not a power of two are possible aswell.
That's strange. In my demo it runs with 60 frames in fullscreen. You only need to call bmap_process once, to copy one line of text. It is not needed to call it every frame. perhaps it is another function.
Ok, my code need 2 frames to paint a word. because of the wait(1) between the bitmap writing and the shader processing.
Well, then there's a little misconception, I guess...
In my example I'm doing exactly the same (just by using bmap_blitpart), but if I want to change the text every frame (like when typing in some text) it needs to be done a lot, so I looked for something faster.
I'll probably use a totally different methot when this is the case..
Using the way Felix proposed, with bmap_rendertarget should work perfectly for you!? Nobody forces you to redraw the bmap every frame. Just do it once and you are done. And if you want to draw additional letters into it there should even be some way to keep the textures content and render on top of it.
Using the way Felix proposed, with bmap_rendertarget should work perfectly for you!? Nobody forces you to redraw the bmap every frame. Just do it once and you are done. And if you want to draw additional letters into it there should even be some way to keep the textures content and render on top of it.
Edit: also why aren´t you using true type fonts?
but hoow? ;_;
In Felix's example he draws it once every frame which works fine, but if I only draw it once it's not working.
might be obvious but I'm really not getting it or it's just some dumb mistake, I don't know
Edit: I'm not using tt font's because in my case (low-/mediumres game) it just works better this way and I'm quite sure you can't get the same effect/accuracy/whatever with truetype fonts. (like the one in the screenshot I posted some pages before)
Edit#2: forget all that, acknex is just doing some annoying s*** again. After the startup I have to wait two frames instead of one before I can write the text, otherwise nothing happens. -.-
If you only add one character per keypress you can send the position data through a material skill. That saves a lot of time. You don't need a wait(1) anymore
Edit: I'm not using tt font's because in my case (low-/mediumres game) it just works better this way and I'm quite sure you can't get the same effect/accuracy/whatever with truetype fonts. (like the one in the screenshot I posted some pages before)
I am sure you can get that effect in a much better quality with true type fonts, BUT you need a font specifically designed for your needs, so either you find one online or build it yourself, however that is done.
Edit: I'm not using tt font's because in my case (low-/mediumres game) it just works better this way and I'm quite sure you can't get the same effect/accuracy/whatever with truetype fonts. (like the one in the screenshot I posted some pages before)
I am sure you can get that effect in a much better quality with true type fonts, BUT you need a font specifically designed for your needs, so either you find one online or build it yourself, however that is done.
Well I haven't seen a truetype font that used transparency so I don't think they support it.
Also I am very happy with the way my font looks like so ttf isn't going to make it look 'better' But thanks anyway
They do, there are different flavours! True Type Fonts are actually quite the amazing things, in reality they are byte code that is executed by a VM. And yes, that is turing complete.
However, the biggest advantage they have over bitmap fonts, aside being scalable, is that they contain information for kerning, ascender, descender and leading. Basically all the information needed to do proper typesetting of the text.
@^: Thank you for the info, I didn't knew that. However I think I'll go with MQ32's or jenGs' approach ( thanks you two ), especially because acknex's way of rendering truetype fonts is slow as hell (every character is a single drawcall afaik)
I just implemented colorgrading into my pixel-graphics renderer:
I'm using this to achieve a 'hue-shifted' look, which is a very often used technique to make pixel art look better. Hue-shifting means you shift the dark colors towards colder tones while bright colors get warmer tones.
Usually it's already applied when painting the textures but I'm using color grading for this in order to achieve a more unified look. Another benefit from this is that the hue-shifting now correctly reacts to the dynamic lighting that I'm using, which means that the look will stay consistent regardless of the lighting conditions.
----------
Moving on to another project, I've been also working on Rigid Force Alpha (which most of you probably already know). However, in the past days I didn't manage to make a lot of improvements on that one and Turrican did most of the work that happened over there, mainly speaking of the new user interface he created:
He also made an IndieDB post in which he explains some stuff about the interface in more detail.
I really thought about programming something similar but I decided to go with a less complex lighting model at first. And another reason why I'm using my current approch is my lazyness because I didn't want to create a model for every sprite to apply multiple textures.
Edit: now that I think of it, maybe ent_setskin works for that
Maybe I'll use something similar later, but at the moment it is more important to get used to the artstyle ( making lots of sprites ) and things like this.
We did very similar work for Barony when we implemented books so they didn't have to be manually formatted for page size in all the textfiles. Our books don't have color formatting like yours, though... maybe we should add features like that sometime!
uhh, these are really old and the code is pretty dirty I guess D: maybe I'll write some better ones in the future... the 'old' shaders don't even use proper hdrr afaik...
Edit: I just had a look at an older rendering pipeline of mine and noticed that the composition actually looks better than my newest stuff, even if I didn't care about keeping the renderer kind of 'physically correct' At least in terms of postprocessing (esp. hdrr, overpowered bloom and lensflares) and also lighting (supports ultra unefficient spotlight shadows!) I really like it's look more...
currently changing the windows (created 4 window "opened states"). After this ill create some smaller Stuff (one more lamp / plugsocket / switch / cutlery)
More furniture/household items. Couches, tv, bed, washing machine, etc. All grungy and old Do they have normal maps? And would you consider releasing the pack in two parts? Indoor and outdoor sets.
More furniture/household items. Couches, tv, bed, washing machine, etc. All grungy and old laugh Do they have normal maps? And would you consider releasing the pack in two parts? Indoor and outdoor sets.
Will be made, not all things you listened up but a lot Yes, Normalmaps will be included. No, the package will be released as one part.
my RTS game project is again under development (after editor reworks). for the 1st sight nothing changed, but thanks to the new tile pathfinder (allowing different character sizes and capabilities) and the optimized movement approach (much fewer repathing), a bit more visual quality can be allowed (including unit self shadows, non-limited camera view) beside similar, but more stable performance, on much larger maps (shot made with an 500x500 map). unfortunately total unit quantity currently cannot be more than 400-500...
It just took me some time to get it working, because I'm using 3d-coordinates for the level and lights and I had to swap and invert some coordinates. It was also a bit tricky to get the normalmapping working because sprites only support a single skin.
My inner pixel artist is back. I created a Mini-Game for YRPG Games. Play Cards to win a special currency. There are special vendors which offer items for jetons.
I don't know if the background will be in 3D or 2D.
EDIT: EDIT: Noticed a small mistake - I meant Royal Flush
Ich arbeite gearde an einer 2D-Filtertechnik, mit der ich Photos in eine Comic-Darstellung umwandeln kann. Ziel ist es, 1. innerhalb von 3 Minuten eine Vorlage zum Weiterarbeiten zu erstellen 2. Zeit zu sparen 3. es soll nicht sofort offensichtlich sein, dass ein Photo die Grundlage war
if you look at an image you don't just see some colors, but your mind creates a whole, three-dimensional image based on things like lighting and most of all also experience. f.e. if you've seen the object once before you know exactly the shape of it, or if you look at a photo of a room your mind knows what a room usually looks like (walls are perpendicular to the floor, etc.) and can create a whole image out of it, which is most likely correct.
This also happens to all objects in this photo. You immediately know what belongs to every object and what all these objects are, which is again hugely based on your experience and things like lighting and other stuff.
A computer, however, just sees some colors and cannot precieve the scene in the same way as you do.
I don't think so, it's just that in reality light would reflect off the stones and light up the areas that are shadowed. With a pitch black shadow it just looks a bit weird.
thats a point, but i dont want illumination. i like it how it is right now, but i still need your functions to work. U now see how important it is... :3
@^: no way this is just the filter. Computers cannot detect edges the same way as a person does.
Aber Computerprogramme können Kontraste finden - und hierbei setzt die Technik an. Btw... ich arbeite an einer Filtertechnik, nicht an einem Filter. Beim folgenden Bild ist z.B. der Aquarellfilter im Einsatz. Arbeitszeit ca 8 Minuten für das Bild - keine Striche zusätzlich mit dem Cintiq zugefügt. Nur eben die Kombi von Filtern.
Vielleicht noch als Anmerkung... die Technik soll gerade für solche gruseligen Darstellungen zum Einsatz kommen. Ruhigere Motive würden dabei völlig zerstört werden.
Yea :3 nice and fitted! Take your time Im going to create a fully rigged character with nice skin shader... oh that was yesterday... today Water is on the Plan
@Kartoffel, love that last screenshot with the blueish glowing stuff. I wonder how it'd look without the aliasing, but I also think that it adds to a retroish feeling.
when you hit enemies you build up a damage multiplier which will deflate when not attacking. when that multiplier gets above 1 you can make a special move and another when the multiplier hits 2.
rojart@very nice screens! is there any video to see all that in motion?
Denn15@I'm impressed by that gameplay video keep it up!
I'm working on a small level editor, which will allow to create,save and load 2d levels, plus I've created floodfill pathfinding which makes me proud of myself yet, I need to add some more features into it, that I have in mind, then I'll include it in my map editor.
@ Kartoffel: What kind of raymarching are you using? Distance fields? Because just a few days ago I integrated my own raymarching solution in Unity3d and it works really fine.
It is so slow because I'm rendering the scene, first reflection and second reflection, and all of these with normals, lighting and variable penumbra soft shadows.
I think it looks pretty cool! Have you tried to apply that to more complex geometry, like some (optionally curved) hallways?
@sivan: Looks interesting! As a side note, how do you handle the movement of a bunch of units close together, for example when they go into a medium sized choke? Do you move all of them at once and then check for overlapping entities or something like that?
it is not really handled currently in a special way, it works simply because it is a hierarchical system: just finding a path through it by targeting always a tile in the next cluster (area) of the group path, and simply ignoring other group members in the tile pathfinder because the collision resolving is effective enough (and the tile pathfinder has multiple error handling inside to approximate the target if invalid). it works fine e.g. at crossing a narrow gate by a couple of groups consisting of 20 units, but standing groups in the way can make it impossible... so in future: (1) groups must communicate with each other using a priority property, (2) formations should change their shape in these cases. (3) unit targeting is the other key thing that will be resolved by using the new formation system, by analysing available space and deforming desired unit positions according to this space (currently walking up to walls through narrow stairways cause some units getting wrong targets and separated from the group.) ...and of course groups should realize that they are in a bottleneck area, which is already done.
Later, later. Right now there isn't even "battle". You cannot hurt someone, just spend AP on actions doing nothing. Oh and you can change position as well as skip turns so you can swap between chars.
Haha thanks. Best stuff i ever draw! My C# wrapper Acknex.NET gets some serious progress through this. Right now i've implemented effect so you can use effects written in native Lite-C. But i'll work on a C# version for effect as well.
I decided to create a new tool for YRPG Toolkit which allows you to use all special effects at the same time. All values can be changed during the preview option. Within some seconds you can see the result of using more than one special effect. If you like the result you can save the special effects for this map. The new File Format is called *.hog and you can simply paste it into the YRPG Toolkit event editor. Just a rightclick and choose "Import HOG Event".
By the way - HOG is the abbrevation for the name of the new Tool. I decided to name it 'Hand of Goddess' because using special effects is so super simple now.
Yeah, it's not that i'm coding the wrapper for fun (only). I want to use Gamestudio with C# in a C#-like environment. The other Gamestudio C#-wrapper is complete but a pain to use. I'll (maybe) post some updates later in the wrapper thread as i've changed a lot of internal stuff...
Edit: @HellThunder: Schreib mir mal ne PM wegen HOG. Dein Postfach ist nur grade voll....
The above image is rendered using raymarching, which is extremely diffrent from normal rendering. You can do really cool things with it and you can render MUCH more realistically (you quickly exceed interactive framerates, though) but in general this way of rendering is not useable for pretty much all kinds of games.
There really are lots of reasons why you shouldn't/can't use raymarching for a game.
@^: Err no, sorry. (no one understands that all the raymarching stuff I'm doing is pretty much useless and just for fun >.>)
There are several reasons why you can't use raymarching the way you (and probably most ppl here) think:
- it's a totally different approach to render a scene, like really different. - everything is calculated using a shader, which basically means you hardcode your graphics (geometry, reflections and everything). you can't dynamically add or remove anything. Animations are very limited aswell. - you don't have actual geometry as you know it, you define your geometry with a (or several) mathematical funcion(s), which makes modelling complex objects almost impossible and can look something like this:
i have been working a little on my hack n slash game.
now the enemies arent just white boxes and a new enemy type which shoots projectiles that you can knock back by attacking it. a new special attack which can be seen on 1:06 is also now in the game.
Haha thats cool Emre, maybe you should try to put it on e.g. steam Greenlight when finished, I bet there are people who would like to have something like that .
@FEL: i've passed my game to some indie game sites and waiting for feedback. I have some free time and well... i'm sick of "stupid non-animated win7 desktop"
@Reconnoiter: Yeah. First indie desktop program ever. Must be on steam! joking apart, i agree with you about that there are people who would like to have something like that.
@Kartoffel: Performance is not so poor actually. it's same as browsers.
You guys download it from here if you wanna take a look.
@Kartoffel: Performance is not so poor actually. it's same as browsers.
I'm just saying that a 60fps application made with 3dgs (for example my music player) has WAAY too much cpu usage. Around 17% on my i7 - which is as much as it can get plus the usage for the fmod plugin.
It doesn't seem to be a bottleneck but even the simplest stuff takes up one full core on my 3.5ghz cpu which I don't understand. That's one of the reasons why I say: Acknex' performance is bad.
Emre, why are u working with gamestudio a6/a7? Wouldnt it be better to upgrade your code etc? U would just l0ose the Physics newton. But gamestudio a8 hast physX and we have a "rojart"
Because of financial issues and my plans. I've been waiting for some things and it seems all my plans fell through... but i will not give up of course. The only thing i want is sell my game. For now, A7 fine for me.
@rojart: How do you get a fully bone animated walking robot to work? Always when i try to make an modell do an animation i got sc_material_event Error...
Funny is: When i attach a animation action to an Entity everything brakes, when i unattach it and let the modell do its default anims (es the engine is porposed) everythig works fine?!?!?!!?
I tried 3 different scripts for animation. Nothing works. W T F ???
Workflow Improvement for my 3DGS RPG Maker. New tree-view for a games event-collection. You can create folders and put events inside. This eliminates the chaos where all events were shown in a single list. Mark your events as 'Critical', 'Work in Progress' or 'Tested'. The Main node of every event-collection will tell you if there is any unfinished work.
do you guys think the arm's skin color looks okay , i think the general color does not look to bad ,however i just threw a skin on it to try and home in on the correct color ?
i am walking around in a level like a zombie ,dont mind the blueish light,i am trying to see how lighting affects the color.
stil have to properly wrap the texture and other stuff.
i was thinking a first person style game ,since i haven't tried doing it yet ,the problem however is ideas ,and i had a few i think would be worth a shot for prototypes ,if one of them works out i would likely stick to it.
there is no real need to share the rigged models ,i mainly use exported makehuman models wich i then rebone in blender if there arent sufficiently enough bones ,the arms are just cut from a complete body and extra bones were added ,i did not use the original exported skin though ,actualy i should use it but as i havent decided yet i may add something to the arms/hands that make them more oriented towards a particular style
as there weren't any sword stuff related in my posts I assumed you ment the models , the code just places the physx character and camera and hands ,basic player movement , nothing special there at this point.
@rojart: How do you get a fully bone animated walking robot to work? Always when i try to make an modell do an animation i got sc_material_event Error...
Funny is: When i attach a animation action to an Entity everything brakes, when i unattach it and let the modell do its default anims (es the engine is porposed) everythig works fine?!?!?!!?
I tried 3 different scripts for animation. Nothing works. W T F ???
I'm continuously sucking with my new advanced formation system (in Oct-Nov), fortunately I found this great idea just in time: http://www.relgin.com/ffmm/ that apparently is a nice solution on my problems. my original wrong plans are replaced by its simplified version, and now the basic operations are working:
lighting looks good! too bad it's just static lightmapping.
thanks I like it too , the purple area seen through the windows are WED lightmap , the pinkish room is a different lightmapper , I cannot get WED to the same amount of quality .
static because it's "lighting fast" on my laptop :-)
it's this little shitty program called light map maker. but it's quality is quite good for the first free program i came upon search.
i was looking for gile(s) or something similar, apperantly giles went free or open source or something way back but i cannot find it anywhere.
edit. I think jcl should upgrade the lightmapper , if I can get good looking radiosity like that in gs plus be able to map a model I would definitely use it.
I have somewhere gile(s) downloaded but I had problems with exporting to 3dgs so I gave it up a long time ago. in 1st or 3rd person games UE4 is really great, and its development seems to be unstoppable in all aspects.
I have somewhere gile(s) downloaded but I had problems with exporting to 3dgs so I gave it up a long time ago. in 1st or 3rd person games UE4 is really great, and its development seems to be unstoppable in all aspects.
Thanks for you input Well, Im coding first person shooters so it should be just perfect in that case
my opinion on the "game studio" problem is that ,this engine can be out grown by developers , when you find yourself in need of doing advanced projects you soon realise ,that even though you know the engine well enough there isn't really enough from the engine side to back you up for your advanced project ,I mean the engine doesn't include the advanced functionality that bigger engines already include so you are left trying to code all the advanced features yourself ,which you could just simply spare yourself from by using a different engine .
problem about this is ,that many who outgrow the engine also love the engine ,so we mule around hoping and praying it wil catch up with your growth. .
I don't think the engine is going to outgrow us any time soon , so to me this engine has it uses as part of the first steps in to bigger things ,and well if you have the time and all the general knowledge it takes in different fields then you could most likely code your ass off and make a big project ,the simpler choice is stil to use a more advanced engine for more advanced projects and use gamestudio for projects suited for its maturity.
my opinion on the "game studio" problem is that ,this engine can be out grown by developers , when you find yourself in need of doing advanced projects you soon realise ,that even though you know the engine well enough there isn't really enough from the engine side to back you up for your advanced project ,I mean the engine doesn't include the advanced functionality that bigger engines already include so you are left trying to code all the advanced features yourself ,which you could just simply spare yourself from by using a different engine .
problem about this is ,that many who outgrow the engine also love the engine ,so we mule around hoping and praying it wil catch up with your growth. .
I don't think the engine is going to outgrow us any time soon , so to me this engine has it uses as part of the first steps in to bigger things ,and well if you have the time and all the general knowledge it takes in different fields then you could most likely code your ass off and make a big project ,the simpler choice is stil to use a more advanced engine for more advanced projects and use gamestudio for projects suited for its maturity.
^ this.
plus, acknex does a lot of weird stuff in the background and it's lacking low-level control.
I still like this , if you have the torches and lighting as part of a core concept ,you could really create that mysterious eerie exploration elements for your game , a part of the mechanics could be having to use torches to light up and explore the world ,if you added a sort of progression through the game then later on the player could start throwing light sticks in dark areas etc .
crackling sound of fire and crickets and wild animal sounds plus some scary monster sounds in a darkish nightime game , exploring underground caverns ..
nice idea but I'm looking for something really simple at first as some kind of practise and to get used to this.
I think all that games which need a lot of detailed assets to make a scene, or things that require a complex collision system (I don't have one, yet) are not gonna work... Doesn't mean it can't be somewhat fancy looking.. but I haven't got what it takes to make a detailed and rich environment. Something with a space-theme for example would be suited better... because there's a lot of 'space' instead of textured environments
anyway, thanks! I'm happy to get any kind of suggestions
things that require a complex collision system (I don't have one, yet) are not gonna work... Doesn't mean it can't be somewhat fancy looking.. but I haven't got what it takes to make a detailed and rich environment.
go for quad trees , for your map system and speeding up your collision testing..
most of the best selling 2d games aren't really highly complicated dense graphical scenes , the important thing is to have a good concept with good mechanics.
start with a simple prototype and drop it early if it doesn't feel good , if it feels good then you might just have something , most people aren't the type who would sit down and play something complicated or big ,they just want something to pass the time with and feel like they are able to play and understand it, this is why super mario and simple games like minecraft had wide spread success , no serious gamer looking for eye candy would bother with it but then again not everyone is seriously interested in technical eye candy .
you know what's the problem for me , as many ideas I have ,there is always a game that already did it in some form , which I hate because it makes me feel like I am not original but sadly thats just how the dice rolls , i mean even angry birds used the exact concept design of earlier games ,but it was definitely more successful as those others so I don't care anymore about using concepts that has been used before , now I just say screw it and do it .
I have noticed that ,even though concepts are being done over and over again ,the only thing you need to do is be better at the addictive elements of your game vs the rest , if people would rather waste their time on your game than the rest which uses same concepts then your going to make money ,addiction is key ,got to make people your game junkies.
Yeah, everyone knows that situation, I guess, even if you came up with the idea entirely yourself.
However I don't agree with the addiction part. You've got to make your games in a way that the people want to keep playing it, they shouldn't keep playing because they have to (I also think that it's unlikely they'll do this through addiction). But yeah games like CoD with all these reward things go more into to the addictive direction.. not sure it it's working, though.
However, you have make your Game fun and worthwile and people will play it. This can be done through interesting or fun gameplay/game mechanics/graphics and lots of other stuff. If you don't have an interesting aspect that wasn't there before, these are the the factors that will make your Game unique.
jk. But seriously, I'd really like to see it in motion if possible
Edit: some basic projectile types:
pulse*, laser, bullet, rocket* and lighting*. (* animated, looks much better in motion..)
But I'm currently having some problems with the displacement shader.. I somehow cannot set different shader values for objects that share the same texture
This 3d one's better. Though I would personally make it less obscuring (/it blocks the vision now). Perhaps scale it depending on the distance between it and the camera? Or maybe move it to a corner (e.g. left bottom).
I was thinking more of bending it on the other axis, i.e. bending the HUD in a circle around the player, that way you can also attach other hud elements and they should seem more "in place".
The panel should also not have a hard border! Makes it look way too "hard" IMHO, the look you should be going for is "floating hologram". Maybe even add a scanline/chroma-aberration shader to complete the look
@^&MQ32: I don't think borders look bad, I'm using them too and it doesn't look that bad imo..
(click for full size)
I'm also using multiple moving screen-lines and a moving rgb-distortion effect, you might wanna try that (if you're trying to archieve a hologram effect)
You're basically some 'thing' in the middle of the screen and have to defend yourself against enemies by typing specific words which get harder as you proceed. (I haven't thought about any kind of theme, yet... and it's most likely going to stay really simple)
Nothing unique, I know, but it still is fun.
I'm going to post some screenshots (or maybe a video) in the next couple of days.. altough at the moment I only have some code done.
nothing fancy and gameplay isn't fully implemented, yet.. but word difficulty and randomization works pretty good. I'm currently using 1000 random words generated by some website I don't remember.
@Reconnoiter Looks good! What shadows are you using? I wasnt able to use any sort of shadows for my game, all the shadows are buggy in some way and since Im not good with shaders I have no clue on how to fix it
@Ch40zzC0d3r, thanks! Here it is a combination of stencil shadows (which are somewhat slow but its not extremely bad cause most models are not that high poly) and those shadows for (WED) blocks for rocks/structures/vegatation. I personally don't find stencil shadows buggy by the way, maybe I am just lucky . In the video I am playing the game on a medicore laptop, so I am quite happy with the performance.
I actually wanted to use the custom/shader shadowmapping but I can only get it working with new projects, not with this one. Those are faster with still ok quality but the basic one that comes with the gs3d install only works with outdoor scenes iirc.
Than there is decal shadows which you can try for e.g. top-down games, but I am personally not really a fan of those.
Thanks, I guess Im gonna use no shadows for now.. The GPU bones make it even harder to get good shadows. Even PSSM looks like .. uhm you know Flickers alot and yeah.. I could use LOD shadows but it does not work with my project at all.
Cool a grenade launcher . The network coding looks solid too. My network coding still lack (those future predictions are hard to get right << ).
ps: maybe a dumb idea, but I saw you don't have many models in your map, just many blocks. Maybe could you set stencil shadows on but use low poly dummy models for player shadows.
Cool a grenade launcher . The network coding looks solid too. My network coding still lack (those future predictions are hard to get right << ).
ps: maybe a dumb idea, but I saw you don't have many models in your map, just many blocks. Maybe could you set stencil shadows on but use low poly dummy models for player shadows.
Thanks I was using ANet and made some modifications to the dll to build in some features. The map was a testing map only, but the shadows even bug arround with my muzzle flash, invisible models and so on. Its not the map itself I tried it several times, also got other maps nowadays. I would pay for a fast and goodlooking shadow solution because GS doesnt offer me any Well, I coded this game with anti-cheating in mind, most things are serversided or controlled by the server. Also coded a small anticheat to prevent cheating on the client itself. I will post a download for it later in 2015 when Im done with school Hope to see some servers open then
there is a little lovely editor for gamestudio that includes Slin's flickerless pssm shdows, which looks really cooool with some stencil blur. works but not perfect with 32b textures (vegetation). currently buggy with terrain lod. stencil shadow also can look fine with some stencil blur, but rendered fast enough for low poly models only. I mean they get very slow for higher poly ones. but with a custom shader set you could use transition from static lightmap to dynamic shadows, e.g. how it is solved in the above mentioned editor for terrain (I'm not sure I made it for map entities which would be similar for blocks...).
here's a little update on Rigid Force Alpha. Since my last post I rewrote major parts of the game code and also did some redesign of gameplay systems.
You can now charge your weapon's special function over three levels, depending on how many energy capsules you have collected (visualized by the horizontal bar in the lower screen center). Additionally, you can charge a combo bar by killing enemies. When maxed out, you can unleash a powerful smart bomb attack which instantly kills smaller enemies and bullets around your ship and inflicts heavy damage to larger opponents. This also triggers a combo mode for highscore players. Just two changes of many, but I don't want to bother you with too many details.
Apart from that, both of our composers now finished their tasks which leaves us with a great soundtrack of roughly 40 minutes length. It's a well-balanced mixture of rock, orchestral and electronic elements. And kudos also to Kartoffel, who has built some nice models and also made several improvements to the rendering pipeline. Without that, the shots below would not look half as good.
Speaking of those shots, I also worked on effects and level design/art which can be seen below - these are just some impressions from the Toxic Jungle stage, including the boss enemy, which I'm currently still working on.
It's all wip, but don't let that keep you from giving us feedback.
here's a little update on Rigid Force Alpha. Since my last post I rewrote major parts of the game code and also did some redesign of gameplay systems.
You can now charge your weapon's special function over three levels, depending on how many energy capsules you have collected (visualized by the horizontal bar in the lower screen center). Additionally, you can charge a combo bar by killing enemies. When maxed out, you can unleash a powerful smart bomb attack which instantly kills smaller enemies and bullets around your ship and inflicts heavy damage to larger opponents. This also triggers a combo mode for highscore players. Just two changes of many, but I don't want to bother you with too many details.
Apart from that, both of our composers now finished their tasks which leaves us with a great soundtrack of roughly 40 minutes length. It's a well-balanced mixture of rock, orchestral and electronic elements. And kudos also to Kartoffel, who has built some nice models and also made several improvements to the rendering pipeline. Without that, the shots below would not look half as good.
Speaking of those shots, I also worked on effects and level design/art which can be seen below - these are just some impressions from the Toxic Jungle stage, including the boss enemy, which I'm currently still working on.
It's all wip, but don't let that keep you from giving us feedback.
@Turrican, very good artwork and effects. The graphics are really good, only some shadowmapping/shadows missing but thats the only thing I can think off. And maybe some low res textures like of those caterpillar-like thingies, but I personally don't mind that.
@Reconnoiter: Thanks, that's good to hear! I'll sort that texture resolution bug out soon. Regarding the missing shadows - well, it's really a performance problem. I personally could live without them, but let's see if Kartoffel can perform a magic trick or two to get them implemented in the end.
I guess Kartoffels shader are damn fast on other engines. 3DGS is slowing things down as fuck. I couldnt even spawn 20 zombie models without any code and got alot of lag just because they had 2k polys and the textures were not 5kb -.- (they were TGA, it runs a bit better with DDS)
...instancing, a better subsurface-scattering than I'm using in the skin shader that I hacked together, shadowmapping with actual lightrays (lit fog in the air), global illumination using light propagation volumes (LPV) or even a voxel based approach, etc.
But it's not just the engine that causes problems.. it's also my knowledge and hardware limitations.
Very nice, especially the first two screenshots look amazing! Only the 4th shot really does not resonate with me, the rock structure and texture work looks subpar and so 2004. It looks like the neat mushroom background art and the foreground art changed places accidentally. The seam to the ground texture looks displeasing, the rock texture is too repetitious because of its defined bright and dark parts, the structure seems to lack some polygons here and there, there are no real details like stuff on the ground, small mushrooms or other stuff growing on the walls and so on. What you could do is try to create a subtle highlight on the borders because the light is coming from the back, then choose a less distinct texture, apply triplanar multitexturing and multiply it with detail and brightness maps or use some sin(inPos.x/256+inPos.y/32) brightness modification or something like that.
Why don't you create a stand-alone WIP thread in this Projects Forum and keep us updated there?
Semi-Fake Screen Space Reflections using directly computed UVs by assuming a horizontal mirror plane. Refractions approximated by beer-influenced Acknex 8 magic.
Long live Acknex, long live old school render tricks. Death to all the fancy render monkeys!
looks awesome by the way.. but real screenspace reflections shouldn't be a problem, even with A8
edit: after re-reading your post I noticed you're probably not using the rendering technique I thougt at first.. in that case it's perfectly fine, even recent games do it the way you described
I flip in the first render pass the figurine along the mirror plane and write the UV of the original pose in a RT... then I render the scene, writing out a mask for the mirror mesh and the approximated refraction vector from the difference between the normal and the mirror plane normal. In a PP combine pass, I lookup for each pixel of the mask the UV of where to lookup SSR.. then I apply an additional fade-out value, which I computed from a fresnel term (from the 1st pass) and the re-projected viewport borders + a fadeout on the distance to the mirror plane.
It's not as versatile as SSR, but it looks pixel perfect on polished planes and that is what I want I am happy!
I only had problems with real SSR.. I tried some approaches but I had only issue with artifacts... ... fuck it! - So I threw everything away and played some A4 stuff where we used to mirror models and make the floor half-transparent... that got me going.
I flip in the first render pass the figurine along the mirror plane and write the UV of the original pose in a RT... then I render the scene, writing out a mask for the mirror mesh and the approximated refraction vector from the difference between the normal and the mirror plane normal. In a PP combine pass, I lookup for each pixel of the mask the UV of where to lookup SSR.. then I apply an additional fade-out value, which I computed from a fresnel term (from the 1st pass) and the re-projected viewport borders + a fadeout on the distance to the mirror plane.
Short reply @Superku and @Reconnoiter: Thanks for the comments. Both of you are right and these are all good points. I've put them all on my list. The fourth shot is indeed a bit out of line and I'm going to work again on the scene sometime soon. Good idea with the backlight, I'll give that a try.
And yes, for the next update I'll create a new thread for RFA to prevent going OT too much in this thread.
Yes! Hooray for Good Burger summing up what we were all thinking!
Turrican! Your graphics are wonderful! Very pleasing to the eye! On that note, some very nice works have shown up here at the forums while ive been away! When I left the forum was dead!! Finally have something new to read haha.
Not game related but I have been working on this: http://www.pioneerspirit.ca Its a website for a local business and they specialize in bees wax.
I am wondering why a black person is used in the above pics to depict ignorance? Is this an unconscious result of tv and media manipulation, (they do it all the time), or is it consciously intended? Due to the history of black and white peoples, and the continuing subtle denigration of black peoples in the media and such, maybe a little thought is needed?
would it have been okay to be a white person ? or any race or color then ?
really ,if you have race issues, leave your race issues at home , everyone sees a human depicting ignorance or simply words , you however see color ? whats up with that ..
The image does in no way try to depict ignorance, it is a well known meme for use in exactly the context it has been used here: It literally says "well I do know some of the terms used in this text or statement, and I do maybe have a certain amount of knowledge in this topic, but to be honest the whole thing was pretty much over my head because it's way more in-depth than I currently am in this topic".
Also, it is originally taken from a movie, Good Burger (1997) and only one of many quotes from this movie. The fact that the depicted actor is black does not have any relation to the symbolic meaning of the meme.
End of random trivia, not trying to lead a discussion or something, just thought some might find this interesting. Also, sorry for off-topic, if that's even possible.
2: As a member of the black racial group, i encounter these things even today, mainly in the media which is done deliberately to maintain a sort of division, but also by those who are not aware of it, thereby becoming subconsciously affected by it.
3: Therefore i talk about it, hoping that others will see the deliberate attempts of certain groups of people in power to keep the groups apart, and nullify their attempts.
@pararealist: Please stop this discussion here for now. If you want to continue it, open a thread in Morbius (which I don't find to be necessary as there are no races, only good and bad people, but this is yours to decide).
... as there are no races, only good and bad people, but this is yours to decide.
That's right. Today there is ONLY ONE human race living on this planet: the Homo Sapiens.
In the Stone Age there were several human races, but only Homo Sapiens has survived to this day. All people living today have the same origin: East Africa.
I'm cool with that. But i still do not get what makes any one think i see multiple races when i particularly stated "black racial group", group being a part of a whole. Anyway, it seems that race is a forbidden subject, or there is some kind of hangup around it, or taboo!!!
I'm cool with that. But i still do not get what makes any one think i see multiple races when i particularly stated "black racial group", group being a part of a whole. Anyway, it seems that race is a forbidden subject, or there is some kind of hangup around it, or taboo!!!
Not so much any of these things as it is just a completely unwarranted reaction, and not to mention very off-topic.
The original algorithm is pretty simple: an 'ant' moves on a square grid. Every cell on the grid is white by default. Now if the ant enters a white cell it turns left and colors the cell black. If it reaches a black cell it turns right and colors the cell white again. The interesting thing about this is that you'd expect some kind of repetetive or simple pattern from rules like these but the result is pretty chaotic.
You can also extend this algorithm with more cell states than the default black/white and different movement directions resulting in new behaviours. Some appear to be totally chaotic while others are perfectly symmetrical, create crystal-like structures or have distinctive patterns for example.
In my implementation I can choose up to 32 cell states with the directions left, right, forward and back.
I read shorty about it, that it can simulate emergent behaviour, but still don't know how it could be useful in a game. maybe simulating some crazy enemies?
terrain texturing is easy (finally added a simple menu to next MapBuilder): (the terrains shader calculates underwate/shore/grass/snow/cliff based on 6 parameters, and supports 2 textures with hand painted coverage mask.)
Holy shit, that is awesome! I only have a few real complaints: first, many of the textures seem to lack a significant amount of detail, the bloom on some of the objects (ie the beer bottle) is too overpowering, and the FPS is obviously quite low. But these are mostly minor issues, I'm otherwise very impressed.
Sidenote: if you don't like using FRAPS to record your videos, you could check out Open Broadcaster Software. I've used that for months to record footage for Barony and it works great, I'd recommend it to anybody.
Also, you got Cinemassacre themselves to respond to your video! That is just sick!
Thanks for your kind words! about your complaints: Lack of Details on Textures: Well, not every surface in real life is breathtaking isnt it?! Bloom on bottle: It is planned that the beer bottle will be an item later in the game, thus the highlighted apperance of the bottle (later it will pulsate) Low FPS: My Laptop isnt that powerful, but Iam able to keep 40fps without Fraps running at the same time. Thanks for the tip on Opfen Broadcaster Software, I will check it out!
About Cinemassacre respond: Yes, I couldn't believe it myself at first!
I have actually quite good results with Fraps, but you should have a hdd/ssd that is fast enough for capturing the video stream (FullHD@60FPS Raw takes quite a lot space...) But yeah, OBS is also pretty good, it renders the videos into MP4 in realtime thus requires a more strong CPU than fraps.
NVIDIA ShadowPlay is by far the best capturing "program" Its actually a built in driver function and doesnt limit the FPS by any degree. It also captures in mp4 and avi quality.
I always loved to hack around on the limits of 3dgs. Today I decided to play around with litec again after a long time. After playing around a bit, I found a way to call struct methods and pass 'this' to it(without actually passing it...). Difficult to explain, here is an example
it needs iron sights ,ammo clip , ammo clip release button , hammer , the hand grip/bud area needs to be shaped like a palm and thumb can easily fit on it , it needs the brass ejection thing aswell ..
did use c_trace before 2.5 to place the entities? ^^
in my editor I basically convert 2d mouse coordinates to 3d, set the z axis (most often 0), and use a direction vector (or whatever it is called) to get the right x and y coordinates. works like a charm
ps: you should outline the object info text of the selected object, its kindy hard to read
I used different c-trace based placement onto terrain, or terrain+terrain models, or terrain+terrain models+buildings etc. what is post adjusted according to should it snap to tile xy or z or not. It is different for each entity category (4 kinds of foliage, 2 kinds of building, and terrain models).
Now I use Carlos3DGS's contribution to determine mouse pointer's view line intersecting the zero level plane if c-tracing fails. Some adjustment is still needed but works fine.
And thanks, yes, that temporary info text looks badly sometimes, but it is worthy to be kept, I will add a black rectangle or something behind it...
nothing special just a grid showing level_ent size - of course, its size is auto updated on entity placement outside actual boundaries. a bit ugly on far distances but good enough for us that green church is just a preview following mouse position, thus you can see exactly where it will be after a click. I'm happy with everything, but too lazy to make the ground and water terrain placement menus... currently the only option is that ugly startup wizard, so it should be made anyway...
@Sivan, but don't you think it is bit of a waste of your time to make an editor for gs if you are going to migrate to ue4 probably soon? (excluding the fact of honing your programming skills ofcourse) Or does ue4 not have good mapping tools (yet) and your going to translate it too ue4?
for the first sight it is waste of time, but it has given me a lot of experience how to and how not to do things, my work is much more organized now (still not enough). and after a long period without update, I decided to release a probably final version with a good enough source, then I can concentrate on my game again. of course I will add bug-fixes and minor features later.
since I develop my game within MapBuilder, a better editor is better for me too it has a simply separated code package not included in the released version. the engine itself is suitable for making my game even if no more updates come, I stay here for a while because I can work fast with it, but ue4 seems to be the right choice of future works for me. it is a heavy but very inspiring engine, being developed with light speed. e.g. visually debugging game/AI logic required a lot of work with 3dgs, and it is much simple in ue4; complex outdoor scene rendering also more optimized etc. but it depends on your needs/targets, any engine can work
, lol , don't get me wrong though, I like gs3d and am planning to stay some time here. Besides the most annoying thing I had with gs3d was WED which I 'fixed' by having my own map editor (like Sivan did but less profi looking). Now I just need stop making crappy fps games do something where I am good at .
@sivan, well you can always ship your editor with one of your future games as a game map editor. The beauty of that is you probably only need to change a few things (e.g. add the units and stuff that users can place) while possible adding lots of value to your game for mappers/modders .
@txesmi, thats pretty cool, I could also see that being handy for a rts game that doesn't focus on graphics to much but more on tactics or such.
today I added the most simple menus to my editor (can be extended in future) now ground and water terrain placement/removal is simple. water is always scaled to ground.
here's a little update on Rigid Force Alpha. Since my last post I rewrote major parts of the game code and also did some redesign of gameplay systems.
You can now charge your weapon's special function over three levels, depending on how many energy capsules you have collected (visualized by the horizontal bar in the lower screen center). Additionally, you can charge a combo bar by killing enemies. When maxed out, you can unleash a powerful smart bomb attack which instantly kills smaller enemies and bullets around your ship and inflicts heavy damage to larger opponents. This also triggers a combo mode for highscore players. Just two changes of many, but I don't want to bother you with too many details.
Apart from that, both of our composers now finished their tasks which leaves us with a great soundtrack of roughly 40 minutes length. It's a well-balanced mixture of rock, orchestral and electronic elements. And kudos also to Kartoffel, who has built some nice models and also made several improvements to the rendering pipeline. Without that, the shots below would not look half as good.
Speaking of those shots, I also worked on effects and level design/art which can be seen below - these are just some impressions from the Toxic Jungle stage, including the boss enemy, which I'm currently still working on.
It's all wip, but don't let that keep you from giving us feedback.
I do not know how i missed this five images, but looks really great. Kartoffel what kind of shaders you use to make such a good background effect? Any advice how i can get better background effect in my game:
@Iglarion, I think your background matches in style to your game, but of course it can be always better - or find a 2D artist for a lot of money, but players won't see the difference
it would be a stupid thing to switch engine in case of a nearly finished game. sometimes developers do it if want to make some sequel episodes...
@Sivan Yes, using 2d art is good way to create a nice backgrounds. I do not have some big problems in levels where my camera pan/tilt/z is locked like here:
But in levels where camera is much more dynamic, i need use a 3d art style and i never satisfated with results. @Kartoffel thanks for info, i was almost sure that it was a shader Now i know i just need play more with fog and my models. @FEL About "swich engine virus" which circulating around here, i'll be immune to it until i can achieve my goals with Acknex. At this moment Acknex in current state is more than enough for the type of games for which i'm interested. I'm thinking that for the time i would have spent on learning a new engine, it is better to make the 2-3 games and publish it. Why pay penalties for UE, when we know how to ride the dead horse
I do not know how i missed this five images, but looks really great. Kartoffel what kind of shaders you use to make such a good background effect? Any advice how i can get better background effect in my game:
@Iglarion: Kartoffel is right, it's pretty much the oldschool approach that we use, no special shaders involved. The key is to divide the background into several layers. For Rigid Force Alpha I mostly try to seperate my scenes into foreground, midground and background layers. I make sure that there's a certain space between all these layers. This way you get a nice parallax scrolling effect. And furthermore, when you activate distance fog and simply play around with color and distance, you almost automatically get nice results.
The last screen you posted almost has the right direction. I assume that the high sand dunes assets in the background are 2D images. I'd move them further away from the camera to create more depths. Then I'd create some mid-poly sand dunes for the midground and some more detailed ones for the foreground, closest to the walkable area. I think that's pretty much the approach that Frozenbyte used for Trine 2: Trine 2 Desert. You should also take a look at other recent sidescrollers like Donkey Kong Country Returns, they all do it in a pretty similar fashion.
Also throw in a bunch of effects like sanddrifts or tumbleweeds passing by for that level and you're good to go!
@Turrican: Thank you, definitely now i'll play more with layers distance, and fog color/distance to get good effect. Example that you gave me is great, desert level in Trine looks perfect. I miss this addition, but i found something interesting in one level where we can see over 50% of blue sky on screen, beacuse the biggest problem i have with the background on the places where we can see the blue sky. In Trine i see some details are painted in sky layer, like bluried mountains. Maybe is point, paint some fog color and some very bluried mountains in Skycube to get better transition with lastlayer. Btw if i put image instended of skucube i do not know did there are some engine function or shader, where we can control strenght and position of fog on this sky image, i never try this, but this also might work. Or simple ignore fog on this layer and make fake fog effect in Photoshop.
About our desert level, yes i already have some tumbleweeds, heathaze, dust,... all models and concept is there, i just need put this in level, beacuse this is last part of game and has not yet come on the order for polishing. Trine desert level will give me new inspiration .
Here is one picture with forest. I thinking about bacground image and fog color
@Iglarion: That's the point. You need a good transition from foreground to background. You can achieve that by directly painting some translucent "fog" or clouds or whatever into your background/skybox, or by giving it a general color tint that matches your fog. You could also put a separate colored texture with an alpha gradient just above your background image or skybox to create a better transition. Then you can easily adjust its color to match your fog using the RGB values. Like this:
You can also do lots of other cool stuff with a layer setup like this; maybe you could paint clouds on your transition layer and move them using UV shifting.
And regarding the screenshot (beautiful by the way!): Have you tried to just use a fresh, slightly saturated blue fog that matches the sky? That could easily work.
Thank you for image, this transparet layer sounds like the best solution. I almost forget that we already use something like this in our menue screen: https://www.youtube.com/watch?v=ybZBta6mISo&feature=youtu.be I will try make this tranparent layer with moving fog and clouds, i hope this will looks good All these things require a lot of time to adjust, but i'm sure it's worth it.
Originally Posted By: Turrican
And regarding the screenshot (beautiful by the way!): Have you tried to just use a fresh, slightly saturated blue fog that matches the sky? That could easily work.
To be honest i not playing to much with fog, i just try on speed some blue fog before but that do not looks soo good. I must adapt back layers for blue color. Dark green fog looks good but only in dark swamp levels.
I just made my first tests of animated sprite lods in my little RTS test-bed with 500 characters, powered by Fragmotion. Hopefully writing a sprite animation shader will enable proper shadow casting (currently the whole sprite would cast shadow instead of the actual frame), and easy direction handling without skin morphing...
Im currently working on a small multiplayer game where the goal of the game is to knock other players into the lazerzone and kill them and be the last man standing.
@Denn15, looks pretty fun. I like these kind of gamemodes where you can only do 1 thing but it still requires skill (like ut instagib). My only suggestion (but it is something you probably already though about) would be to add maps with more props (like boxes) and things like more edges.
@Reconnoiter thanks. it is inspired by shootmania which is also that kinda thing where there is not much to do but you have to do it well and a map in warcraft 3 where you were pushing each other over the edge. i have thought about having some objects in the map but right now you can very easily get stuck if you try to jump on top of each other so i think i will make the primary goals of the mode work with score and such and look into obstacles when that is done
@Superku thank you. i have posted a little response
Posted By: Anonymous
Re: What are you working on? - 04/23/1515:30
@Denn15 This is my idea of what Sumo Wresting would be like in the high tech future. Cool concept. I would make many unique weapons with maybe only one or 2 ammo per weapon. That will make it a real game of skill.
Hm.. I remove things and prepare everything that it fits into my game pipeline. But you can take lightfrog from AUM resources and use it with A8.. except the skycube everything works good!
Ohh, I see. I asked because I tought that you are going to optimize it, in order to use in your game. Cause I've heard people complaining over the forum that it's not suitable for games (way too slow etc).
I don't know - it works out of the box and is beautiful. The best solution for Gamestudio for all important light types in my opinion. And I don't think it is slow.. soft shadows are always time consuming and I really just need a few dynamic lights. So I don't know what the others did, but I feel comfortable with the performance.
I will optimize it as far as I can, sure. I'll see what I can do about an intermediate techdemo, before I pull it in my game. BTW: I use the standard-material, I don't use normalmaps/specularmaps.
Two quotes to show, why I thought that it's pretty slow:
Originally Posted By: MasterQ32
Also lightfrog is deadly imperformant as it uses simple shadow mapping with 6 views per shadow map (image your game takes about 6 times additional render time PER light)
Originally Posted By: Kartoffel
It is slow and it does use 6 views for shadowmapped pointlights but even the newest engines use this approach because it's simple, fast (through clever culling and similar stuff), easy to blur, etc.
With an engine like acknex, however, performance is pretty bad due to the slow-as-fuck rendering and rather small freedom for optimizations.
If you think 6 views are too much you could use dual paraboloid shadow mapping (2 sm. views) but this results in badly distributed pixel density and higher performance cost because of culling reasons.
I never worked with 'lightfrog' so unfortunately I can't say anything from my own experience. I only run the demo and really loved how it looks like, it reminded me of indoor scenes from STALKER and METRO 2033.
To get something straight: re-rendering the scene is not as costly as you might imagine.
In my oppinion classic cubemap-shadowmapping is the simplest way to perform point-light mapping with lowest degree of shadow aliasing - this is similar to Kartoffels statement. Hummels blur implementation also works very well, so it is a good compromise between simplicity, quality and performance.
Best thing I ever saw for Gamestudio - use it!
What I dislike is the framework/integration friendliness, however, I will fix that for myself.
Well, rerendering the scene is quite costly, especially in terms of CPU overhead in Gamestudio, but even then the cost is usually worth it if it adds nice shadows. And just rendering the depth is a lot faster than calculating complex shading for each fragment. From my experience VSM suffers a lot from light bleeding and there is no good solution for it, except using exponential shadow mapping which mostly fixes that issue, but only when using two 32bit floating point channels, which from my experience is still quite costly on laptop GPUs and often slower than a small filter kernel for percentage closer filtering, even without hardware filtering (with vendor specific extensions or newer DirectX versions it is possible to get 2*2 percentage closer filtering with bilinear blending for the performance of a hard shadow). Also the possibility to use a slope scaled depth bias without having to implement it in your own shaders is quite an advantage of single depth texture based approaches.
Okay, I am not sure what I wanted to say Other than that it depends a lot on the hardware how good things work and that DirectX 9 sucks .
I used to back the static objects depth and only render the dynamic objects for points lights on the run. A double depthcubemap check inside the shader speeds up the thing a lot.
Lightfrog allocates shadowmaps for every single light source; they are not shared, which is a huge waste. I wish I could stop people using it. It was only meant to win me a free pro upgrade/date (not really but that's what I keep telling myself to not feel too bad about that mess).
making a little dynamic array helper to publish soon as a contribution. functionalities: create, create with default values, fill with values, set all bytes, destroy, rotation of 2D and 3D arrays. element addition is not finished yet. I don't know whether to implement element removal or not... what do you think, is it okay in this form?
the header:
Code:
///////////////////////////////////////////
// functions
/*
allocate memory of a new array with the given number of elements,
fill all its bytes with zero for safety
example-1 : MYARRAYTYPE* MyArray = array_new( sizeof(MYARRAYTYPE), NULL, 10 );
example-2 : int MyArrayLength = 10;
MYARRAYTYPE* MyArray = array_new( sizeof(MYARRAYTYPE), NULL, MyArrayLength );
example-3 : int MyArrayLength = 0;
MYARRAYTYPE* MyArray = array_new( sizeof(MYARRAYTYPE), &MyArrayLength, 10 );
(now MyArrayLength is set to 10)
(initially MyArray should be NULL)
(arraylength can be NULL)
returns the array, sets arraylength
*/
void* array_new(int elementsize, int* arraylength, int newelementscount);
/*
same as array_new() but initializes all bytes to (char)defaultvalue (instead of zero)
example-1 : MyArray = array_new_initbytes( sizeof(MYARRAYTYPE), NULL, 10, 255 );
example-2 : MyArray = array_new_initbytes( sizeof(MYARRAYTYPE), &MyArrayLength, 10, 255 );
returns the array, sets arraylength
*/
void* array_new_initbytes(int elementsize, int* arraylength, int newelementscount, int defaultvalue);
/*
fill an array with default data given by one locally created element,
example : MYARRAYTYPE NewElement = 10;
array_fill( MyArray, MyArrayLength, &NewElement );
returns nothing
*/
void array_fill(char* arrayin, int arraylength, char newelement); // by loop and =
void array_fill(short* arrayin, int arraylength, short newelement);
void array_fill(int* arrayin, int arraylength, int newelement);
void array_fill(var* arrayin, int arraylength, var newelement);
void array_fill(VECTOR* arrayin, int arraylength, VECTOR* newelement); // by loop and memcpy
void array_fillbytes(char* arrayin, int arraylength, char bytevalue); // by one memset (fast)
void array_fillbytes(short* arrayin, int arraylength, char bytevalue);
void array_fillbytes(int* arrayin, int arraylength, char bytevalue);
void array_fillbytes(var* arrayin, int arraylength, char bytevalue);
void array_fillbytes(VECTOR* arrayin, int arraylength, char bytevalue);
void array_fillbytes2(void* arrayin, int elementsize, int arraylength, char bytevalue); // name should differs from the other overloaded functions !
/*
reallocate memory of an existing array with a given number of new elements,
without element value initialization, or decrease array length
example : MyArray = array_realloc( MyArray, sizeof(MYARRAYTYPE), &MyArrayLength, 10 );
returns the array, sets arraylength
*/
void* array_realloc(void* arrayin, int elementsize, int* arraylength, int newelementscount);
/*
reallocate memory of an existing array and a given number of new elements,
and fill it with default value given by one locally created element,
or decrease array length
example : MYARRAYTYPE newarrayelement = MyDefaultValue or MyDefaultStruct;
MyArray = array_add( MyArray, sizeof(MYARRAYTYPE), &MyArrayLength, 10, &newarrayelement );
(where new_arrayelement is a local variable or struct created/allocated and set before calling it,
if NULL then new elements will be filled with zeros)
returns the array, sets arraylength
*/
void* array_add(void* arrayin, int elementsize, int* arraylength, int newelementscount, void* newelement); // WRONG ! -> char/short/int/var* newelement because of memcpy
/*
free array memory, set it to NULL for future usage, and set length to zero
example : array_destroy( MyArray, &MyArrayLength );
(arraylength can be NULL)
returns nothing, sets the array, arraylength
*/
void array_destroy(void* arrayin);
void array_destroy(void* arrayin, int* arraylength);
//----------------------
/*
pseudo 2D array operations
*/
/*
2D array helpers
*/
int array_getx(int sizex, int sizey, int abspos);
int array_gety(int sizex, int sizey, int abspos);
int array_getabs(int sizex, int sizey, int posx, int posy);
/*
reposition array elements as it would be rotated as a 2D matrix
to the given direction (angle/45), assuming currently looking towards 0
example : int* MyArray is the existing array to be rotated
int MyRotateArrayLength = MyArrayLength;
int MyRotateSizeX = MySizeX;
int MyRotateSizeY = MySizeY;
MYARRAYTYPE* MyRotatedArray = array_rotate_...( MyArray, &MyRotateArrayLength, 2, &MyRotateSizeX, &MyRotateSizeY );
(direction 2 means rotation by 90 degrees)
(no diagonal rotation supported i.e. 1,3,5,7)
returns the new array (it can be used to overwrite the old one, without using new variables).
*/
char* array_rotate(char* arrayin, int* arraylength, int direction, int* sizex, int* sizey);
short* array_rotate(short* arrayin, int* arraylength, int direction, int* sizex, int* sizey);
int* array_rotate(int* arrayin, int* arraylength, int direction, int* sizex, int* sizey);
var* array_rotate(var* arrayin, int* arraylength, int direction, int* sizex, int* sizey);
//----------------------
/*
pseudo 3D array operations
*/
/*
3D array helpers
*/
int array_getx_multi(int sizex, int sizey, int sizez, int abspos);
int array_gety_multi(int sizex, int sizey, int sizez, int abspos);
int array_getabs_multi(int sizex, int sizey, int posx, int posy, int posz);
/*
same as array_rotate() but supports 3D arrays organized level by level
*/
char* array_rotate_multi(char* arrayin, int* arraylength, int direction, int* sizex, int* sizey);
short* array_rotate_multi(short* arrayin, int* arraylength, int direction, int* sizex, int* sizey);
int* array_rotate_multi(int* arrayin, int* arraylength, int direction, int* sizex, int* sizey);
var* array_rotate_multi(var* arrayin, int* arraylength, int direction, int* sizex, int* sizey);
I will add them a little later ,I am using large base textures plus detail textures for it in my tests so its stil ugly , i just finished adding texture support today ,I didn't want to redistribute my textures so I just added some of my other test textures. . I still have to atlas map ..
thanks .
@txesmi , I like it , looks great , what are you using to do it ?
@ kartoffel thanks , yes this test also is far from being memory efficient and/or low on memory .
if you fire up windows task managet -resource monitor- memory , you will see it eat up a lot of ram .
the reason for this , at this moment even though I destroy entities outside of range , is that the arrays are pre filled for the test , I have to switch to dictionaries and other methods to only create what I need for visible range instead of filling invisible range data ..
to do this I would need to serialize to disk so I do not store un needed data anymore , and that goes more towards endless generation of visible data..
I have a few experimental updating conditions I am testing now , for instance only updating when frame rate is above a few frames below frame cap , this however needs more thought because a user can easily travel beyond update range in that time , so I need to manage it such that this update condition is overidable at the correct time.
then I am also leaning towards pausing and resuming the update thread , but I haven't even gone beyond this in theory only as of yet .
tust was great help to me in studying voxels plus a certain unity tutorial aided me in forming the basis of understanding voxel worlds , so thanks to those people , however I suck in lite-c , I am better at bringing my concepts to actuality in c++ and now c#( which I am new at ).
I have just finished completely rewriting my chunking system for my polygonal representation of a voxel world .
my new chunking system now , is fully dynamic ,in every axis direction , fully 3d in to negative and positive space, but now comes the distance management to add and remove those chunks as needed to give the impression of infinity, this is where it could get hairy if I dont manage it efficiently .
Note: this test ,does not include any distance based management so frame rates and memmory could drop badly under stress .
the log may slow performance down aswell as d3d line display mode , currently theres a couple of un needed performance stresses because of the testing process and because I just finished re-coding ...
bug I know of : when creating a tile on the inside of a chunk border , it needlessly creates a new chunk which is only supposes to be created if the border block already existed..
space -remove shift -add x - remove chunk f6 - d3d_lines off f5 - d3d_lines on
@Txesmi: Looks good! To bad that's the last one @MasterQ32: That looks awesome man! very beautiful style you have! @Wjbender: Sweet. Are you planning on combining 3D Models with your world?
I've been working on my website.. which I've been cramming with resources... http://www.cgforfree.com
@dlively thats great , do you also allow people to sell there ? you could say allow users to sell but by agreeing you take a percentage of their sale for using your web services .
I am not sure yet where I am going with it , I thought I like to split the project in to two or more branches fir separate types of projects based on voxels , like one minecraft type terrain ,one where I try and turn it in to real terrain surfaces , and what ever else comes to mind..
with the help of the log and c# debugger I finally figured out the bug ,turns out it produced a hell of a lot of unseen un-needed chunks and stresses.
glad to stil see gs users go through such extra lengths
@txesmi he's gotta have a rocket pack or back pack right ?
@DLively, I like the frontpage, the other things could use some work. Some pages are not consistent, the different pages differ from style. Also some text is a bit hard to read. Never do black on a dark background (dark green in this case). I would make all the text black except the title+subtitle of the site (the one in the top left corner) and make the background completely white like you did on the news section. Lastly change button appearance when the mouse hovers over them . So in short, I would suggest making the 'shop' and 'free' pages similar to the frontpage. gl and good job so far
He is the driver of the first ship I showed, the buggy shapped one. Been sit on the ship only shows the head, the arms and the upper front part of the model.
Wjbender@ I tested your "voxel" demo, and I have to say it works great! I've been toying around removing/creating tiles and I got constant 60 fps (well, sometimes I got fps drop to 59 for a few seconds, when I was editing blocks too aggressively).
Originally Posted By: txesmi
Taking into account that the concept flow over big heads and small machines, I feel the head too big. What do you think?
It looks awesome to me. BTW, would be nice to see that big head leaning from side to side while rotating vehicle fast.
do you also allow people to sell there ? you could say allow users to sell but by agreeing you take a percentage of their sale for using your web services
@Wjbender: Yes. I just have to update my EULA first. I really hope you figure out the real terrain and the like.. that would be very cool to see!
Quote:
@DLively, I like the frontpage, the other things could use some work. Some pages are not consistent, the different pages differ from style. Also some text is a bit hard to read. Never do black on a dark background (dark green in this case). I would make all the text black except the title+subtitle of the site (the one in the top left corner) and make the background completely white like you did on the news section. Lastly change button appearance when the mouse hovers over them tongue . So in short, I would suggest making the 'shop' and 'free' pages similar to the frontpage. gl and good job so far laugh
ps: I like the first pic with the editor stuff.
@Reconnoiter: This is awesome feedback. Which pages do you find inconsistent, more specifically; which two pages are you comparing? In short; The main page is mainly its own thing. The rest of the website will be similar to the resource and webstore pages. On that note, all resource and webstore pages are consistent with each other.. afaik And the rest of the website isn't entirely styled yet. TBH, I had an old design and domain 5 days ago -- So on monday I purchased my new domain and gave my website a facelift. By tuesday, I was uploading models again. Its not completely styled yet either
I like the green background, however I'll take some of your advice and change the text color to white instead, to make it easier to read. After reading your comment and looking at the text I have to agree with you its straining on the eyes.
and I'll change button appearance on hover .. Thanks for your input and suggestions! You've inspired me
I've started a project this week called SolidLayout.
It's a layouting engine combined with a custom layouting language like HTML/XAML. The current status is: Attributes are bindable or statically assignable, layout supports two basic layouts: Stacking or Docking.
Here is an example of how to declare a UI layout as well as how it looks:
I know the example is pretty lame, but imaging declaring Gamestudio PANEL/TEXTs with a layouting engine with bindings of properties via Lite-C variables.
Right now the project has no Lite-C compatibility or even windows, but those will be my next steps as i think gamestudio integration will be pretty simple, but also useful.
I will keep you updated
Also here is an example how a Gamestudio layout could look:
Code:
# Simple main menu with stacked elements
Blank {
Panel(width=150, height=300, horizontalAlignment=center, verticalAlignment=center, bgcolor="#888888") {
StackLayout(verticalAlignment=center,horizontalAlignment=stretch) {
Panel(height=30,bgcolor="#FFFFFF",event=[startGame]) {
Text(text="Start Game",flags="CENTER_X|CENTER_Y");
}
Blank(height=5); # 5 Pixel Distance between elements
Panel(height=30,bgcolor="#FFFFFF",event=[loadGame]) {
Text(text="Load Game",flags="CENTER_X|CENTER_Y");
}
Blank(height=5); # 5 Pixel Distance between elements
Panel(height=30,bgcolor="#FFFFFF",event=[quitGame]) {
Text(text="Quit Game",flags="CENTER_X|CENTER_Y");
}
}
}
}
Which pages do you find inconsistent, more specifically; which two pages are you comparing?
, glad it is usefull, I will try to answer your questions even though I lack real website building experience (I only have made a simple site once and did some tuts for lateron).
I was comparing the main/frontpage with the other pages. The frontpage had a 'less is more' feeling with most of the text clearly readable (black text on white background = nice contrast). The pictures switching is nice too. Nice logo too, but ugly buttons (I would remove gradient difference on the buttons and instead go for something simple, maybe black+white, to play it safe ). I see now that you have added some stuff (/buttons) which makes it a bit more messy in my opinion (is too clustered beneath each other while having a very big empty place to the right that you can use). I do like how you positioned the menu/shop/free buttons at the very top of the page
Now the resource / webstore pages. Click on them and you get a similar layout with extra green bars beneath. I would make that white. You can always make eula text bold or color it so it is more striking (/not easy to ignore). Perhaps add some images here like you did on the frontpage for a nice look.
When you click a resource, let's say bathroom page 1 it adds additional stuff beneath the page. I find this awkward, I think it would be better if it goes to a new page or places the bathroom stuff more to the top so it can be seen instantly without having to scroll below. Or maybe auto scrolls to the right position. So when you click e.g. Bathroom, you see the available items. What I think is nice is the information given here for each item (preview images, link to watch it on youtube etc.). What I think would make it more clearly would be some thicker border around the descriptions and thick borders around each item (+ a bit more whitespace below and above items). Speaking of borders, the buttons could use some nice solid borders too .
Quote:
Its not completely styled yet either wink
, yes most of my suggestions has to do with styling and visual layout (or whatever it is called) I think, so some points you probably already thought about My experience is too limited to give advise on technical points
I am currently working on a collection of mini games. Each game will be very simple and basic but can be used as a good start for new game projects. Some of the games are based on the Mario party games from the Nintendo system. The games will not be complicated and very basic in setup. The first mini game is almost finished and many more ideas are on the shelve. Thanks to Malice the first one is nearly done. Here is a clip that shows it all in action :
Reconnoiter. I ended up taking all your advice. It actually looks pretty good now
EDIT: I'm not sure what Im planning for the empty space on the right. Maybe new stuff... Ideas are still brewing. I changed the layout because the old layout didn't allow enough room for what I wanted to do -- So nothing was planned :3
You are the end user, so your advice is necessary.
I am expecting bad performance , but I don't know and I would appreciate anyone giving feedback on the amount of fps they get ,so I can estimate what to do , because on my laptop things aren't running that great yet (without distance management abd other important features) currently.
when ran , this test will be active a small amount then be busy for a while , after that it should read 512 entities on the top left .
the fps I would like to know is during the active rendering part of the entities , and perhaps some modifications to the world. .
f5 d3d_lines -on f6 d3d_lines -off x - delete chunk space -remove block shift -add block
yellow coloured highlights - shows chunk updates
red coloured highlights - shows newly created chunks
@Wjbender Contant 57-59 fps. d3d lines activated or not. It sometimes get freezed for a tenth part of a second when creating or deleting blocks, but not always.
okay thanks alot txesmi , I was expecting the worst , so I should be able to make a rough performance estimate compared with my fps .
Thank you for taking the time , and yes your model looks just great in gs , to be honest I had a lot of difficulty finding a good workflow for models in to gs from both 3dsmax and blender , modeling was one of my fav past times but when it comes to texturing I suck badly , and I really should'nt because I am not to bad at airbrushing (not the digital kind) and drawing.
wjbender: 50fps without d3dlines and 45fps with As txesmi said, there is a slight lag when deleting chunks.
intel Core 2Quad (4cpus) ~2.3Ghz Geforce G210
Posted By: Anonymous
Re: What are you working on? - 05/18/1523:50
@wjbender - I would test it if you real like, but I'm running a 3rd gen i3 Intel HD 4000 with 4 GB ram on a laptop. Based on the status of the 2 machines above with dedicated GPU, I would expect poor performance.
Let me know if you are interested in the numbers. Mal
@Dlively: Less is more. Try displaying as few as possible one one page. Put the impressum/EULA stuff at the bottom into a separate page. Remove the black bar at the bottom, it's too much distracting from the rest. Use less color (only black/white/gray/accent-color), add more contrast I'm feeling completly lost on where to focus on the site, i need to search for everything. Try creating a design that guides the user, not with text but visual abstraction
little time, little progress: no more concave areas, and much faster map decomposition, approx. 45 sec for 400x400 tiles in case of this relatively complex level (enabling procedural terrain in future) this debug shader set is a bit heavy on performance... click to enlarge image free hosting
@sivan that looks cool , I would play it in debug mode just because it looks so colourful and interesting .
I have added some greens and browns , and using the estimate for good performance view range and chunk amounts, I have started working on clipping etc.
@realspawn looks great , you could totally make such games sellable as a bunch together.
@txesmi it looks game ready , and definitely fits a certain cartoonish style , I could see a sellable game too.
@dlively thanks and I am using opera mini on a Samsung ..
this is the first workable test for an "endless" world , its inefficient ,crappy ,jerky ,low view range, no serializing so world is non persistent ,theres a bug wich constantly creates and updates and I think flush chunks wich costs alot of my frames wich I cannot figure out yet, but it works .
first I had it on a thread , had constant 60 fps ,and it was as smooth as ever ,but nooooooo ...Engine functions aren't thread safe ...#@##@#
I went up to 90,000 quants before I got tired of jerky blocks heading my way .
I have a lot of ideas for mini games so i will try to create them all then perhaps make full games out of them of make a total package with resourche and models so new people can start right away
@txesmi: very nice indeed. I agree with Superku, an F-Zero style game would be absolutely rad. Add network+local multiplayer for ~8 players with quick and easy matchmaking and release it on steam for free (or maybe 1-2 USD), and you'd have a hit.
@Kartoffel - I'm assuming that the background and the bars are animated to move with the music playing, if that's the case, I so badly want to use this for myself. I love the background, and would love to see it move to my favorite music.
aww, damn, that is a bummer. I look forward to seeing how it turns out though!
I tried working on my own a while back, I got the bars moving up and down, but I had, and still have no idea on how to get it to react to the frequencies of music.
Working on my fps ( like the last years ). Interaction between weapons and zombie - bodys right now. It feels a bit like Dead Island ( had luck while i wrote the pain blending lol ) ^^ Melee weapons and first slow zombie nearly done now. Zombie can grab player biting into his face ( but not if player wears the iron-mask btw^^ [mean he is able2 grab player but bites wont harm him then] ). Background is only a dev / test room, so no level shot for now. Will show more "soon^^" ( and Videos ).
Hope u like the look i managed (via ShadeC) of weapon and first ai, much as i do. ( will reduce normal mapping strength of player's hands )
Was a long n hard way. Please keep in mind its all WIP, and this is the first posted shot ^^
another weapon, nailed wood piece
( btw, base of used models from TheGameCreators [mod. versions] ) edit: of course player can mod the weapons. Selectsystem is like in MaxPayne ( thanks for this idea [and the pathfinding system] SuperKu, again ^^ ). Story in short words: player must go pillaging staying alive, of course.
Thanks! Just installed AudioTech Mic with amplifier and extern 32bit sound card. Now ill make Zombie sounds ( good tool (love it) CoolEditPro2 btw ). ^^ ...and dont forget to close the window neighbours might call the police lol
I like ntract , but that cooleditpro2 looks cool , I can just imagine the look on peoples faces if you have to record grunts .
with my audio i had once, hooked up a 2000w car amp & a sub woofer plus 2 of my 5.1 channel audio tweaters plus its sub then I run my guitar through line in , in to a digital amp model and whack out some screetching hendrix or a bit of santana with a backtracts , luckily never had the neighbours conplain .
then along came windows 7 and they totally ruined my mojo with theire multiple audio playback problems , totally something I hate about win 7
I forgotten to say, need2 record female Zombie sounds lol ...even more freaky
and yes, windows7 audio sucks ( btw, sndrec32 now totally sucks, loved that tiny app [buts its gone since XP] ). iam lucky i can run my Phase26 at all, with some drivers a hobby - guy made for 64bit Systems and this hardware.
Posted By: Anonymous
Re: What are you working on? - 06/05/1514:39
Have you looked into Audition CC, it's a month rent software, but it is the evolution of Cool Edit 2(which I loved in 2003). If you are using a 64 bit OS and system you, might benefit from Adobe's evolution software.
Looks good, you should disable the button Start Server once its started and same with Stop server. For some design improvement stuff you might add a manifest to the file so you have a nice windows 7 gui, rest is really great. My server console is actually really just a console, I should add support for a gui too later
Thanks I tried the manifest thingy but could not achieve it. I tried all ways but could not link the manifest to the runtime file. Other than that the buttons get deactivated once the server has started, but thanks for pointing that out.
IIRC Simply name your Manifast YourAppName.Exe.Manifest If you do this, you don't have to add it to the resources of an Application. Just place it into the same folder as the exe with the Namestructure above
Made some final adjustments to the lighting and postprocessing of my pixel-style renderer. (Also, thanks to Superku I managed to fix an important "bug" (aka. my stupidity))
Rendering is pretty fast, around 2-3 ms/frame at 720p and there's still some room for improvements.
However, I'm having huge problems regarding the interface. Nothing works. at all. There's not even anything being rendered. ...and since nothing happens I can't seem to find the problem which is making me go insane right now.
I don't wanna sound selfish but I'm starting to think that the renderer is capable of rendering some pretty nice scenes. However, my artistic skills still need some work (actually quite a lot of work ._.)
And I FINALLY managed to fix the interface issue. The problem was that I'm setting up all the views before loading the level - which works just fine - but the parameters of the interface-view always got reset when loading the level (because it's position is not updated during gameplay) and it ended up somwhere, but not where I wanted it to be.
There are still some minor issues but fixing these will (hopefully) take less effort and cause not as much of a headache.
Working on a tutorial based on the one Unity has :the space shooter I was like hey i can do that with 3DGS as well so no 2D scroller but using real 3D models and all
it's almost done and will be there for download soon
it will contain :
- Window size - Scrolling background - Player movement - Blasting astroids and enemies - Score - Random wave of astroids and enemies - Particle ship engine flames - Explosions/animates sprites + FX - Restart game - Show sign at borders space - Enemy fire
When game is done full documentation will come with it so it can be easy used as a tempplate for scrolling shooter games
I've made some improvements to my font-to-bmap script featuring colors now:
It only supports bmap-fonts, I had to hardcode the width for all 256 characters and I think I've written the worst if-comparison of my life but somehow it works.
However, although Alibaba showed me a workaround on how to create sprites from bmaps I'm still having scale problems with these sprites in my interface. ..so I'm thinking about switching to a draw_quad-base approach.
Using draw_... commands is perfectly fine (I use them all the time). They, as well as regular true type fonts, used to be super slow in the past but aren't that slow anymore.
I like the lighting and the bloom in your screenshots!
Guys, Ive used the draw_text command and draw_quad command for my playerlist, and the FPS for arround 32 players suddenly gone down from 60 to 20. I replaced them with a normal TEXT object and there was no FPS decrease at all.
Im still working on melee weapons. Heres the first version of the "Electro Bushwacka with Pink Glow Slime (machete mod level 1)" ( inspired by Dying Light^^ ). Not the final version, still WIP. Will add some more (a lot) cables, led's and stuff ^^ . Sparks etc already working.
@MasterQ32 What happend to your cut - through shader btw ? ^^
Now I'm at the point where I'm stuck due to missing assets, storyboard and more... I don't know if this thing will ever finish. I feel pretty much left alone with this.
Started with the last scene - still a bit empty. I want to get this thing done now. Even if I have to remove all voice samples because they are not recorded in the end...
Is this working like your lighting system in SuperKu? I mean where you just have cubes and therefore can use some fancy calculation to improve the speed?
Well they, the two systems, work completely different but yes, it's restricted to a rather blocky or primitive based style (cubes, planes, spheres, cylinders). The good thing about this is that I can add different light types like spot- and flashlights, illuminating surfaces, rotating warning lights and other stuff rather easily - at least in theory!
Pretty much, Brutal Doom (/ II), Turok 1, maybe a touch of Metroid Prime and Resident Evil 1 Remake serve as inspiration. EDIT: Oh, and something else, but that's a secret!
this week I implemented this behaviour tree (inspired by migration to UE4), replacing the old state machine, and I'm quite happy now. sucked means stucked code is more simple, thus performance is slightly better.
basic description: the group manager defines desired targets, and units start to move if it differs from their position. sounds simple
conclusion: it is better to use a clean easy-to-debug system for AI design; and without visual tools it is hard to make a complex one (like in UE4).
warning : this is unrelated to the use of gamestudio engine , but related because I am porting my voxel code to irrlicht (yes consider it the devils work if you want) , performance is quite good compared over gamestudio , little perks as threading and good mesh buffers above others makes it a joy to port .
if you download this (little over 1mb) , be prepared to wait a little while the meshing turns 3556629 solid voxels of a 7077888 voxel field ,in to blocky meshes , around 500 updated chunks you should be able to start seeing meshing in front of the camera and you could start moving around.
thus far with the project being very simple still , I am quite happy , I had done tests of up to 10 million solid voxels and 32×32×32 voxels per chunk quite successful , although ofcourse there wouldn't be that much chunks in view at once when I am finished , especially not the same kind of noise function.
edit..I see out of haste I didn't thread lock so the timing things happen at will crash , just restart until it functions...alt+f4 exit
Hi, would you consider the following art-style / graphics of the weapon + arm in the pic below be good enough for a free fps game? (I ask cause I always find graphics hard to get right)
Im using a custom object and custom functions to edit and handle unicode strings. This also includes saving to and loading from text files. I did this because not all of acknex's string functions work with unicode STRING* objects.
Im using a custom object and custom functions to edit and handle unicode strings. This also includes saving to and loading from text files. I did this because not all of acknex's string functions work with unicode STRING* objects.
So, did you maybe wrapped those strings/chars together with other data into structs and did you wrote some special handlers for them?
Not really. I used the native str_ commands. The only problem I found is str_cat_num that messes up the format string when it carries more than the number format. Solved by adding the text and the numbers in different instructions.
I also used file_load in order to be able to open UTF16 files from an offuscated resource file.
I present the next 3dgs contest. It is the the " put up or shut up challenge"
With in the next next 18 months, put aside time , money , recruit team members , start a kickstarter(or EU version), and create the best mother fucking game you can.
At the end of 18 months , put up a near finished version or shut the hell up because your just playing at a dream of game creation. Bam mofo's
Hahah Mal
Posted By: Anonymous
Re: What are you working on? - 08/09/1504:27
Submit fully detailed design docs to the forum within 3 months
P.F.S if you ddoc is less then one 20 pages . it is not a ddoc it's a worthless brief
Just say if you in or out. Will have the mod's create a locked thread for us. We do need to keep outsiders out after all.
Posted By: Anonymous
Re: What are you working on? - 08/09/1505:22
You know why this engine is not being developed, why jcl can not fund improvements. it is simple, you lone wolf ass holes don't produce shit. Can't decide on a idea, can not draw up designs, can't work with or develop teams. Can't produce shit??
Fucking hobbyist. Nothing and I mean no fucking thing in this world was ever made by one man. You jack ass' ! Group up. team up, get a producer like me and make something. Any fucking thing! 20 damn games and 5 fuck success rock this engine back to life. People once asked jcl if they could help develop the engine and he said " just focus on making games" well he was right! Where the fuck is Sid's and slin's engine and why the fuck can't they lend there God given genius to producing something fucking substantial from this engine.
You all a bunch of noise to me. 10 years trouble shooting cry fucking babies code. Where did I learn, solving bitch ass problem for bitchs that have fallen by the wayside. I know the names for many member long loged off this forum.
Make something , sell something and do it soon or just kiss the fucking engine go bye.
lol, if the engine wouldnt be that buggy my game would already be finishhed No seriously, theres alot bullshit happening when your using some more advanced models and textures. Im just preloading my weapons so they dont lag like shit when a guy drops them after a kill and this steals 1.3GB RAM. Now tell me why the engine cant load a fucking level, ah yeah Im using non standard textures, thats why. Its a 32bit engine and it cant use more then 1.5GB RAM before it simply crashes wow And to do some more advanced stuff I had to hook and manipulate some functions too ..
Posted By: Anonymous
Re: What are you working on? - 08/09/1506:37
Advanced model... 1.5 gb ram.... blah blah blah
Heres a idea, don't produce this game with this engine. instead produce 3 games with out advanced anything"" that the engine can run. 32 bit engine built for win xp so fucking build game for that target or move on
Put up or shut up, there are other engines that will run your game.
Of course there are other engines, but I wanna finish that game before leaving this engine. Actually I wanna finish a game I started 3 years ago and I will. Like windows XP this will be gone soon...
Aww, did Malice get a boo boo? It's all gonna be ok baby, daddy is here.
Patronizing comments aside (just jokes man), the last thing we need in this community is someone spitting venom and throwing out obscenities every other sentence to make a point. If your goal was to motivate people you're doing it wrong. I can appreciate your reasons and understand the motivation behind your words. But there is a much better way to convey your message like an adult, and if you really want people to take your requests/comments seriously I would recommend finding a different way to do so. Personally, I use 3dgs as a hobby and not a career since my real career is in enterprise level web hosting (but I digress). There isn't anything wrong with wanting to use a tool like 3DGS as a learning experience. If you want to get a team together and try making a real solid game then by all means, go ahead! But don't call out or belittle "hobbyists", they are the spark that one day may turn into an inferno.
"Nothing and I mean no fucking thing in this world was ever made by one man" I certainly hope you're being hyperbolic, but in regards to video games I can think of a few that were made with one or two people (Jazz Jackrabbit and Minecraft comes to mind first).
As someone who works in the support industry, people can be ABSOLUTELY stupid. But someone who steps up and helps a person grow and learn is, in my book, the most important thing to this community. I personally have to thank Superku (Felix) and Darkinferno for helping me so many times in the past when I was still learning the ropes. Honestly I wouldn't be at the level I am if it weren't for them and I thank them truly from the bottom of my heart. The point is, you have done great work in the past and have helped a lot of people (even me ). I know you may be frustrated and possibly jaded, but I think if you were to take a step back and look at the good you have done for this community you would be surprised.
Anyway, I am going to end my rant as this is not the place to have such discussions. If you ever need to vent please feel free to message me brother. A community sticks together, and dammit you are a pillar in this community! Be the change you want to see in the world, put together your own team, pursue your own dreams, and show us lowly hobbyists how its done! I believe in you man
@Malice: Calm down, seriously! Those insults go way too far and don't do anything. Additionally, I don't agree with anything you said in your last 3 posts. If you really think you need a (20+ or even 1+) page game design document to make a game you don't know anything about game dev but just try to imitate AAA studios and their workflow of assigned tasks. For hobbyists or indie devs IMO such a document is counterproductive instead, you should change and adapt your project and the idea behind it at any time, based on your progression and possible new and better ideas.
Thank you Superku, I needed a bit of help and some sleeping hours to take courage and throw away yesterdays work to the trash. Your sentence is fully brightening
Finally got the shader based antialias properly working from a 2^2 render and a simple downsample shader.
I did not define the zbuffer bitmap and the shader was messing up the coordinates. I almost went crazy xP
@Reconnoiter I know, but there is an inherited 'high quality' button I have to fill, although there is no way to improve those bunglers models with post processes.
Might be a bit blurry for some ppl but I prefer the SSAA'd version. However, I'm not sure about the hdr. It's pretty subtle and doesn't seem to be using actual hdr.
This second blur is much better I think. I personally like the subtle hdr, feels not to overbloomed. But you could always add a more strong hdr as an option perhaps?
- the menu and the GUI is very nice. - I would add some more lights to the scene (lamps switched on), mixing the warm Sun light with colder artificial lights would result in more realism. after daytime matches the finals could be played in the evening or at night. - surfaces are very flat, I would add some noise by normal maps at least to the red-white-green field dividers (and to the field to get a slightly worn look). also some more reflections (cube-mapping or at least speculars) would be great on close items. - the gameplay seems to be simple, I mean targeting the ball and making a straight and accurate rolling, it could be disturbed a bit by some human breathing movements or something similar... in a darts game the hand moved from left to right and backwards continuously, so the speed adjustment and aiming were not so easy. otherwise, really nice job!
Its good. Suggestions: the points Sivan mentionend and maybe some slightly different gamemodes (2-3, e.g. a bit difference in rules); could increase the game length I think. Also maybe add the option to play with 2 people if you haven't already.
I got some feedback too! - It looks like the people are floating a few centimeters over the ground (and I am quite sure that they really do). - Some real shadows for the people and balls would be awesome, but I get that it would be a bit tricky to mix them with the static shadow map and make it all look good together. - It looks like you are trying to do some kind of SSAO, but it causes halos around the players which doesn't look good. So maybe don't use it or try a much shorter depth bias. - The UI graphics are not pretty. Making them flat and tweaking the colors would improve the general look a lot already. - Some dust particles in front of the windows with the sun would probably look awesome.
Thank you all for your reviews and suggestions! I will seriously consider all of them. How hard it is to suit all tastes for a hapless designer like me
The truth is that the plan was only to give the game a face scrub while giving multilanguage support. It went far and pretty far I guess. It has been reworked for a full day working long month and the budget is probably depleted. My kind boss is confident that I know what is necessary and where to stop. Fortunately the main goal has been reached fast and with no significant problems.
Now is the time to find the things can improve the game with the less effort and not get lost on the way. The player characters models and animations are a complete disaster. They fly and skid! It is a must but hard for me.
It's all looking rather good but I seriously hate SSAO/ other ambient occlusion, I just hate it. I'd put an option into the game to deactivate it, which is what I always immediately do in all (few) (AAA) games that I play. Apart from my hate it doesn't make a lot of sense that there's AO in let's say the first shot around the player. Additionally, the default click sound from the video is a little obtrusive.
Hi all! Some of your suggestions are on the rails - There will probably be new throwing mechanics based on mouse movement. I have to think a bit and look for examples and ideas. Any good reference? - Hot seat multiplayer and two different game rules were implemented already - Real time clip shadows for the characters. I deeped on this task in the past so it will be easy. - Flying lighted particles. It seems tricky. Never did this effect before so I will left for the end. I wrote a fast and good looking animated fog based on two skycubes. May be its chance. - A sound effects rebuild. I was aware of their low quality but got used to them. Someone pointed it out on other forum and taked me from blindness.
The AO is not really an AO. It has a lot to do with the AO algorithms I played with in the past but decidedly it is an edge detector based on a 16bits depth map. Like the HDR that is not exactly a HDR but a fullscreen soft bloom because there is no dynamic gamma correction in the way.
The AO is not realistic and it was not intended to be realistic. I received totally contrary reviews about its convenience. There is people who loved it and people who hated it. I can see good points on it. Gives deepness contrast to the scene, darkens the crowd surrounding areas, stands out the player characters carrying them to a clear frontplane. I am in the group of its lovers. The producer liked it so definitely it will be there.
Anyway I did it less noticeable and there is an option to disable it at the same time of the full screen bloom, so not a problem at all.
Your critics are welcome and help a lot. Sincerelly, thank you all. Salud!
Those two windows above the wood area look a little off to me, do people put them on different heights compared to the rest (in such or a similar case)?
BTW, iam still working on my Level conversion from a Unity Project to 3dGS, post some shots soon, i guess. And playing around with process injections for a few days now ^^
Awesome! maybe gig the timing but I have no clue the overall feel. IT feels slow then enemy jumpy, strong contradiction. However that may very well be the point of the game. Without playing I can not say how it truly feels. Maybe look at reducing the margin between the two. Beautify look, music and clip provide opposite reactions, with the music enrgizng and very fast paced kind action and the game slow and creepy. Suggest creep slower music to being fear anxiety. Environment sounds as well for same point. May simple feed back.
@Dico: good work, I like especially the blinks. I had to blink as well as I watched the video, maybe it should be introduced in Windows 10 to decrease the bad consequences of drying eyes caused by the steady eyes during watching the monitor.
Sivan , this video is recorded from windows 10 , the scen gameplay is a slow that cause I have an old graphique card. But soon I will buy new one 😊. Txesmi , this is just a scen from gameplay . Thanks you guys 😊
Awesome! maybe gig the timing but I have no clue the overall feel. IT feels slow then enemy jumpy, strong contradiction. However that may very well be the point of the game. Without playing I can not say how it truly feels. Maybe look at reducing the margin between the two. Beautify look, music and clip provide opposite reactions, with the music enrgizng and very fast paced kind action and the game slow and creepy. Suggest creep slower music to being fear anxiety. Environment sounds as well for same point. May simple feed back.
Really great and interesting.
Thanks Malice , it's an old project and i will back to it soon .
Ofcourse it will be cheap . The converter just need java jdk , no need to android sdk or ndk and in 30 second you get your game ready for web and android .
Posted By: Anonymous
Re: What are you working on? - 09/03/1515:56
I would love to see the pluggins opened to users. I would wait a bit till better users have vetted it. however, I would surely buy it at any reasonable price I could afford.
Also I would like to point out to you that you will be opening up yourself to a flood of support questions from users. And not just the really good questions or possible bug reports, but also the newbie playing with advanced toy questions. The LBGUI thread comes to mind.
I would love to see the pluggins opened to users. I would wait a bit till better users have vetted it. however, I would surely buy it at any reasonable price I could afford.
Also I would like to point out to you that you will be opening up yourself to a flood of support questions from users. And not just the really good questions or possible bug reports, but also the newbie playing with advanced toy questions. The LBGUI thread comes to mind.
Originally Posted By: sitrep
If it's affordable and allows 3dgs 8 to export to android - sounds good !!!
when is the release ?
Thanks , i will release the first converter soon , and yes it support a6,a7,a8 and can move to any engine easly This plugin created by my engine.
with debug shader showing clusters: without debug shader:
but at least no major issues, better behaviour for the price of worse performance, allowing lower unit quantity. so now a bit different game-play is planned...
That speed modeling video is pretty cool and interesting!
Had an idea for a technique on the river in Canada yesterday to display mountains full of trees like I see them here every day. The idea is to only transfer basic data to the graphics card (a 32x32 RGB texture in the following example and a simple mountain mesh) and use that to create and render a bunch of (what I call) virtual sprites. Runs on almost 300fps on my laptop, depending on how much area of the screen is covered with trees. Maybe that technique could be extended to render fur but that will get more complicated and slower, and right now it's still full of bugs too.
Posted By: Anonymous
Re: What are you working on? - 09/07/1501:06
Awesome last 3 post, and here I was just listening to some fool tell me how usual this engine is.
to be fair rvl's post is not engine spec, but still a awesome use of this community.
Thx for the Feedback guys @Superku and Anyone else... Interested in some more Speed Art? But only with smaller assets, cuz one fully modeled / textured Building with anything on it will be to time consuming
@txesmi
Quote:
but your texture colors are so unsaturated...
For one only Reason... I Like it I dont like this Colorfull rainbow tasty happiness lucky... Style. I like more the: Darker Abandoned Dirty used... Style
@Malice
Quote:
to be fair rvl's post is not engine spec, but still a awesome use of this community.
Thanks Malice for these kind of Words Youre Right, all images ive shown so far are made Directly out of Blender, without any Lighting / Shaders or what else
One more Image, Just to show the Modeling(Solid).
(Yeah ive created the Frontdoor, but the Texture is not finished yet )
EDIT: ive already noticed that the door is to broad... Changed it
So Far: Tris: ~15.000 Textures Used: many, havent optimized it yet
@rvL_eXile: Looking good but something about the glass area bugs me a little. Maybe it's the broken glass itself (maybe less detail/ "bumps" could look more natural) or that someone with a house like that would not spend that much time on cutting out perfectly fitting wooden parts to replace those 2 windows. But that's being overly picky.
@JibbSmart: Thanks! I use the pixel's world pos of the simple tree mesh (which is a z-extruded copy of the mountain mesh) and read the mountain height from a 32x32 bmap (without filtering, resulting in 32x32 trees). The idea now is to create a "virtual sprite" facing the camera (which is just a position and an x(y)z normal (DX coordinates), calculate the intersection with the (worldPos-vecViewPos) directional vector and tex2D a "clamp"-tree-texture (the next step is to use a tree texture atlas for variation and clamp and shift the Tex coordinates manually).
After some time I finally could work on my framework again. The framework (named Inca) now contains a simple drag and drop system, but still work in progress.
Fun fact the framework originally started as a chart framework. But at the moment I still haven’t implemented a working chart.
16 files chart related 131 files inca framework
A little more than 12% of the files for the actual planned feature
working on a little doomed template from scratch almost all doom game aspects are in it. One final thing to figure out and i will upload it.
- movement keys and mouse - First person view - Animated weapon - Exploding stuff - Enemies attack - Doors (With and without keys) - Lives/score/health - Shooting bullets - Elevator (not working yet)
^ That's a cool start bro. You'r gonna get the big issue question with this template - namely enemy AI, pathfinding, enemy control functions, center of screen cross-hair hitting, enemy firing offset(always hits player or random misses), there are more, but you get the idea.
This will be a simple mini template so people can click together a doom kind of game
for first part i need the elevator to work first. From there i can expand in parts. It will hold all basics so user build a level places hero crates enemies assign them their action ands voila play on !
thanks to the help of Superku, finally the animated sprite lods of my soldier models casting proper shadows, resulting in sensible performance increase.
in action (only the few very close units are models, but distance can be tweaked smoothly):
Hey there guys, impressive to see that some of you guys are still here.
I thought maybe it would be cool to show you some of my most recent work. Its not really connected to 3DGS anymore. Just some Modelling with Cinema4D.
I wasnt around here much because I finished MediaDesign school over the last 2 years... but I really got a lot into 3D Modelling ;D totally loved it. And totally love what you guys are doing her
Hey Aztec Your alive? Like ure work, but i have to criticise she shape of the spoon. Why is it so "sharp" at the top? Also the fork should be a bit more bend in my opinion
At the kitchen, the "Worktop" the Edge of the Front / Top Faces, is rounded? Could be look better with an clear and sharp edge. AND where are the drawer / cubboards?
In the "TV Picture", please make it a bit brighter it looks too Dark (Black TV / Loudspeakers / Alarmclock / Table under Alarmclock).
The Last Picture looks great so far, but could maybe looks much greater with some additional Lights.
So Far, Great to see u again here with some nice work
tree models having billboard lod thx to txesmi (1500 pcs without model merging) (soldiers too, but they switch from 3d bone animated model to a separate animated sprite)
Hm since you cannot really post images in Showcase threads and my old projects thread is rather dead, I will make a Superku development post here for now:
I had this idea for a level where you'd activate some kind of water pump or boiler, which would then lead to rising water and you could reach heights and areas which were inaccessible before. Obviously, you'd release a bunch of piranhas as well!
The water (and to some degree the piranhas as well) posed a big problem for me, I just couldn't get it to work. It was supposed to move or rather flow around the corners of a tile based level, if possible in all directions, meaning downwards too. A simple wavy but level water line was something I was trying to avoid. One of the first (~3) tries of which I took footage, based on multiple tile based models:
Realized a water line for the first time, sadly with issues and the limitations of not being able to really let flow water left and right or even downwards:
I was trying to use the calculating "normal" direction to realize current, foam and such, but with less than desirable results:
Getting the Piranhas to work was a whole different thing, too (they are still far from perfect):
(EDIT: Oh and there was another again tile model based approach where I would deform high poly meshes so it would look like the water had breaking waves (towards the camera). It was by far the best looking stuff but I had huge problems with connecting the tiles at different water levels nicely.)
Then someone told me about a thing called Metaballs which basically are distances to positions (like the low res tile map in my level) summed and compared to a threshold, resulting in liquified contours. Final result:
I'm not 100% satisfied but I say it's good enough to move on. I spent way too much time on this level but at some point I just had invested too much already to give it up completely. I think the metaball concept can be expanded into many more effects, maybe even gameplay related stuff!
EDIT 2: The final result is just one big 2face quad mesh across the whole level, no need to place all those small tile models manually.
Today i started to create a game without a plan but with the goal to answer a lot of questions about basic coding in Lite_c.
While i create i write down all i do and explain the scripts used.
So far i've got :
- Player simple movement with animations. - Teleporting from a to b and back again. - Losing lives in graphics and numbers. - Restarting a game. - Using variables the full way. - Enemy bouncers that kill on touch. - the use of Emask. - Picking up stuff. - Doors. - Collecting keys and use them to open doors - Using a timer - Exit level and load next - Using animates sprites - Using particle trails - Creating fire particles - Adding simple gravity (move up and down slopes) - Moving platforms (Elevator) - Throwing/shooting/blasting bullets or spells - Exploding stuff when shot - Add and use ingame music - Getting more spells
As i am writing the doc new ideas come to me and will be inserted.
it is a nice little demo , add some spot shadows under the pickups and the black "enemy" ball , in that perspective it is hard to guess exactly if the player is in line with the objects without shadows ,just a suggestion.
@superku , i am always amazed at the effects you come up with for your game , it is so much part of your game style
Some more screenshots of the finished version. At day time you have to collect wood and snowballs and at night you have to defend it against the evil gnomes in a first person shooter like mode.
I found some more wait(1) issues which caused particles and shaders to double up. Now all is left to find out why the trees are so extremely performance greedy and I will release another version with improved framerate.
As I kept telling you: lots of overdraw. Not much I can do about that besides using different tree models. Maybe just the same but making sure polygons are ordered top to bottom or something.
I created a less detailed version of the tree using smaller textures and half the face number. Hardly noticable in the game. Still some slowdowns here and there, but framerate ismuch more stable. Uploading fixed version now, same link as before.
@PadMalcolm: Its simple you just create an entity in WED or during runtime with some parameters for the lighting (5 types) and for entities you just call a function with some properties and you get a material. It just writes into a GBuffer so all I add to it will be tied to a unified approach - so if that is ok then you can throw it in your game, yes. On the other hand side - compared to what I did before - I only develop for my own needs, so everything is subject to change I want to add glowmaps, static lightmaps/AO to the light/shadowing at least, then it would be a complete approach to what I want to have: everything between all dynamic and all static DOF/Tiltshift is planned after I integrated it into my reflection demo. I try to build something that works for my next 1..3 games with Gamestudio so I'll try to decouple it and, yes, it should be easy to use at least I hope so
@FLE: some fancy stuff is planned, yes. That whole VSM stuff and that light accumulation is already making my pants wet, though.
@Kartoffel: any chance for seeing that Bokeh DOF code? Or was it Hummel who talked about that? However, thanks!
Yeah we did, Slin, Firo, my brother and I took part to advertise Gamestudio Btw we scored 2nd best game in "Best adaption of the topic winter dreams".
Yes, Devmania was great. The fun part is, I only actually tried playing the game full length when I was back home. Before I was only do short tests and debugging. I'm surprised that it's actually fun to play and has a nice rise in difficulty. Seems like slin or PadMalcom or both did a great job in balancing spawn rates and speed while I took a nap
Hi! Yesterday night I knew a cool board game and decided to build it with liteC because is really funny. Its name is Carcassonne. I found it difficult to perform because of complicated relationships between pieces and definitively it is getting me mad. Here are the rules.
With no lack of effort I just finished pieces location. There are so much exceptions that I needed to restart it four times because the initially planned structure failed on the road. I hope I am able to finish it with actual data approach...
What do you think? Even though the game is not as ready as shown in the trailer ^^ Still some polishing needed.
* good graphics for indie (though the ground texture in the hollow sphere-kind map at 0:36 could be improved a bit) * features like 'Epic Story' and 'Team Battles' show only like a second of gameplay before switching to the next feature, so its not clear (to me atleast) what they really add and perhaps more important its easy to overlook them. * does it have bots, even if they only serve to fill up servers a bit?
The cut is a bit too fast. Give the people a bit more time to get the sellingpoint of your game (crazy movement).
When will you release it? I will buy a copy. Always wanted to play something like that. Nice disorienting shooter using all dimensions. Great work.
Yeah, I thought that maybe if I cut it fast it will show how fast the game itself is. Release is still not in sight. Soon I´m going to release 1.0b and put that on Steam Greenlight to see how much interest there is. Depending on that I´ll continue with the development.
Originally Posted By: Reconnoiter
* good graphics for indie (though the ground texture in the hollow sphere-kind map at 0:36 could be improved a bit) * features like 'Epic Story' and 'Team Battles' show only like a second of gameplay before switching to the next feature, so its not clear (to me atleast) what they really add and perhaps more important its easy to overlook them. * does it have bots, even if they only serve to fill up servers a bit?
Thanks for your comment! Yeah as stated, I tried to get an action-like trailer (But I´ll consider that point when I redo the trailer) Yes, it does have bots, but only offline at the moment. Online bots are on my To-Do list though and will be implemented in the 1.0b version.
cool that you will have bot support, I think I will buy it too . If it will have enough weapons/replayability I might be able to convince some others too.
Quote:
I tried to get an action-like trailer grin (But I´ll consider that point when I redo the trailer)
, perhaps do first 50% ingame action / cool shots (50 sec?) than something like 5 main features with each some ingame footage/shot? (6-10 sec per feature?)
Posted By: Anonymous
Re: What are you working on? - 11/05/1503:52
Maybe for me personally - Everyone post you current game project and estimated Release/greenlight date.
Also for my personal curiosity- Out of the known forum member currently active please post your dream team to work with and the roles you'd assign.
Thanks Mal
p.s. I f needed I can post screens from my current and last 10 worked on projects to meet the requirements of the thread.
I just managed to create a somewhat working first person playercode (finally)
In all my previous attempts something went horribly wrong when I tried to implement stuff like max walkable gound-angles or being able to traverse actual stairs (movement between higher and lowere surfaces without a smooth transition). Some weird things happened like the player constantly bouncing up and down a few centimeters or flying up into the sky when hitting an object at the wrong angle (yes I'm talking about you c_trace(, , USE_BOX) ).
...Acknex's collision system was a pain in the ass to deal with but I could somehow work my way around it.
Right now there's pretty solid walking and jumping mechanics, you don't fall through small holes in the ground, you can walk stairs and other objects with small height differences without flying all over the place and you slide off surfaces that are too steep to stand on while correctly keeping the velocity you had before 'hitting' the surface. The stair thing was the most annoying one... c_trace with USE_BOX doesn't give you the exact ground normal and you needed to take a run-up when trying to go up some stairs . However, I kind of fixed it and I hope it's solid enough now.
The next thing to implement would be standing on moving or rotating objects aswell as jumping off of them and the way the player's velocity is retained while in the air and maybe some quake-like bunnyhopping mechanics .
Well, long post and mostly ranting... Maybe I'll upload a small demo soon.
Yes, please share a small demo for us to test it. As you may already know, I tried to create solid first person movement by myself too, so I'm really interested to see your results as well!
..aaand collisions & traces are f*cking me up again.
The cylinder is the player, I'm tracing straight down with USE_BOX and the glowing dot shows where the trace hits the "ground". This completely messes up the ground normal and it's a pain in the ass to try find a workaround (the one I found doesn't work in all cases).
On top of that the player can easily get stuck in thight corners (angle < 45°) because the trace somehow ends up in the wall again. And apparently it matters whether the collision is happening in x+ or x- direction because I sometimes get a ground angle of 90° (same with y).
The only solution I can think of are different models for the traces to get a smaller or a bigger volume to trace with but I believe this will lead to even more problems.
Another possibility would be using a physx actor but I'm not sure if I want to do that.
The ellipsoid collision model just plain out sucks. It has always been a pain to deal with and always will be. In two dimensions you can find some workarounds - although this gets a lot more tricky with rotated models - but in 3D this gets a lot harder.
What you can do for the gravity trace is make min_z and max_z (in absolute terms) really small, then perform the USE_BOX vertical trace and undo those bbox changes. This will make the downward trace rather box-like or cylindrical, and will give you less stupid target and normal results.
I think I managed to fix the issues I talked about.
Also, standing on moving platforms now works, aswell as Jumping off of them. If you stand on something that moves, for example, and jump straight up you still move in the same direction as before instead of suddenly standing still in the air and falling off (I hate games that with mechanics like this).
Posted By: Anonymous
Re: What are you working on? - 11/08/1520:31
Quote:
If you stand on something that moves, for example, and jump straight up you still move in the same direction as before instead of suddenly standing still in the air and falling off (I hate games that with mechanics like this).
Yes even in the code examples we have here, there is no though of a added vector for inherited movement vec of the surface.
Ed- Is anyone currently thinking of doing a footsteps-like bone driven animation system.
System that allow active modification of the animation do so by adjusting the base hard coded bone animation with factors such as the ground height and slope.
I'd hate to try to start a add-on if someone with much great skill is already on it.
@Iglarion the art is beautiful. Is that your talent, A team-mate's or store bought shop art? You've been working this project quite some time now, I recall the first 'Showcase' posts.
Thank you txesmi and Malice. Complete art coming from my friend Markus and me, and how our skills have progressed we improved the graphics in game. And yes, i working a really long on this game, the problem is that the last two years i work alone on project and it's a painful job to paint, modeling, animate and code .
Posted By: Anonymous
Re: What are you working on? - 11/13/1520:45
Quote:
it's a painful job to paint, modeling, animate and code
A little tease of the upcoming Multiplayer (Shooter) Template in the new AUM (complete rewrite compared with the last version, started with a blank script):
28 bots in that shot, all shooting like crazy, as well as the server client and regular client playing.
A little tease of the upcoming Multiplayer (Shooter) Template in the new AUM (complete rewrite compared with the last version, started with a blank script):
28 bots in that shot, all shooting like crazy, as well as the server client and regular client playing.
I am working on my addictive game Endless. I got almost all basics working planning to add more special boost weapons and ships + suprises to unlock enjoy and hate playing my own game gheheheh luckely i can cheat to test stuff.
@Turrican: wow that is pretty cool. Can the game be completed by newbies? (I am personally not so good at sidescroller space games but I do find the game appealing)
@Reconnoiter: Thanks! Yes, I'm aiming to make the game as fair and "learnable" as possible. Playing it through might require some practice though, but that's what shoot'em ups are all about anyway.
That is true, but will there be difficulty options or perhaps a simple cheat, incase one cannot finish a single map? Quite alot of times I played cool side scroller games but got stuck somewhere midgame e.g. cause of a boss or minigame or such. Its just annoying when it happens.
Difficulty options are planned and already prepared to be integrated. And I'm seriously playing with the idea of implementing a real beginner mode, as I'm constantly getting questions about the difficulty of the game. Many people really seem to be scared by shmups just by their looks, maybe because they instantly think of the hard-as-hell arcade games from the 80s/90s or japanese danmaku/bullet hell games. RFA should hopefully take a different route and offer good playability without being boring/unchallenging.
Difficulty options are planned and already prepared to be integrated. And I'm seriously playing with the idea of implementing a real beginner mode, as I'm constantly getting questions about the difficulty of the game. Many people really seem to be scared by shmups just by their looks, maybe because they instantly think of the hard-as-hell arcade games from the 80s/90s or japanese danmaku/bullet hell games. RFA should hopefully take a different route and offer good playability without being boring/unchallenging.
, great! For me personally I am often wary of difficulty in 2d games cause of the limited evasion (/only 2 dimensions) and one hit is one kill games.
I'm working on a pixel shader that uses a flowmap for water. Since it is more or less working the next steps are: -Resolve the pulsating effect -Improved perturbation of the water. It'll keep me busy researching and tweaking, but I'm enjoying it.
Working on a fighting game workshop its almost finished and has all the basics to create a mortal kombat/killer instinct/tekken soul calibur kind of game
I'm currently writing a custom 2D sprite system to be used for 2d games or interface rendering. At the moment I'm aming for plain graphics without shaders, although I'm planning to add some optional ones if my rendering method supports it.
The basic render code is kind of done and works with acceptable performance. Rendering 2000 sprites takes around 1.0 to 1.2 milliseconds on my machine, which isn't that fast but I think it's good enough for most 2d games.
The next thing is most likely going to be a collision system (and basic physics) or some extensions to be able to use this system to create a basic interface (windows, buttons, etc.).
I might post some screenshot's in the future if there's something worth seeing.
I had a crazy idea 2 days ago: Create a profiler for lite-c(measuring time that functions need to execute). Even tough I'm no more developing in 3dgs, it seemed interesting. Soo today I got a simple version working.
To use it, you just need to include my c file anywhere inside your project:
Php Code:
#include "AKProfiler.c"
and it will log all function calls/waits into a file:
How it works: AKProfiler.c has a startup function that execute immediatly when the engine starts and stops the game. It then looks for the main file of your project(First command line parameter). Next it parses that file and all files that are included into that file. It creates new temporary files for all these files with profiler functions injected at function starts/ends and wait's. Then it executes the main temporary file in a new acknex process. That process is now your game with profiling functions included in the scripts. When the process ends, it deletes all temporary files.
It's magic!
I'm still not sure how to exactly represent the profiled information. Currently I'm importing the file into excel and generate some charts(pie chart of functions that needed the most time f.i.).
Maybe I will relase a first version this evening if anyone is interested.
Maybe I will relase a first version this evening if anyone is interested.
This might be very useful tool for debuging! Keep it up man. But you only have to make sure that it doesn't skip anything, otherways it will be useless at all.
The map editor for my RPG Creator (YRPG-Toolkit) reached some milestones.
- Tile based map creation with the possibility to use normal maps. - Different predefined tiles - Single placement - Fill Areas with tiles - Map Properties customization (music, sound, enemy groups, map size) - Individual map format, created for YRPG Toolkit (*.ymap). - A8 standard shaders for entities (Waving Gras, normal map) - Usability changes (marked tiles, entities) - Skies
HellThunder@ that looks outstanding! I'm looking forward to test it, is it going to be free to use? Or is it for your private project only? Keep it up anyway, it looks so great!
About the project... Actually this is a commercial project. But I really thought about releasing different versions - a free version would own just limited functionality.
Old Camera Tool - more or less just an engine window:
BlackJack@ I really loved the style you have in your game, it's awesome! Models, levels, sprites and effect they looks so cool! But you really have to consider on making everything more smooth! F.e. add blending between animations (it really don't fit overall quality of the game, when NPC switch from IDLE to TALK animations etc), add recoil for weapons (to make them look more powerful, double-barrelled shotgun seems to be really weak to me... also there is that strange lag when you shoot enemy and it spawns blood particles and disappears, maybe you could also add ragdolls!), add draw/holster animations for weapons (simply morphing/swaping models also doesn't really fit the overall quality of the game), footsteps seems to be out of sync with the camera bob effect (the bob effect itself sometimes looks like the player is actually limping as if he was hurt). Such simple things will significantly improve the way game looks and feels! Anyway, I'm really happy to see such projects going on with Acknex engine! Please keep us updated, or even create a thread for your project, I'm pretty sure forum members will be really interested in some project details etc.
you have a really good contemplation in terms of our project ok well most of it is known to us that with the npc animations do I look like I solve it best to ask I will soon use this forum to solve the problems the project imagine I would , however, only dan whom the project so far as is currently located there indeed in development
and I need to improve my knowledge of english.
where we require assistance , is the area reserved shader since we have only recently become the com edition it really difficult for us to deal with it
should see even more interest for the project have , I have in my banner a link to the facebook page.
@Reconnoiter Are you serious ? this is the 4th version of the weapon speed switch system , but this is good for happiness , the fast selection of weapons, to clear selection you have the inventory system ps .. which must face are smoothed exactly ? the shotgun I immediately fixed because of your kritick
thanx guys.
here is the real inventory system , when you dont like the fast switch system
we hafe 2 types of switch system´s for the inventory 1th: is the fast switch of the weapons the 2th : type iu the real switch system from the weapons and the items .
ps here can you see the first version from the fast switch menu
@Reconnoiter Are you serious ? this is the 4th version of the weapon speed switch system , but this is good for happiness , the fast selection of weapons, to clear selection you have the inventory system ps .. which must face are smoothed exactly ? the shotgun I immediately fixed because of your kritick
, is this game commercial or is it free? To put it direct; whether it will be commercial or free the gameplay seems like fun, but if it is commercial than the graphics look a bit dated (stiff/wooden animations namely. e.g. the animations of the enemy monsters).
As for the reloading animation (please note the weapon switiching system is fine, I am talking purely about the visuals), the idea of the factory clock thingy is not bad its just covers way to much screen. So I think if you decrease its scale a bit it will look better. Also just saw your older video, I think the reloading sprite is their better since it is transparent (so it covers less view). Perhaps the current reload sprite could be used for a game saving icon? (/decrease its scale, let it rotate and put it in a corner and I think it will look good )
BlackJack@ I agree with all that Reconnoiter said. Perhaps you could take a look at the way weapon changing cycle made in DooM 4.
Quote:
DooM4:
Your older version:
Your newer version:
As you cans see, in your case using model (animated) looks pretty nice, but it's not useful at all, in fact it actually disturbs you, by blocking most part of the screen! So while you enjoying all the fancy look of animated cycling things, enemies can get by you and punch all the health points out of you Look how it's made in DooM4, TRANSLUCENT! and the main part of the screen is not blocked (even if it's blurred). Anyway, it's your game, keep it up! It has some nice looking style (borderland's style), but I really hope it won't be one of those projects that nowadays made, where visuals are pretty nice, but the gameplay part is not that good.
FLD@ it looks good, but it's not made with physX, am I right? I'm saying that, cause to my taste the way it climbed up the hill was a bit not realistic.
oh no ^^ this is a free Game we are all hobby developer about the team can i say My team is currently about 5 people , we have a musician, a sound-desinger , me of as model and textur-maker and 2 coder then 3 dubber fan as synchronous speaker, and i know for this stiff/wooden graphics and i really hope i can fixit with some shaders, 3 years ago we used shade-c but it was really buggy. and yess you have really right with the weapon switch system, so i worked of a new desing , i make it smaller and better.
for that I can say I am working on a 7nd opponent we have another enemy type , a mix of destroier and Berserker.
i hope do you like it
oh yeah , the shotgun has also been fixed , they besizt a new animation and the recoil was revised
@txesmi LOD for terrains is a A8 PRO feature, right? Does your system work with a lower version?
I really like the normals and the lighting. Would love to see some decoration on the ground - some cactus plants or mountains. Would be interesting to see a much larger map and the resulting fps.
----------
Here some impressions of a new tool of YRPG Toolkit - Quantum Manager aka Particle Editor A Quantum Scenario will be added with the Event Editor and can be placed on a certain camera position (created with YRPG Camera Management Tool) or an entity of the map editor.
Yes. It is a model and a shader so it should work in commercial version too. It really has nothing to do with 3dgs terrains.
Captured scene takes around 0.3ms to render on a NvidiaGTX660 but it computes pixel normals in real time so backed normals will speed it up a lot, I guess.
Next step is to generate some tiled nm+hm in order to build an infinite system but L3DT does not finish the process after hours of computations Does somebody know any other free program with erosion and so?
@HellThunder looks great i like it . just some more updates from me. and here a another video ^^ sorry for the stuttering , and the poor cut of the video.that is not my fault : D ! https://www.youtube.com/watch?v=gjMsvfq1hVA
I know it is an old abandoned library. But it works it is relative lightweight. acknex is old too, so it fits.
All gwen objects are accessable through interfaces for example (window->base)->setSize(100, 100), window->setTitle("Title") etc. All base element mthods are accessable through the window->base struct. I don't need to redefine every base method this way. The real c++ class object is saved as a void* in a lite-c insterface struct. The interface points to methods like this->base->setsize = gwen_base_setSize; This method calls the real DLL wrapper method.
I had to extend the gwen sourcecode, to use lite-c methods in event calls.
So far this works very well. At the and it is easy to use and a very plain easy system. But it is a pain in the ass to wrap all of the gui objects this way.
yes I will share it. But my progress is not very far right now. There are ca. 30 (complex) Elements to wrap and I have 6 right now. Then I need to wrap the skin system. The general implementation is done. I will create a small video this evening.
yes I will share it. But my progress is not very far right now. There are ca. 30 (complex) Elements to wrap and I have 6 right now. Then I need to wrap the skin system. The general implementation is done. I will create a small video this evening.
, well in any case, good luck with it , and look forward to the video.
So I have been working with Superku and George's amazing multiplayer system. I have been trying to make the system feel more "real" and polished by tightening up the movement and updating the UI/graphics. I am no artist, and map design is not my forte. My one and only skill is programming and scripting, so here is what I have so far;
I would like to bring more people on board for a project, specifically I would like to make this the "end all" multiplayer template/example. If anyone is interested in contributing please feel free to shoot me a PM.
@exile: This is looking really cool! I like what you have done to it and the level. Only the FOV/ camera.arc is a little low for my taste. I think I may have unpublished changes regarding the template somewhere but I assume you've changed way too much in the code already for that to be worth it (?).
Kartoffel, man what I would to have shader skills like yours. Would make so many of my own projects actually realistic lmao. Great looking work dude!
Here is another video showing some new things I added into the game. Namely, basic weapon switch animations, clambering (climbing ledges), tightened up the double jump, and a few minor improvements to the movement.
(Note there is a bit of an audio delay for some reason, seems to be a youtube thing since the delay is not present in the source video) https://youtu.be/6gBFrvr-B0o
Feedback is always welcome. Just wanted to note, the weapons are either free models I found online or they are ripped from UT2004. They have not been created by me and any credit to the models should go towards the awesome team at Epic.
Man, it looks even better than it was! It reminded me of Xspire AI, but it looks more dynamic and intense. I would suggest you to add ragdolls with blending (I made a contribution of converted into phyX version)! Same for the player, imagine him being killed by explosion and tossed away with camera being attached to the head bone (this will look really awesome I can imagine ). And it's really sad to see that the audio is messed up with delay, I guess it looks much better in the source video.
BTW, about weapon models, as far as I can recognize them, they are from FPS Creator original weapons, but I might be wrong.. I'm not sure about license, so I don't know if you can use them for free or not.
If you will need someone to test this out, please let me know! I found plenty of bugs back in original version (while playing it with friends).
BTW, about weapon models, as far as I can recognize them, they are from FPS Creator original weapons, but I might be wrong.. I'm not sure about license, so I don't know if you can use them for free or not.
I reckognize most of them, these are from unreal torunament 2k4.
BTW, about weapon models, as far as I can recognize them, they are from FPS Creator original weapons, but I might be wrong.. I'm not sure about license, so I don't know if you can use them for free or not.
I reckognize most of them, these are from unreal torunament 2k4.
Most doesn't mean all F.e. this one is from FPS Creator (Assault rifle) for sure! I recognized it from the first look.
Lol yup youre both right. I got the shotgun model from another website, t3dfm I believe. I wish I was a good modeler or animator but I guess you cant be good at everything lol. Thanks for the words gentlemen, and 3run, I think you will look forward to the next video. I am in the middle of developing a gore engine which I will implement in the next update.
Lol yup youre both right. I got the shotgun model from another website, t3dfm I believe. I wish I was a good modeler or animator but I guess you cant be good at everything lol. Thanks for the words gentlemen, and 3run, I think you will look forward to the next video. I am in the middle of developing a gore engine which I will implement in the next update.
After some paused time iam back with an simple FPS project. Working some weeks now on it. Made a Project-thread too btw. Hope u like it so far. Goal is fast arcade with high FPS. Heres a WIP of the weapon System ( Video linking is not working sry )
BlackJack@ it looks pretty nice, but I have few advices: 1 - disappearing monsters after death with some blood splashes doesn't really look that well.. add at least some body parts etc (add death animation or ragdoll would be the best). 2 - I don't really like the recoil you have for your weapons, it looks more like an explosion than actual weapon recoil, but it's just my taste.
@rayp, why ironsights with fast arcade gameplay? Otherwise looks fun.
@BlackJack, cool. The thing that annoys me is the constant shaking cam when firing, better to reduce that or maybe add an option to change that it in the game options (e.g. to reduce it by ~50%). Also might be an idea to add normal/bump maps to your models for better lightning-looks.
all right , I have the recoil of weapons revised berreits repeatedly it would be good wen you could me make an example .
the gore factor is still expanding . there are actually panned the many flying around body parts .
the normal map I can only say it is a pity that provides game studio there. the standart game studio normal maps are not supported by the static lighting . which is quite a shame .
shade- c has the problem that is uses its own zbuffer and unripe it
I have the recoil of weapons revised berreits repeatedly it would be good wen you could me make an example
I could only advice you to change more TILT and PAN randomly (don't touch ROLL), depending on the weapon (also note that you have to move TILT more than you spin PAN, cause recoil should go upwards but with a little sideways movement). Set max value for TILT and PAN (f.e. for TILT 10 and for PAN 3), and randomly set their values within the range (like 0 and MAX for TILT and -3 and 3 for PAN). After each change, lerp back to zero (or lerp only when stop shooting). Add those results to camera movement and you are good to go.
@BlackJack, for the recoil I think 3run's idea is good. For normal map, that is a pity indeed. I still like the graphics. Perhaps you can spice things up a bit by adding a little bit extra shininess/hdr effect at lights and metal-like parts?
...very handy function The following Video shows a Explosion init with fire burning and a Zombie - arise ( arise from floor ). I can say "PEB" - tool is great! https://youtu.be/fiJEz2Llh_A
Sry i dont understand the question ( english-german ) ?
, Dunglish . What I meant was, you mentioned it is fast arcade gameplay and in a video I noticed you used iron sights when firing some gun. Iron sights seem to be more used for shooters like Call of Duty and not so much for fast arcade fps games (where "hip fire" is more used). Not really important though, just some detail I noticed. Perhaps a good idea to make the iron sights not toggled though, if you haven't done already.
You mean the "empty" bullets flying out ? If so there part of the shoot animation And yes...they would not be very fast ( flying bullet ents ). If you meant something else, sry iam tired
Nothing against iron sights in arcade shooters, though.
, I like it when its not toggled but hold, like sniper rifles in e.g. UT. That way you can switch fast enough between e.g. hip fire and a quick headshot. Iron sight on itself is not bad, its just that the games that have them make hip fire so bad.
So i implemented it. You can hold mouse right and Switch in realtime while firing. Today i want to implement some gore for the working pumpaction shotgun edit: Already featured: Zombies fly away when hit near by shotgun ^^ edit2: But cause of the still existing bug i CANT use ent_animatefrom with ent_morph ( in combination with decals of course ). That sucks! JCL said the bug will be fixed but...nope! In this case my damage models cant use the animations of the Zombie models. Makes me sad and angry, this is a mighty bug for my taste. And we all know that it might stay forever now...
edit3 - Bug example ( for code example check out Bug-Hunt section ) ( not all of us are common with shader programming. means those of use [ like me ] cant fix such probs by script on their own ) Situation () You want blood decals on your ents ( <-of course. done via [c_trace-ent_decal] ) () You have for example damage models ( ent_morph ) To be (a bit) efficient you would do the following. Share your animations for your entitys via ent_animatefrom. Lets say we have 2000 frames in our model. Means without the bug we could have hunderts of damage-models with only 2000 frames. With bug we have to add 2000 Frames for every damage-model. Heavy resource limitation. Maybe some shader dude would be so kind and upload a ent_decal "replacement"/workaround ? Would be nice.
@Kartoffel The Ramboanimation is really cool. The Impression of the face...mega cool Looks like his Body is firing the bullets Theres another funny Scene in Rambo. When he fires the bazooke out of the helicopter. Funny how his hair moves from the rocket air
The second Scene of the first link, short before the end! YES! lol
if i remember right theres another hair-scene with Bazooka and a heli...
Quote:
(I will stop spamming now)
Me too
edit: First Video of "gameplay", better said first Show of script playing all together ( music is not mine / placeholder, a friend will make some new those days ) https://youtu.be/4MWpSExyZio
edit: Just wanted to show this screenshot of my old shadeC project i found yesterday
Wow, all of a sudden everyone is coming out of the woodwork with their own projects. I love seeing this community active, it inspires me to keep going more than anything else. Seriously, I have been using 3DGS for years as a simple hobbyist and every time I see the awesome work from other members it brings a tear of pride to this old lumberjacks eye lol. But back on track, compiled a small video of the state of the game now. I'm at the point where I have made my own sounds, my own HUD and visual effects, and I already have the music done (not showing it off just yet). Take a look and leave any feedback you can think of.
exile@ it looks epic! I love those gore effects you've made, they look awesome. The only thing that I can tell is the explosion effect, it still doesn't look 'powerful' enough. Anyway, this is one of the best (if not the best) multiplayer fps this forum has seen in awhile!
Created my own alternative to Helghast's ellipsoid collision
The problem with Helghat's snippet was, that removing entities used to cause pointer errors, and I couldn't really figer out how to remove them from structure list. So I decided to create my own version based on his simple idea, to get the closest NPC and push it away, with few improvement. I handle everything (all npcs) in one loop (thank's to Superku for his wonderful idea) and I check for special skill NPC_PUSH to make pushable 'level' + only NPC with FLAG8 set can push others. I'm planning to use this in one of my future projects, where bigger enemies will make their way straight to the player through the crowd
First "gameplay"/"killthrough" video ( running around killing some Zombies ) https://youtu.be/U4xX_Te7BIY ( At +- minute 1 of the video u can see that Zombies can standup again when fallen to ground )
edit: first camera/Player intro https://youtu.be/aB8f1W_el6U ( camera starts to far behind player model - fixed already )
looks really cool! at 1:39 there is a shadow error... do you use a lightmap with stencil shadows? it is better than the earlier dots shadows, but a nice shadowmapping would improve the nice artist work.
Well here is the last bit of gameplay footage for a while. At this point I am in need of artists since my artistic skills don't even belong on mom and dad's fridge I will be in need of someone who can animate, model, and texture. Music, sound, visual effects and gameplay I can take care of, but the art department is seriously lacking. Until I get an artist there isn't much I can do at this point. This is the problem I always run into, I make a good looking project only to be stonewalled by the lack of artistic assets. Ah well, enjoy guys!
Well here is the last bit of gameplay footage for a while. At this point I am in need of artists since my artistic skills don't even belong on mom and dad's fridge I will be in need of someone who can animate, model, and texture. Music, sound, visual effects and gameplay I can take care of, but the art department is seriously lacking. Until I get an artist there isn't much I can do at this point. This is the problem I always run into, I make a good looking project only to be stonewalled by the lack of artistic assets. Ah well, enjoy guys!
, Maybe you could go for a simplistic graphics style? I currently like the colorfull lightning of the level. I could see adding some lines etc. with some perhaps some neon/vivid colors could work well. For weapons you could than go for a simplistic futuristic style with some nice lights/lightning. Good animations would still be a bitch ofcourse. Anyway gl with it, looks good.
Working on some lens flares (converting them from C-Script project and fixing pointer and other crappy issues):
Special thanks to 'Prete aka Priest'! He is the one who created those lens flares, I've just converted them into the LITE-C and adjusted a little bit.
Edit: changed screen! Edit2: lights are destructible, all lens flares get removed etc Edit3: Download link (UPDATED LINK! VERSION WITH MULTIPLY LENS FLARES!)
probably not the best idea to use a game engine for that stuff but whatever... the maximum fps is set pretty low when no animations are playing so it's not too bad
Cause it was a lot of work i have shown nothing off, i want2 do this now. Some fast taken vids from my (for now) frozen ShadeC Evo Project The small "Errors" like flashlight "light/glow" tex are only some optic things not done yet.
And you can see "interaction with Zombies" is much more intense. They grab you and you must shake the mouse. Ive to say, i really love this script https://youtu.be/a6QxJaqmCsM
Well this was some totally different style of game like the "Simply Dead" Project. Ill make some good vids some day (soon).
Here is the first public build of the game, I am currently developing a "project" build and an "Open" build. The open build will contain the source code and basic features as an expansion on Superku's amazing multiplayer system. The open build is designed to give the user all of the features of a modern FPS. The Project build is an actual FPS project I am making based on this system, sort of as a proof of concept. You can download the game using the link below. Let me know what you think.
Here is the first public build of the game, I am currently developing a "project" build and an "Open" build. The open build will contain the source code and basic features as an expansion on Superku's amazing multiplayer system. The open build is designed to give the user all of the features of a modern FPS. The Project build is an actual FPS project I am making based on this system, sort of as a proof of concept. You can download the game using the link below. Let me know what you think.
However, I don't like packed .exe files like this one. They increase the startup time, you don't know what stuff it contains and can't be sure if the program will clean up all the mess that it left somewhere on your hdd. Besides that, they don't really prevent other people from looking into your project's files. Although I know that in this case it's made using the old gs-packer, I generally don't open these files.
Ah gotcha, only reason why I used gs-packer was primarily because it's one self contained file. I didn't really care about the content being "hidden" since I know gs-packer can be hacked with trivial effort. Here is an updated version which is not packed into one exe file.
Extremely and unforgivably slow on my machine (AMD X4 641 QuadCore 2.80, 4 GB RAM, AMD Radeon HD6670), couldn't even play it (fps below 15-20). Is there any ways to disable shade-c?
@3Run, Hmm well I am actually not using shade-c at all. Does it start off slow (Before player control enabled) or does it get progressively slow as time passes? I will upload a version with the shaders disabled, I am also working on a menu system which will help tweak visual options. As a comparison, I get 60fps on my htpc (AMD A4-5000, 6G RAM, and an AMD R7-200). I will actually build a PC meant for testing based on your specs, I actually have all the necessary parts just lying around lol. Thanks for the info!
I thought you used shade-c, when I saw all those 'pp' .fx files, but now I can see that most of them are just from native shader collection. Nah, it's nothing related to the script part (it just slows down as soon as everything got 'drawn'), it's just visuals (shaders 100%). Let's see how 'non-shader' version will go on. Plus, I couldn't get LAN to run, but I guess it's just not supported yet.
On my machine it runs quite smoothly: - Intel i7 4770K 4x3.5 GHz (+Hyperthreading) - AMD Radeon HD7850 - enough RAM (shouldn't be a problem anyway)
My CPU is probably above average but I think my GPU should be quite a bit worse than what most people (gamers) have nowadays.
I didn't play the demo for very long, yet, but this is what I noticed:
> the input seems a bit laggy, might be vsync, delay caused by windowed mode or it's engine related > like in superku's demo, there still is mouse_force.xy used for the camera's rotation. I highly recommend you to use mickey.xy instead. This one doesn't cap at ±1. The values are a lot higher, though, so you have to scale it down a bit.
Hmm I need to look into that, I am able to run LAN fine from two of my test machines without any problems so far. Let me see if I can get a more "stable" version working.
EDIT: Instead of cluttering this thread with my stuff I am going to make my own project thread.
mouse_force.xy used for the camera's rotation. I highly recommend you to use mickey.xy instead. This one doesn't cap at ±1.
, is this an important issue with mouse_force? Cause I too only used mickey.z and have problems with it (the value gets stuck sometimes, than it doesn't return to 0). Could be its a specific problem with my laptop cause I didn't had it on other computers. I am wondering if mickey.xy could give that problem to.
@3Run, pretty good, I mainly like the last weapon even though I dont know what it is meant to be .
Quote:
Edit: oh, but I know that keys remain pressed in an 3dgs application when changing to a different window and back or tabbing out and in again.
, that could be it, since it is was in both cases with a windowed game iirc. Perhaps something in the background like an update popup than caused windows to switch tabs or such. I know I only have had it with mickey.z, but that is maybe caused mouse scrolling sometimes keeps rolling a very short while after releasing it instead of instantly stopping like releasing a keyboard key.
Thank you guys! The last weapon is going to throw spinning circular saw blade I just done idle and walking animations for the alien, I'll create another thread for my current 8 bit project as soon as I'll have something more to show!
I finished my game battle brix all is done and error free working accept online playing i need to dive into that shit
Battle brix is a fun uptempo mix of break out,pong,arkanoid and other old school retro games thrown together See how fast you can destroy your enemies bricks! The soundtrack is created by moi and astronaute voice and voice in soundtrack jup its me
Thank you for an idea, I already knew about such tools, and I even do own one. But thing is, that those tools are used for animating vector characters , and they aren't suitable for low res. pixelated animations. So no easy way for me, but it's fun anyway.
Kartofel, that looks very nice! BTW it reminded me of the shader pipeline that was made by txesmi, take a look at this video: Ligths in the fog
Kartofel, that looks very nice! BTW it reminded me of the shader pipeline that was made by txesmi, take a look at this video: Ligths in the fog
cool stuff! Pretty sure he's using the same (or a very similar) method as me. But he most likely did some good optimizations.
In my case it's kind of a bruteforce approach... the lightrays in the screenshot I've posted increase the sample count from 6-9 (light + shadow) to 206 (light + shadow + lightrays) samples per pixel ...per light ;_;
That's why a single light at a 1600x900 resolution causes the framerate to crash down to 50-60 fps. On second thought, however, it should be waaay faster with some clever optimazations like using a lower resolution for the lightrays and adding dithering (greatly decreasing the sample count at the expense of some visual noise)
acknex's rendering is just horribly slow, and there's almost no flexibility so I can't really bypass this and do things my own way. On top of that, good old (antique) directx 9 is not only a bit slower than newer versions but I'm also missing a few features to improve performance...
Thank you for an idea, I already knew about such tools, and I even do own one. But thing is, that those tools are used for animating vector characters , and they aren't suitable for low res. pixelated animations. So no easy way for me, but it's fun anyway.
I used my software package to do this.I think I remember the process as - I create the vector layer image in a high resolution. - bone and animate it. - export to image files(frame to image) - Then batch scale down the image files in gimp.
Don't get my lying as I don't recall from so long ago. I'm not sure the total outcome or the tweaks and workaround's that I used.
I always worked and lager resolution and dpi and scaled the end result, just because I couldn't work in tiny resolutions.
Posted By: Anonymous
Re: What are you working on? - 06/18/1606:02
Not sure this counts as a "What are you working on.." post.
I've spent about a week(less as I didn't work solely on this) creating a qBezier Curve system and movement along it.
I'm not a math guy so this was a lot of work for me to puzzle out. Well not the curve, I kind of got that right in the beginning. For that I had to just step out of all the crazy math and think it out mechanically. It was solving collision movement along any shaped 3-point curve, at a continual-speed of the entity(No fast takeoff or slow end), that was heck for me. Edit - So to make a blue dot travel the curve is easy, but to make it's speed synchronize with the moving entities speed was just crazy hard for me. -end edit
So I took way to much time creating a helper_function in a way that is almost humiliating...Lol
c_moveCurve
The screen-shot button and gif process makes it seem jumpy Both car and box use some simple repelling vectors to make the pushing out of way action.
@Malice: If I understood right, you're using actual movement vectors instead of placing the object at the calculated position on the curve? cool stuff.
At the moment I'm working on my 2D renderer. I just made some improvements to allow for pixelart rendering:
(too many scriptfiles ;_;)
Edit: next up will be an interface system
Posted By: Anonymous
Re: What are you working on? - 06/18/1615:51
Quote:
At the moment I'm working on my 2D renderer. I just made some improvements to allow for pixelart rendering:
Awesome! Will it be faster than the engine 2d rendering?
Quote:
@Malice: If I understood right, you're using actual movement vectors instead of placing the object at the calculated position on the curve? cool stuff.
Thank you! Yes the white entity is moving with the normal c_move on a the relative x vector. It's using vec_to_angle to look at the blue dot, which is an point on the curve somewhere around 1-2 frames ahead of the entity. Because points on the curve are found by a factor of time_to_length of 0-1 * curve_length, The entity can't simple chase a factored point. The next chase point must be found each frame.
Because t=0-1 creates different distances with different length curves, simply stepping through at a set step_interval would cause the look_to_point to move faster or slow dependent on it's place on the curve and the curve total length.
Using a standard movement vector the entity can join or exit the curve at any point and has standard collision. Because the look_to_point(bluedot) is calculated form the entity at the entities speed, even during a slowdown or collision(or being pushed backward) the look_to_point stays relatively 1-2 frames ahead of the entity at the entities movement speed.
So the entity is never attached and can move at any speed or speed variation (acceleration/deceleration).
The blue_dot is actual about the entity.xyz+entity.speed*frame.
Lol I think I'm just sound more and more confusing lol
At the moment I'm working on my 2D renderer. I just made some improvements to allow for pixelart rendering:
Awesome! Will it be faster than the engine 2d rendering?
I think so, yeah.
the rendertime for 2000 sprites is around 1.0-1.2ms on my machine. it could be faster... but performance shouldn't be a problem.
however, things like collisions are going to be a bit tricky. I'll have to write my own collision system.
Posted By: Anonymous
Re: What are you working on? - 06/18/1618:00
Quote:
the rendertime for 2000 sprites is around 1.0-1.2ms on my machine. it could be faster... but performance shouldn't be a problem.
Could it possible work as 2d particle system? Or rather 2D visual fx system..
Quote:
however, things like collisions are going to be a bit tricky. I'll have to write my own collision system.
I would assume this would just be a simple matter of modifying a collision system already available. 2D collision, after all, isn't new, and I would assume that it's "state of the art" hasn't advanced much in years.
However I assume my way into a lot of falsehoods...lol
@Malice: well it's 2d only, which means you can't really combine it with 3d rendering (if that's what you meant)
and yeah, 2d collisions aren't that hard to do, but I have zero knowledge in this field so it's gonna be a lot of trial and error.
it's gonna need a somewhat sophisticated approach, though. simple bruteforce methods (checking all objects for collisions) get extremely slow as the number of objects in the scene increases.
@Malice: well it's 2d only, which means you can't really combine it with 3d rendering (if that's what you meant)
and yeah, 2d collisions aren't that hard to do, but I have zero knowledge in this field so it's gonna be a lot of trial and error.
it's gonna need a somewhat sophisticated approach, though. simple bruteforce methods (checking all objects for collisions) get extremely slow as the number of objects in the scene increases.
Well that's why usually you have a broadphase first, where you test all pairs but with a cheaper test (distance only for example) and then only test interesting pairs in the narrow phase.
You can improve on that by putting the objects on a grid and letting the broadphase only check for potential collisions within a cell of the grid (and keeping in mind that an object can span multiple cells)
You can improve on that by putting the objects on a grid and letting the broadphase only check for potential collisions within a cell of the grid (and keeping in mind that an object can span multiple cells)
I actually tried that some time ago. It worked pretty well (sorting objects into the grid, moving them) but there were some bugs and the whole thing was dirty a.f. since it was my first try to code something like that. In the end I threw that part of the code away. However, I gained a lot of insight and my next approach should turn out a lot better (and probably maintainable so I don't have to trash it again )
Posted By: Anonymous
Re: What are you working on? - 06/18/1620:01
Quote:
@Malice: well it's 2d only, which means you can't really combine it with 3d rendering (if that's what you meant)
No it's not quite what I meant. So if we use a 3d engine analogy, A 3d model use smaller 2d sprites and/or 2d particles to create FXs. The programmable nature of this FX system allows for some randomness and "noise" in each instance of the FX.
So in a 2D system, where 2000 sprites can be rendered but probably only 20% n400 will account for the main entities, than maybe the other 80% can be used for a FX systems that allows for randomness and "noise" in each instance of the FX.
In the gif below.. the gun-flash is always the same... But with a programmable FX system and using(for ex) n10% total sprites, couldn't you achieve some pre-instance unique FX for the gun flash...
Is it a waste? Probably. I just really love the idea of adding chaos and variety to all games. My mind just dislikes the repeated exactness in the visuals of all games. To me it can be the greatest, most beautiful 3d rendered game with hyper real textures, but my eye and mind crave those little bits of dis-uniformity that bring me closer to real life. EDIT- That is when I'm not running for dear life and I stop to admire the vWorld -end edit
So I was thinking with 2000 sprites a vFX system is possible. And of course in a vFX system collision isn't really needed. Or at least pre-sprite collision, a volume for the whole vFX would probably do.
Shouldn't be too much of a problem, unless the number of particles gets really high. Although (in the case of a muzzle flash, like your example) I'd prefer to "bake" the animation and use a single or at least a lower number of animated sprites.
Posted By: Anonymous
Re: What are you working on? - 06/19/1621:14
Tried working on a Father's Day eCard using the engine build as deployment.
I was making sprites and models act out a cut-scene basically. But it's quite bad looking. So here is a bad looking sprite for you... lol
I managed to port the pixelated (deferred) renderer that I made a couple of months ago to my new rendering pipeline:
I'm now using my own 2D renderer, instead of using acknex's VIEW* and ENTITY* objects. It works quite well and the performance should be a lot better than before.
Like the old one, it supports dynamic lights, normalmaps, high-quality HDR and some other post-processing effects. However, the visual effects aren't properly adjusted right now. I'll need a full scene to do this.
Iam bored. Working on a funny animated desktop shredder / trashcan. Eyes follow cursor, eyebrowns are moving, and maybe it gets more fat with more files in ^^ When drag and dropping files, it opens its mouth ^^
edit: Of course the window frame wont be visible later on Desktop edit2: If some spriteguy is bored, i would love to Change sprites to a "fat Comic rat"
I am currently on a major bug hunt in my game. I want to make a stable, working demo that I can let people try out.
Questions for experienced developers: 1. Does it ever end? 2. How do you stay positive? 3. How/When do you decide that it's easier to just take what you've learned and start from scratch?
Thanks. The eyes are really funny, guess deleting files becomes fun ^^ Redone sprites
1.No 2.Good question iam asking myself a few times a day since some years. Me for my part i just want something that makes me feel "ahhhh". 3.When u feel its DEAD ^^ ..but: every new start is gettin better!
Iam stuck in design with the (City)level too. Guess iam going back to ShadeCEvo and a scary indoor design attempt.
Questions for experienced developers: 1. Does it ever end? 2. How do you stay positive? 3. How/When do you decide that it's easier to just take what you've learned and start from scratch?
I'm not experienced developer, but here it goes: 1. I guess no. I think it's just same as life routine, it just goes on and on. 2. I try to motivate myself. Think of something positive, set a goal for yourself. F.e. to sell your game and get good income (or at least get audience attention, for your next projects). Everytime I start losing motivation, or start loosing my belief in my project, I just look at those games which are sold on steam, it helps 3. This question is not clear enough for me. But maybe you meant that situation, when you rewrite your project (start from scratch?) over and over again, as the whole progress goes on and you find better/faster/easier ways to implement all that you've already done. My main problem on this one is, that as the project grows up, it becomes more spaghetti like (I guess I've read about this in AUM or MANUAL) and it's really hard to keep things organized. So there are some points of development, when I just stop and rewrite everything on the paper, then I 'copy&paste' those parts of the project, which shouldn't be changed into new .c files, and after that I just rewrite the rest as I drafted it on the papers. It happened already twice with my small 8 bit shooter game So far I'm happy with results, and it seems to be very stable right now.
So we finally numerous updates in this video I wanted to introduce you to our special turret code , I think next month I can imagine an official announcement what it is in my project. and do not be surprised , the whole is still in development , so the whole bugs have Fun !
Hey Great thanks that you like it , I never expected to get such a good SO kritick , for spring-loaded sound and EVERY other synchronous project Must you my thank speakers , Currently there are 6 dubber involved in the project
I make grants some advertising But here please the player speaker
@rayp -- That's really cool! Looks like a fun project. @BlackJack really nice looking. goning to be a really cool project.
@Dooley -->Malice Lecture Warning<--
Quote:
Questions for experienced developers: 1. Does it ever end? 2. How do you stay positive? 3. How/When do you decide that it's easier to just take what you've learned and start from scratch?
I'm not count myself as a "experienced developers" I say long time trouble shooter.
1) No not really. It helps to have more eyes and hands working on problems but even the AAA teams never fix it all. In fact in my last software development course the instructor told us "Code-projects are never really finished, never truly perfect, and can always be made better." Of course he wasn't talking about the 'Hello world' app lol. A few things can help you. Design design design, plan, research, plan-the-implementation, and write modular. *So taking a good time to really think out the whole project, it's parts and how they fit together can be worth months of designing before one line of code. *Planning out each one of the parts and how they will act and interact again is worth a lot of time before coding *Research how other people have made the parts work - if there is a better way - if there is a quick solution you can just grab and modify. *Plan-the-implementation- This confuses people, but don't just have an idea of a system(ex. movement) , how it should act, and what the end result should be - instead map out the code design in flow like charts ( This assumes more complex than simple movement). *write modular - Keeping code parts separate and organize is very helpful. Basically this way you can quickly plug-in or unplug a code-part system, with little effort and without break other code-part systems. Also don't work in a vacuum 2)LOL... Me stay positive!! --> positive or negative these come and go, it's a goal/a job , so positive or negative you've got to get up and do the day's labor. You've negative will turn to positive when your progress starts build and your problems stop piling up. However you feel, just keep getting up everyday and working. 3)Never really! Starting new with what you have learned will simple mean new game and content that reaches the same limits of the unknown. I would say if you like/love the project then never stop working at it until you feel you've completed it. Maybe it's about taking a week to redesign it. Or sometimes I've moved so fast I need to stop coding new stuff and debugging it, instead I spend a week organizing and commenting it all. Once I clean the mess the bugs jump out more obvious and while I'm cleaning house I just keep a running note-pad of bugs and issues I find. Then I go back to the bug fixing or new code.I'd say if you have to keep tearing down the house while you build it, than you didn't focus enough on the prep and design work.
Let me know what you think, any feedback is welcome.
Posted By: Anonymous
Re: What are you working on? - 06/29/1616:12
^ Very cool. I love the glowing grid texture like Tron. I used the same kind of visual in a very different project type. If I suggested anything it would be a visual at the bounce point.Like it's bouncing off an energy wave... Very crude example..
Questions for experienced developers: 1. Does it ever end? 2. How do you stay positive? 3. How/When do you decide that it's easier to just take what you've learned and start from scratch?
, that feeling sounds familiar lol. Lost count of the projects that I scrapped. Atleast you learn something from it for the next one.
@Malice, consider it done my man! I was toying with an idea like that actually, your feedback has convinced me that it's a good idea. Actually just got done adding a few improvements here and there. Will post a demo soon.
I added some new stuff to my 2d renderer (an additional forward pass for flexibility; allows for transparency and other things) aswell as lut-colorgrading which I'm using for color correction and hue-shifting (= "warm" (yellow-ish) highlights, "cold" (blue-ish) shadows).
(I'm already annoying you with this boring, technical stuff, right?)
No screenshot this time, as the differences are just minor.
By now I'd really love to put my new 2D pixelart graphics-stuff to some good use... but as always I have just no f-ing clue what to make. Gotta hate this specific lack of ideas.
currently this is adapted to the Zeal project , and was also for created because I have passed a pay even him
Posted By: Anonymous
Re: What are you working on? - 06/30/1621:20
Quote:
(I'm already annoying you with this boring, technical stuff, right?)
Not really. However you do add a bit to my study list. I must know what are these things... Lucky for me there is a web of information to learn from.
Quote:
but as always I have just no f-ing clue what to make. Gotta hate this specific lack of ideas.
This is kind of a tragic curse.. All that ability and your mind just blanks on how to execute with it.
Section that might be better in a Rant.
Click to reveal..
This community has always lacked working artist and working designers. As for the designer, we have tons for qualified programmers who just fall blank in the idea department. They have the programming and organizational skills, the studying and planning skills. However, they just blank on a great idea path. The other-side of the coin are the people who just think they are designers.This is like waking up on day and just claiming to be the boss of your work-place, even if you lack the skills to be boss and knowledge of what those skills even are to be in charge. They have simple ideas or pieces of other games they would like to mesh together.They tend to lack the programming skills to think up the design and the organizational skills and planning skills to fully flesh-out and document a project. So you end up programming for someone who took 2 hours to think out your next 20 days of hell and rewritten code. Because they have no idea the path and are trying to think it up or correct it as fast as you can code the last idea-part of their blind mans run.
Of course to the solo developer this does not apply. You need not plan and document a path if you are the whole team... However, I would work on project teams with a real game designer and a true documented design path for me to work from. And as I've taken on trying to be the Boss , I've been both types of the above designer , and in turn.
END Section that might be better in a Rant
Quote:
A9 ? huch you get the news to a Release ?
I do indeed but will only release it to those people who send payment to maliceisfullofcrap@paypal.com - (Joke implied - Do not send payments to this account)
I like that logo. Is the white gradient like dither just visible on my display?
However, I think that this is something that can be "trained". What I mean is that doing some game(play) design, game art etc. will not magically cause you to come up with awesome stuff, but the experience and knowledge you gain can definitely give you a "good feel" about what works and what doesn't and it can also spark some good ideas.
Just uploaded a new gameplay video to youtube, this time showcasing the "Rush" mode. Feel free to take a peek and let me know what you think. Actually, you know what, why not just download it and take it for a spin yourselves? Check out the Showcase section for a link to download the game in it's current state.
Working on sprites pre-rendered from models directly in Acknex with ShadeC EVO. Creating the animation files with my old editor CED ( see "Third party tools" ) Looking really cool in game.
edit Started Friday to work on a topdown "ultra arcade" shooter. Map comes from my zombie-fast-fps. Took some time to adjust the workflow, all enemys are sprites, means a lot will run around on the map ( first test around 60 on screen, pathfinding and 60fps in fullscreen )
Looks great! Is there an automated way to create sprites from 3D models? I have done it with Print Screen and photoshop...
Posted By: Anonymous
Re: What are you working on? - 07/19/1617:32
^-- You can do this with your 3d modeling software.
I max I set the key frames to be rendered. I set the camera. Then a render out the frames all at once. After there a piece of software called glueit , that I select the rendered frames and it stitches the sheet. My final step is to use gimp to turn the background color to a alpha channel. However with any single color bg color the engine can turn it to alpah.
The problem with malice way is u have different looks, better to do it in the engine itself, just my view.
No no autoway. I wrote a simple script that handles it for me. Simple sayed it makes scrshots while animating with fixed cam view.
From my shitty handy Greets
Edit Ill make a project thread soon showing that stuff and the project in detail, but too much real work those days to do. Hunting dollars like a * ...sadly
Posted By: Anonymous
Re: What are you working on? - 07/19/1620:08
^-- In this case(Which is fully true), the script can both capture the screen shots and glue them into a sprite sheet(strip) with bmap_create and bmap_blit.
Again alpha can be handled at runtime with a color key or maybe there is an engine bmap function for replacing a color on the strip with a 32 bit alpha before saving it out.
Instead of doing something useful, I spend a few minutes hours (definitely hours) to implement Normal mapping for the tiles in Dungeon Deities. Of course, the normal map shown here is not a great fit for the actual image, but it worked great for testing.
When I did it I was using MED exclusively, and hadn't learned any other 3D software ... that would make it a lot easier!
Posted By: Anonymous
Re: What are you working on? - 07/20/1617:04
@Error014 - Nice! Go to see your working it more. @rayp - Sweet! You over came the limits and have a window to make something really great... Eyes on this! @Dooley - Green Light! Really cool.. There is a momentum here again.
Posted By: Anonymous
Re: What are you working on? - 07/21/1600:21
An attempt to use autodesk character tool and my bad skills to make a character.
The purple 'dress' will end up being more of a 'Deaths robe', the 3ds max materiel hasn't made it into the engine yet.
The character has a real size but, will end up being squished scaled to have the cartoon n64 kind of look. Legs, feet and arms are left open for large comical gloves and boots. Yes, I'm bad at this and only the autodesk cheat tool makes it look ok..
@Realspawn - Thank you - you know I hate the content creation stuff...So Slow so hard to do, so very much to learn and know.
@Reconnoiter- Thanks also - I just keep going. It's fine in max for now, but no way I can put it in the engine as is.. Maps, and face counts and all that.. We'll see how it's comes in the next few days. Time permitting..
I'm not even sure what form the chains will be in the game ready model. I can use 3ds max and a physics simulation but it won't be clean and looping animation. I could simple assign them to animated splines in max and the exported model can be a second model using point animation. Or do the same and convert the aSplines to bones. I could end up turning the chains to a texture and just assigning that to a alpha skirt like the bottom in the very first pictures, this can be bone or again a second model with point animation. It's a bit ahead for the moment, so I have time to figure out a better plan for them.
Edit/Updated Decided to make fewer large chains for both polycount and animations reasons. 2 front chains have the reduced polys that will most likely be game level. 18 polys for link 180 per chain, 8 chains = 1440 polys for chains alone!!!
Posted By: Anonymous
Re: What are you working on? - 07/21/1623:29
@rayp - That's looking very cool and BLOODDYYY! I looks like a great game in the making.
Already really nice. Shadec is the best contribution ever for me. I like the specular settings.
, tnx, yeah its such huge difference compared to the default lightning. I think Shade-C should have been supported more by JCL. My only problem is that I can't seem to get some models to ignore dynamic lights and sun etc. (would be nice for e.g. 3d interfaces), and that it removes a bit a few options like tinting entities through their rgb vector I think (or maybe I am doing something wrong).
The dungeons aren't created by me in WED or something - players create them themselves on our webpage (which allows you to login whenever (say, while at work ;)), and change something). Then, in the client (=the part made with Acknex), we randomly select one of the available dungeons of an opposing player for you to explore. Make it through and steal one of their magical artifacts granting powers such as summoning horrific beasts to serve you!
Since those levels are designed online, they're stored in our database. And I thought a bug in the Room Editor script had destroyed parts of this particular dungeon. Looks like it hasn't though!
By the way, this screenshot also features Normal maps that now fit the tile imagery better.
, just noticed your reply, but unlit doesn't seem to work with shade-c. I tried also material settings and commenting some out in the .fx files, but so far not being able to disable impact from dynamic lights. I temporarily use particles to color the 3d buttons, which works until fps changes.
Hm, not sure if I remember this wrong, but wasn't UNLIT just setting an entities' material to mtl_unlit if set in WED and basically do nothing regarding dynamic lights otherwise?
You could try setting it's materials Diffuse-parameters to zero, but I'm not sure if this will also stop the calculations from taking place (or if its all calculated only to be thrown away by multiplying with zero later).
Though I have no experience at all with Shade-C, so maybe all of this is wrong and useless.
Posted By: Anonymous
Re: What are you working on? - 07/26/1618:45
@Error014 that really cool!
Quote:
You could try setting it's materials Diffuse-parameters to zero
This will do nothing... In sc_entity when setting up the entity they change or plug in values to sc_skill (NOT my.SC_SKILL !!!), the answer many be in change or adjusting those -- If you knew what they all were and what they do.
I love to see people still doing good stuff with 3dgs it proofs that it's the creative mind that makes good stuff and not only the engine.
Since I understand Lite_c more and more I be able to pick up old projects from my shelf and try to finish them so here is marble aventures. Still tweaking some stuff. Thanks to Malice I had a good start and idea how to do it all. I needed to add and change stuff but for now it almost works all needed is to add more levels and enemies with their own unique attacks.
^-- Great work and good job... I can see I left a bug in the 'face the mouse' part... Sorry.
Quote:
Since I understand Lite_c more and more I be able to pick up old projects from my shelf and try to finish them so here is marble aventures. ........Thanks to Malice I had a good start and idea how to do it all.
Great that you are getting the programming down. I'd say be very careful with those old code project I wrote on. We did a lot in a little time and it was all very QnD and I think they used some very crude code to keep the logic simpler. So you should start to see that it's has bad habits and poor code. That's when you know your really getting good at lite-c, pick up those old scripts and you 'just see' how they should be written better.
I am working on a RTS with First World War setting for a lot of month now (I am not sure anymore but I would guess 1 ½ year by now). Some parts are already done, like the models for most of the units, the User Interface, most of the units AI and control etc. Right now I am designing the first mission/map of the game, so the level design is still WIP.
@Jerome8911, will it have skirmish? Is there some form of base building? Looks good, though the terrain is a bit flat looking, do you already use a normal map for that?
@Benni003: I like the overall color composition of that scene, giving it already some kind of a look. Just the resolution of the grass texture seems to be a little bit too low in comparision to the building. I am excited to see your progress because I loved these old city building games like Caesar III and Pharao.
@Reconnoiter: By now I am just working on campaign missions against the AI. I have not yet started to implement anything for multiplayer or a skirmish mode against the AI. Maybe I will later on, but I don't have any experience in programming anything multiplayer related.
The game won't have base building. The gameplay is less like Company of Heroes, but more like the old Codename: Panzers or Blitzkrieg Series (limited amount of soldiers and reinforcements).
The terrain is using normalmaps, anyhow I see what you mean. Maybe I can tweak that a little bit.
Posted By: Anonymous
Re: What are you working on? - 08/01/1607:18
Realspawn is a thief
He stole from many to make p-slide and he's stealing now for his android port's -- we all coded that!
yup 15 years... I know it all lol
p-slide he promised all the coders money then told me in skype he got checks and paid no one... Got prof , use your mod powers to erase it this time lol...
<<--- Been screwed over so many times..
Posted By: Anonymous
Re: What are you working on? - 08/01/1607:44
You can try this time to discredit me Realspawn, but I'm beating all will see, like before I've give answers and help. And you've taken all so. 50% of you so call tutorials have my code !
You don't make a dime this time without paying or everyone knowing your a the if thief ..
Not sure if serious, trolling, drunk or account has been hacked. Anyway, thanks for spamming the forum. That'll surely help!
Posted By: Anonymous
Re: What are you working on? - 08/01/1608:54
Quote:
Not sure if serious, trolling or account has been hacked. Anyway, thanks for spamming the forum. That'll surely help!
Serious... He deleted all my post those years ago and banned my account, I waited this time silently to expose it. Spam yes, truth yes... thanks for noticing he can't use his mod power like before now!
Posted By: Anonymous
Re: What are you working on? - 08/01/1608:56
I understand and accept your need to remove the post and me to some extent... But please don't repeat this. He stole and got paid and I got nothing. They all should know and he shouldn't be allowed to erase it this time..
What is this all about? What is p-slide? (first come up in my mind was powerpoint slide but that doesn't make much sense.) And you are saying Realspawn banned you earlier, when was this? Since I see your current account is from 2007 which is quite old, was this ban before that?
Quote:
I waited this time silently to expose it.
, why waited than? You now know for sure or something like that?
it does not make any sense to me at all and i don't appreciate this. Frankly i don't care if this is drunken talk, misplaced kind of humor there is no excuse for this weird behavour. Back to what are we working on. cause that is the topic of this post.
So, I've written my own input system which supports mouse, keyboard and gamepads (both direct-input and x-input). However, a function to bind keys ingame is not implemented, yet.
I also started to do some art stuff for practise and to test the new graphics pipeline:
click for full resolution
There's not really a game idea behind this, I just wanted to do something. (and I love space so...) Obviously it's still work in progress (only half of the lineart is done and a single object is colored).
Great. I like it! Looks like a much better Space Quest ^^
edit: Working on a single global ai Controller, one while one cmove call per frame etc. Thanks 3run for the tip. Nice enemy amount with pathfinding and high fps.
I love the afford of time and quality that you all put together.I love to check the forum here from time to time. Keep up the good work. I allways wished to get deeper into gamedesign etc... but thats really a long time ago ;D
As i create a game i write everything down that i do. So an other workshop is in the making. till now 8 chapters that show how things work including all models and sounds needed. 2 more chapters to go and it will be out there : Space shooter II
An arcade scrolling space shooter with different enemies,weapon upgrades and wave attacks.
As i write it while creating the game, you can see mistakes i made and corrections I add later all kinds of scripts are used so always handy to have it as a lot of scripts are used in many games. Working now on the eyecandy arcade chapter. Working with animated and non animated sprites sliding in out/ growing and moving + Particle use
As a side note, is anyone of you creating GIFs of their game? If so, what are you all using? GIFCam? There's so many tools, and given that some people used to distribute GIFCam (which seems to be the tool of choice) including Malware, I'm a bit hesitant to download.
@ painkiller, yeah environment mapping for shade-c would be cool. You try to decrease the light range a bit and do more spotlights. You could also maybe try some soft light beams through particles or through transparent sprites.
Here's one more link for you guys, hope you like it! This is about the effect when enemies are spawned due to the player triggering a trap. The game automatically finds an empty tile/space for the enemy, regardless of how horrible the room you've designed looks and is made for that.
@ painkiller, yeah environment mapping for shade-c would be cool. You try to decrease the light range a bit and do more spotlights. You could also maybe try some soft light beams through particles or through transparent sprites.
I've changed light's positions and color, now I get a different result:
normal maps on that white panel walls look inverted, could look better if you flipped them.
yeah, actually not inverted but that normal map gives a weird result. I've generated a new normal map from the diffuse texture and I think it looks better now:
It does seem a little bit better. Is CrazyBump worth it? I am currently using the nvidia normal map plugin.
I created it using the demo, as I've just installed it. I've taken a look at the pricing and an indie license costs 99$. I'm not sure if it's worth it. The good thing is that it has different settings to get the result you want:
I'll take a look over the internet to see if I can find a good plugin for Gimp. Currently I have one based on the nvidia plugin for photoshop but it doesn't allow to customize as good as crazybump.
Edit: after taking a look I think there's just one plugin available for gimp, but I've seen some videos in which they later modify the result using different filters, maybe that's the way to go
Well, if you work with mixed material and you only have "regular" textures, generating Normal Maps is the last resort, as it always produce just a sophisticated guess.
I liked PixPlant: https://www.pixplant.com/ ... because it generates you from arbitrary images good repeated textures *and* creates normal maps and so on for you... so it's more for the repeated texture part. That is why I always liked that program.
Sry for going a bit offtopic but this might interest some: anyone knowing what #define PASSSOLID does in a shade-c .fx file? When I commented it out, I noticed no visual difference but it did gain me +5/6 fps (e.g. was 23 fps and is now 28/29 fps, - on a 840m).
^ From what I member looking through it's the same as No-back-Culling or the two_sided standard 3dgs material..
**However, if you are picking up fps and can see on screen transparency disabled than Kartoffel is more likely correct. It seems to set bool PASS_SOLID which is used in sc_lights.c
Opposite of TRANSLUCENT: enforces rendering on the solid pass, i.e. together with non-transparent objects. For entities with alpha channel textures that are used for other purposes, such as a height map in a shader. Entities in the solid pass are not sorted, and thus render faster than in the transparent pass. This flag can also be set by defining a bool variable within the .fx file (bool PASS_SOLID;) A
And PotatoMan saves the day Again... He's correct!!
Selective uses in object shaders may be to force that object alone into the solid pass. I could also be a switch for alphatest or alphablend(???). Sorry I can't recall now, I may have know more in post past when we talked transparency before, if you can find those..
It is weird that the quote says that with solid pass it should be faster cause of entities are not sorted (which makes sense), but when I comment it out the game runs actually faster. What kind of weird Acknex wizardry is this?!
Posted By: Anonymous
Re: What are you working on? - 08/27/1614:59
If I had to take a wild guess it would be... -- The system and rules that 3dgs was built under may no longer hold true with newer GPU design and on-board programming - This may mean that you GPU gen or even more current ones now use better systems to run the transparent pass, as this would be a logic place for each gen to improve optimization.
Other notes begin that using transparency and solid pass are pointless opposites and my lead to more work as the system or shade-c attempts to set-up alphatest/alphablend and then drop down to a solid pass. It's like saying "You must be transparent" but then saying "you must be rendered none-transparent. "
But hey guess is fun right... I wonder what it feels like to actually know... hehe
, because he is the father of gs3d isn't he? So I quess he would know alot about these kind of things (I mean creating a game engine sounds like needing to take alot of things in consideration and requiring lots of knowledge/experience). And also cause I was primed that day by watching Deus Ex 1 videos where people joked about JC (Jesus Christ) Denton, so I thought JCL sounded like Jesus Christ.
Posted By: Anonymous
Re: What are you working on? - 08/29/1615:11
I'm still using 3dgs as a rapid-p test box well learning other engines. This is a video for a tower defense like game. My odd-take on the subject is to use a stationary base and a 3d area which you can place your defense devices into. The aspect is to defend a 3d space from an oncoming waves of varied attack ships and machines., whom adapt to the topology of your defensive network. They remember bad paths to the base and exploit open paths when they find one. The idea came from watching the original Starwars 3-pack, I wondered what it would be like to defend Lando Calrissian cloud city from attack.
The video is older and not a current state of dev. Of course this is very early and exploratory so it's not much to look at. Just a base in a bad 3d cloud atmosphere and so dumb attack ships.
If you've seen the video, then you'll know why I didn't post any screens..
Anyone looking for an idea or who would like to run with this idea is welcome to it, but please notify me. It's yours if you want it I just don't see any reason to work on the same idea when I've many other projects under dev as well.
I managed to get cubemaps with mipmapping to work. Using this, rough and smooth cubemap-reflections are pretty easy (and quite fast), allowing for more realistic materials :
( I know, the reflection doesn't match the scene )
Im in a contest and have 3 days to make my game. my idea was to make at a camp in arctica( name of my game ist"arctic camp") but i lost my insperation. I wanted to build fences, a water supply and solar cells. Anyway, I wanted to be picked up by a helicopter and get it eqiptment
I managed to get cubemaps with mipmapping to work. Using this, rough and smooth cubemap-reflections are pretty easy (and quite fast), allowing for more realistic materials :
( I know, the reflection doesn't match the scene )
@Kartoffel: looks cool! approaching physically based rendering?! I would think it is another game engine, if the guard was not there maybe if you use more cube maps created design time and storing their positions, you could get better reflections by selecting always the nearest somehow...
, you may want to try out a draw program called 'PD Howler'. I recently tested the program and noticed it has particle drawing/painting (mostly nature related stuff) and also includes things like editing in frames (for sprites and movies). Its UI is not so smooth though and buggy undo function.
Fixed some Bugs, and updatet the Radio/Subtitle code
, cool turret <3, cool blood dripping effect on the screen and cool that gibs fly around when killing an enemy. Also cool to see a fps sp game on gs3d in 2016.
Suggestions: for shadows, have you tried the shadow mapping from Slin on this site? Iirc is pretty solid alternative if you cannot use Shade-C. Also there seems to be a bit of lag at some points (but maybe that is the recording software). As for gameplay; its nice what you have so far, but I personally feel its missing something to keep the attention for a longer time. Perhaps throwables or special melee weapons maybe or just more equipment like the turret (that would be sweet).
Fixed some Bugs, and updatet the Radio/Subtitle code
looks great! About the npcs, I would add some animation smoothing when you begin and end talking with them, as now they turn to you instantly and the same when they start takling again. I think it would look more natural if you add some blending between those states.
Suggestions: for shadows, have you tried the shadow mapping from Slin on this site? Iirc is pretty solid alternative if you cannot use Shade-C
yes its true,the shadows is actually the biggest Problem in the Projekt, tried the Shade-c shadows, tried stencils (30fps) and alternatively the ssao shader with the same Problem
Quote:
lso there seems to be a bit of lag at some points (but maybe that is the recording software)
Not really, the video has 45gb ,:D And my hard drive had not enough space. so comes the bit lags,
Quote:
but I personally feel its missing something to keep the attention for a longer time. Perhaps throwables or special melee weapons maybe or just more equipment like the turret (that would be sweet)
lets hear what idea´s your have , ^^
Actually is this not required. this game have lots of traps in the map,
@Painkiller
I see what I can do, which is with the acknex version not so easy ^ ^
Not really, the video has 45gb ,:D And my hard drive had not enough space. so comes the bit lags,
, ah that explains it
Originally Posted By: BlackJack
yes its true,the shadows is actually the biggest Problem in the Projekt, tried the Shade-c shadows, tried stencils (30fps) and alternatively the ssao shader with the same Problem
, have you tried setting lower resolution for the shadows in the respective shader files? Although it will always eat away some fps, it can make a bit difference. Also an idea is to only enable shadows for actors (enemies and npc's etc).
Originally Posted By: BlackJack
lets hear what idea´s your have , ^^
Actually is this not required. this game have lots of traps in the map,
, from the top out of my head; variations on the turrets and traps with blades or flame jets etc. And cool that it has traps in the maps, didn't noticed that earlier.
@Reconnoiter hats off,Already alone that with shade-c to link,!
@Painkiller Oh well.I was at the age of fifteen,Where I started with Gamestudio.Without internet and other tools, I have then fought somehow.On the other hand I find the Med pretty good
@Reconnoiter hats off,Already alone that with shade-c to link,!
, tnx! though this some simple scene and simple lightning, next I am going to try polices sirenes (and other moving lights) and indoor scene.
Originally Posted By: BlackJack
@Painkiller Oh well.I was at the age of fifteen,Where I started with Gamestudio.Without internet and other tools, I have then fought somehow.On the other hand I find the Med pretty good
, nice, fifteen is quite young indeed. I personally started at fourteen with modding but gamemaking is more difficult (I think). Anyway, good job with the model (which reminds of me of a low poly Blizzard model for some reason).
3run: Looks promising, but maybe you should add some kind of suspension to the wheels, this gives the tank a heavy and massive feeling
Thank you! Yes, I thought about adding suspension on the wheels, but I was too lazy to add them, as far as I understood I need to use additional joint for that, between the base model and each wheel (maybe PH_SLIDE joint?).
, cool, what shade-c shadows settings do you use to keep shadows at a distance look good? In my shade-c project shadows seem to get clipped away to soon and pixelated-like at distance.
, cool, what shade-c shadows settings do you use to keep shadows at a distance look good? In my shade-c project shadows seem to get clipped away to soon and pixelated-like at distance.
actually I haven't still changed the shadows settings, I have them as they came with the shade-c demos:
Code:
sc_screen_default.settings.lights.sunShadows = 1; //enable shadows for the sun
sc_screen_default.settings.lights.sunShadowResolution = 512;
sc_screen_default.settings.lights.sunPssmSplitWeight = 0.7;
sc_screen_default.settings.lights.sunShadowRange = 5000;
sc_screen_default.settings.lights.sunShadowBias = 0.001;
actually I haven't still changed the shadows settings, I have them as they came with the shade-c demos:
Code:
sc_screen_default.settings.lights.sunShadows = 1; //enable shadows for the sun sc_screen_default.settings.lights.sunShadowResolution = 512; sc_screen_default.settings.lights.sunPssmSplitWeight = 0.7; sc_screen_default.settings.lights.sunShadowRange = 5000; sc_screen_default.settings.lights.sunShadowBias = 0.001;
, I think I have fixed it with setting sc_screen_default.settings.lights.sunShadowRange manually tnx (I had it on 0 which than uses camera.clip_far iirc)
I'm amazed with the results you guys are getting! Could you help me out with setting quality of point/spotlight shadows? I only found sun light parameters.
, sort off but it doesn't have reflections and no waves, but if you have sun light (and probably normal lights to) moving over it I noticed the effect of the normal map somewhat mimics movement a little bit. Particles may also help, haven't tested that yet. Anyway the water is just a thin model with a transparent water texture and a slight emission effect to brighten it a bit.
Originally Posted By: 3run
I'm amazed with the results you guys are getting! Could you help me out with setting quality of point/spotlight shadows? I only found sun light parameters.
, I have tested indoors only a little bit yet, but see the shade-c example for spotlights or sc_lights.c. Important are sc_light_create (lots of choices here cause of function overloading[?]), sc_light_update, pan/tilt, sc_light_setRange (I use this for lights that go out and on rapidly ), sc_light_setColor (nice for disco's ) , sc_light_remove. Dont forget that lights are entities, and so add them to a group that is ignored in your collision and trace functions.
Oops I misread your question, didn't notice the 'shadow' part. Okay now to the interesting part. I have been digging into the sc_lights.c script and noticed some size vars and black bmaps created for the shadows. So I changed them from 256 (fairly small which explains the ugly shadows) to 1024, and get alot sharper shadows BUT the shadows are displaced or something like that and the're weird grayish doubles of all objects beside them . Maybe some value I forgot to change? (sun shadows are off)
I noticed that when you reduce the spotlight's arc, you get better looking shadows too (probably because of the smaller/thinner area). So atleast that is a way for creating decent indoor lights.
But we should try if we can achieve a higher resolution too cause with a fairly high arc and 2 spotlights (everything needs to be the same except for reverse pan value) I noticed you can create 1 point light but with shadow support (since I think normal point lights dont support shadows). The shadows now dont look too good but I think with a higher shadow resolution they will look ok:
I don't remember exactly of Shade-c EVO shadowmapping, as I remember it utilizes an exponential shadowmapping. If I am right, you should increase the exponential shadowmapping constant (or factor), or to decrease shadowmapping depth distance, to eliminate that bad fade out at legs. Increasing resolution, and decreasing shadowmapping area also should help. (in reality, shadows should be sharp at legs and blurred at more distant parts) The best is to create a tool where you can tweak all parameters in real time, that would save you a lot of time. I did it in MapBuilder, it really does worth the time! Or ask JCL to integrate it into WED
Hi Sivan, I can't find depth distance or exponential constants factors, are they maybe in the .fx files? Also this the previous pics had an extreme arc value which makes the shadows look alot worse (but you can create points lights with). This is with an normal spotlight arc (60):
@BlackJack: I think your project's visuals would heavily benefit from using reflections, on most of the materials. Cubemap and maybe even screen space reflections (a combination of both approaches usually gives best results). Additionally, you could do a view space normal lookup highlight/ shading thingy for at least metallic and skin materials. I don't read too many tech papers but I found http://www.adriancourreges.com/blog/2016/09/09/doom-2016-graphics-study/ and a Killzone: Shadow Fall paper very interesting and inspiring in that regard. The metal highlight or skin lighting is as simple as OutLookup = mul(normalize(mul(InNormal, matWorld)),matView)*0.5+0.5; in a vertex shader, with an appropriate sphere-ish texture. They used that for instance for character rendering in Half Life 2, if I'm not mistaken.
Looks cool! Do I understand correctly that you are making/designing the map editor? If so, does it come with these assets? What makes it different/better than WED?
Working on super mario run clone Who needs mario when we have SID ! so far so good I am always interested to see how a game is done and if i can repreduce it having fun doing it
Designing a model for my honorable ass ... maybe add some animations for farting, defecating ,diarrhea shooting ... kinda hard since the ass is a hard thing to animate ...
I've been recently working on a movement code (as always damn ), and main goal was to create climbing mechanics for stealth/horror game. It ended up with this features:
Quote:
- climbable ledges, walking sideways on them, jumping from one ledge to another - climbable ropes/pipes, also player can jump from them to grab ropes/pipes/ledges near by - walking tightrope (this bastard was gilty for the appearance of my insomnia) - handrail climging (stealthy as shit! )
Special thanks to MasterQ32 for helping me out with some awesome ideas, as well by testing everything I sent him and especially for writing me an example of 'find position on line' code ( How can I project a 3D point onto a 3D line? ). If not him, I would probably already lose my motivation.. Next things to add is swimming!
Hey! Well, I'm not an artist, so no nice locations from me, sadly.. But hopefully I'll do my best for levels in my project
Originally Posted By: mk_1
3run, are you interested in sharing the code for the annual game jam (ackcon) in Germany?
I'm not interested in sharing the code yet, also I would like to know a bit more about ackcon, since the thread is in German mostly, I don't really know what is it about
AckCon is an annual event in Bielefeld, Germany, where (partly former) 3dgs users meet. We usually end up creating a game over the course of one weekend using the acknex engine. You can download and play most of the games here: https://vksbm.de/ackcon/
Except for the 2015 game they're all in English. I like 2014 (Ackmania: Operation Deadnex) and 2013 the most. Last year's game was accomplished by only five people, including code, art and music.
Wow! I played some of those games, but didn't really know that they where from AckCon, thought it was some kind of games to troll jcl (zorro and other things) Anyway, right now I'm not really interested in sharing the source, I'm really shy to show But if I'll be able to make out a good working example (hopefully), I'm going to publish it on the forum, under contributions section.
you can give the code to the AUM winter contest, but i seems, here are the most peaple are interested in shooter games.
P.S. one Jear after i started with 3dgs i was too in bielefeld, we had no warm water. And I was one of the few to shower more than once hohohohoho I was helping in the race game
I have make lowpoy models, for example the start/finish building and the vegetation, some uv s, I have maked with someone, in windings3d. It was not much, but we were 14 men and every one wanted to do something . edit: I have taken so many memories with me. One actually had his pc tower in the train taken and again one had his old rren screen with. For the soundtrack 2 had rehearsed in the large dormitory. It was also a graphic artist who had modeled the cars, which had a crate mate with, where all felt it tasted after moth XD. Just sid had a nerfgun bought and occasionally börger for all fetched (the system to collect the food for me was unclean, and I still believe that he has not gotten all money back;) Master ## had provided the Gui with shaders,Which impressed me very much. and the topic isometric at least as critically discussed as if all to unity 3d are changed XD
I haven't posted any updates here in a while. But Dungeon Deities is going strong and progressing well!
We've just uploaded a gameplay video of the Tutorial (YouTube link) - tell me what you think! (If you want to skip the introduction and go straight to action, skip to 1m:38s, or click here!
A few more GIFs for you!
(Color changing tiles that require specific movement through the room - good for puzzles!)
(Teaching the player gameplay details in a fun way)
@Error014 Is this Gamestudio ? i like this Projekt, thats looks really smooth, and the Kamera works to good, ! nice stuff. can you tell me what the codec for animatet gifs is ? ___
at the moment when you realize the stencil shadow were not a good idea
and the Town is now bigger
Someone has an idea what I can do with the shadow?
@BlackJack are you using shace-c in your project? In that case you could use shade-c built'in sun and spotlight shadows. Else you could try gamestudio's PSSM shadows
I use Shade-c only partially, I have switched off most of the functions. I think I am doing something wrong with pssm,
Code:
// all objects get the SHADOW flag
for(you = ent_next(NULL); you; you = ent_next(you))
reset(you,SHADOW);
// adjust camera and bias, activate PSSM
camera.clip_far = 32000;
camera.clip_near = 30;
pssm_fbias = 0.0005;
pssm_run(3);
// Shadow
mat_shadow.alpha = 25; //transparency of dynamic shadows
shadow_mode = IGNORE_SPRITES+IGNORE_PASSABLE+IGNORE_PASSENTS;
shadow_stencil = 4;
shadow_threshold = 30; // 80
shadow_offset = 2.5;
shadow_lod = 3;
shadow_range = 30;
Which is located in the main. So I get a strange bug with the Weapon and the framerate does not look better, the Shadow flag I've selected by Wed. The building model is also 100% closed I have before long time see a shadow shader who works with gpu here in the forum, but I can not find him any more
here is the screen with (pssm)
The shadow falls dashed on the floor, there is nothing on the building and is missing half from the Weapon.
maybe gamestudio's pssm doesn't work together with shade-c
Actually there should be no problems as I use shade-c only partly Now I have however an alphakanal bug, or the znear flags on the view camera are not co -generated.. In the menu level is not the Problem,
@Error014 Is this Gamestudio ? i like this Projekt, thats looks really smooth, and the Kamera works to good, ! nice stuff. can you tell me what the codec for animatet gifs is ?
Thank you! Yes, this is indeed Gamestudio. Camera movement and its acceleration and velocity have been carefully tweaked over a long time, so I am very glad you liked it!
As far as the GIFs go - they're just quickly recorded in GIFCam and rudimentarily edited there as well. It works surprisingly well and is quick - and had to, since I do these GIF-things almost every day. I can't tell you what GIFCam is doing when exporting, but these above are just the normal standard save-settings. Just remember to set the FPS to 60 when recording to get the smooth GIF-look
I like your game as well, its got a nice look to it! However, you seem to have some framerate issues there - 34fps on the main menu screen is a bit low. Shooters should not fall below 60, and higher is always better. Now, of course, Acknex is hardly a powerhouse, but perhaps you can find some room to improve that?
@blackjack If ure using shade-c shadows set the stencil flag in main function to -2(not at home but guess it was -2). Activate entity shadows via
Code:
set(my,SHADOW);
In "setting up shadec" topic u can take a look of the main function examples. Maybe this might help. If you wanna use gs shadows with shadec this may cause some troubles. And they are really really slow btw.
And yes theres a problem with shade-c and znear flag i can confirm this. I solved it by making the weapons mdl very small and using other vectors.without znear flag activated. Not the best solution i know. Long weapons like a bushwacka for example will still penetrate close walls. The znear bug only appears in combination with some flags and/or fx settings btw. because iam not at home i cant tell more exact but i think it was glowing(alphachan) and znear together. with this nice weapon i realized the bug
Quote:
Now, of course, Acknex is hardly a powerhouse, but perhaps you can find some room to improve that?
60fps in a shooter in 3dgs isnt possible. 35 to 45 maybe a bit more is realistic. I know that sucks but its the way it is. good performance boost btw is one while that controls all enemys at once i call it mass-ai. i used that trick in my top down project. Thanks again 3run for that nice tip.
Thanks. Its a bushwacka mod level 1. in this level it is electroloaded(u can see the pink sparks) and a budwiser can filled with pink glow slime feeds the blade Texture of budwiser needs to be mirrored of course. Yes i love it too pretty crazy
@rayp The sword looks damn good and stylish! Already alone the pink is very brave;) is used in games, in Japanese Games is the color pink for rare and very strong,
Anyone here working with high poly animated models? I have some high poly enemies (~2/3) and my fps goes rock-bottom fast when I ent_animate them. (~13500 faces and 36 bones) Is Acknex just not optimized for this or is animation like this heavy in general?
Does anyone know how to change camera.clip_far in the middle of a map with shade-c (evo) running?
I want to optimize maps by limiting the clip_far at some places in the map where there is also fog, the problem is as soon I change the clip_far I get weird shadow effects (somewhat similar to setting the wrong shadow bias). Normally these settings are set with or before shade-c settings/init. So I tried simple loading shade-c again but that gives the error "Script crash in sc_forward_destroy" (maybe cause I did not properly destroy the previous shade-c thingy?). (so far I can't find a destroy_sc or something like that)
Yay - still working on YRPG Toolkit. Right now I am testing the result of all tools as union.
Reusability was a keyfactor at any time. Here is a small titlescreen, which uses HOG (Hand of Goddess - for environment and atmosphere), Quantum Management (the simple to use animation/particle editor) and two Entities/NPCs: One for the running girl and one for the animation effect of every lamp.
All tools are working very fine together. For this shot, all tools of YRPG Toolkit were used and alls different types of NPCs/Entities (and combinations) were used (automatic entities, touchable entities, interaction entities and entities activated by switches).
messing around with graphics tablet input and gpu-accelerated drawing/painting.
(veeery wip screenshot)
I need to rework the way strokes are handled, though. (make use of multiple position changes in one frame to fix these extremely choppy lines) But other things like pressure sensitivity work fine.
Finished a small simple game based on my favorite tv cartoon Rick and Morty. I had fun creating the models and turned it all into a fan based little game. 2 player and 1 player option.
Me and my colleague are currently working on a second set of new weapons, Respawnplaces, which will differ according to the breed and the environment (in the picture for the human race) ,further leveldetails (Level 1) and Vehicles.
Me and my colleague are currently working on a second set of new weapons, Respawnplaces, which will differ according to the breed and the environment (in the picture for the human race) ,further leveldetails (Level 1) and Vehicles.
It's been few days since I started this platformer game, so far I'm happy with results! Maybe this time I'll make it till the end and will publish complete game!
Cool stuff I like it. Although, the animations look a bit too simplistic to me. (but the spritesheet is good nontheless, it's just a matter of personal preference, I guess).
(actually, who am I to blame? The wip character animations that I made a few weeks ago are even simpler )
Did you create a spritesheet for each of the configurations (with/without weapon/jetpack) or do you combine them somehow?
Originally Posted By: 3run
Are there any one left to post here his progress, or am I the only one to bumb this thread?
Don't worry, I'm still working (more or less) hard on things these days There's just not very much to show off, yet. I'm still in the basic framework-developent phase. Lately 2D collisions, interface and code-cleanup is what's keeping me up all night
Edit: btw. what program are you using? I've been using aseprite and it's reaally good in my opinion.
Thank you man! Yes, it's simple, but that's intentional (I don't mean that I could create something more complex than this, cause I couldn't but well I don't need anything more comples than what I have atm)
Originally Posted By: Kartoffel
Did you create a spritesheet for each of the configurations (with/without weapon/jetpack) or do you combine them somehow?
Nope, I draw seperate sprite for each configuration.
Originally Posted By: Kartoffel
Edit: btw. what program are you using? I've been using aseprite it it's reaally good in my opinion.
I used photoshop for creating this sprites.
Great to hear that there are still people working on! Keep it up and I'm looking forward to see some screens of your project!
The possibility of endless equipment objects (theoretically - I limited it to a valuable amount) and the equip-properties of elements.
More than that we have the possibility to create special equipment for certain classes. In my RPG Maker game "Final Dreams" the holy sword of dreams - razuna could only be wielded by the main hero. This idea wouldn't be just an idea anymore.
Every equipment can change now the effectifity status of elements for heroes. Our heroes will get now elemental attack types - but don't have to get one. Nevertheless every Weapon could change this elemental attack type now.
It will be more and more tactical - but not complicated. Easy to understand, easy to create and easy to play.
Looks really cool, like you've put together some really solid toolkit there. Very nice! Are you trying to remake your RPG with it (or making a new one)?
Looks really cool, like you've put together some really solid toolkit there. Very nice! Are you trying to remake your RPG with it (or making a new one)?
They are from my favourite game of the 'Tales of' series =D Right now of course for testing purposes. Ludgers Face (second one) is a bit wider. I tried to place and rescale pictures dynamically in order to fit with the box-panel.
To recreate Final Dreams would be awesome. But doubt to get this much time for a hobby project =/.
Actually my target is a deep quality check of YRPG Toolkit Core Engine. And what is better for testing, than create a small game? =) The user is able to combine Event Commands dynamically in order to create NPCs, Touch- or Trigger-Events, Global Events and Battle Events. Much of them can also handle parameters (user configurations). Right now I've 83 Event Commands and I am trying to put them together with strangest combinations.
The blood on the screen is great. The boss's texture could be a tad higher resolution. Rest is cool, though I am not sure yet about the cam shaking with the minigun. On one hand I can imagine while firing with it as the player it enhances the feeling of using a powerfull weapon, and on the other hand I find it a bit annoying to watch from the video. Maybe add an option in the settings menu to reduce cam shaking.
@Ayumi are those shadows realtime or baked? Are the mountains/mesas different models placed together, or one mesh?
i also would like to know.
Quote:
@rayp what will your file manipulator be able to do?
manipulate date time stamps and scan n replace of text strings. Nothing special. Just wanted to do some small app after a long time of doing nothing with computers.
Lets say u modified fex a excel document you were not allowed to open...with this tool u can hide it Its more a training of doing user interface for me. Edit Maybe ill add feature to show passwords of office documents. My code breaks them most in less than 2minutes.
@jumpman & rayp The shadows are baked. But dynamic shadows are planed too.
For the models we have built 10 individual stones. These stones were then combined into 10 blocks (10 models) with 10-20 stones. A good compromise for performance.
@Ayumi I like the atmosphere of your shots. Reminds me of old sith ruin in korriban. =)
Click to reveal..
@jumpan Yeah I will definitely release a demo after hunting all these mean bugs.
@rayp I know a dozen people who would pay such a thing. Did you create this with Acknex engine? I am a friend of this dark / gray themes. The highlighting of the button seems to be a bit to bright. Or is it the border size/width which disturbs me? But I really like it.
How about reading out the current timestamp and reuse it after editing the file.
I've been working on a new dungeon crawler rpg. It still needs a lot of work, but seems to be coming together slowly. I'm trying to keep a somewhat 2D feeling to the game with simple textures and shadows so that it blends with the 2d artwork a little better. I'm currently undecided on the GUI visual and I still need some more quality resources to get things polished.
CONCEPT ART : Guild
The game will be a mix of 2d style dungeon crawling similar to "Arcana" from the SNES era and some new modern features similar to the "Etrian Odyssey" series for DS.
CONCEPT ART : Grid Maze and Battle The labyrinth development is already in progress and a custom editor is being created to develop the levels.
I plan to keep working on this project and hopefully get it to a quality that I'm happy with. I will have new updates on the project once I get some more done.
Jumpman : Thanks. No, not yet. There is still a lot that needs done and I'm currently working on the custom editor for the levels. Hopefully I'll have a good video or demo in a couple of weeks once I finish the editor and battle system and get things more complete.
You could try and add some subtle reflections (simple stuff, env mapping or just view-normal based highlight map sampling) to the scene. IMO that's one of the major things which separates contemporary from outdated visuals.
I will The playground floor may emulate some sort of polished concrete with a reflective material. The railings may reflect a couple of light sources throught an object space normalmap so it will look like a real tubular fence. Same for the playground front limits so it may look glossy.
I want the project run on old machines so I will probably look for a straight rendering that looks good enought with just the backed shadows.
Old animation on a new player model. The secuence is built in a way the running loop can be repeated or obviated in order to fit the player movement ditance.
Working on fixing a database error in my space exploration game.
I created a working database, that stores information about stars and planet that the player encounters. However, a certain bug keeps messing up the data and up until now every solution I've found seems to only delay the corruption, not fix it. I am actually afraid to play it because it's so disappointing when it finally screws up
If the solution I recently implemented does not work, I will switch to a fixed database. This will be generated once at the beginning of the game, and allow the player to access the information, but will not be created as the player explores. I don't like this idea as much as a growing database, but it's better than a corrupt database...
I am going to start working casually on HeroScape like game, instead of some overly ambicious games for 3dgs and a small team I have worked on (i.e. The kID & Spoils of War) which get 60%-75% done before the $$ for the final run dries up.
Anyhow...
A pic & First camera work on the game (.rar, extract and run):
Setting that camera up specifically for a HeroScape kind of board game. If you played HeroScape before and have additional suggestions on camera, I will listen.
Wanna see some really old cool stuff, this game eventually you could change weapons, doing mission stuff. Original music and all, a early beta version that was put out publically, so I can show. This version you can click mouse-right to attack, lol.
Small update. Excel hacker. Fast breaking of passwords (below 3 minutes) via md5 alternativ password. Excel 97 xls and higher xlsx supported. Shot shows beta state. Tested it today on some busines pcs. Now i will clean up the interface.
Working on top-down shooter, video quality is pretty bad, for yet unknown reason.. If you can advice an alternative to fraps (free), I would appreciate it.
Edit: yes, I'm gonna replace sounds, after playing around with realistic sounds, I found out them to be out of place. So I'm gonna spent some time creating 8 bit sounds.
Working on top-down shooter, video quality is pretty bad, for yet unknown reason.. If you can advice an alternative to fraps (free), I would appreciate it.
, Bandicam is pretty good iirc. If you have windows 10 you can also try the build-in Game Bar recording mode. There are a couple more.
Use OBS (open broadcaster studio) with recording, you can get pretty high quality videos with compression as well as add additional stuff like images or overlays to the video
Haven't had much time, but I managed to throw together a simple editor for my new dungeon crawler. It's not much to look at, but it does it's job well and will allow me to quickly create all the dungeons of the game.
Each dungeon is saved out as an individual file and up to 100 floors can be created for each dungeon. I only added the basic features needed to develop the dungeons and everything else will be controlled by the Event blocks that will call whatever code I need.
Now I need to finish the code that reads the dungeon files and generates the dungeons in-game.
Dungeon Deities is still going strong - we've invited a few people to partake in a private beta, to see what breaks and what doesn't. So far, people seem to enjoy the game - pretty reassuring
However, we've noticed that we're suffering a bit from a lack of easy dungeons that are still large and interesting, which frustrates beginners as they often have a hard time finishing the harder dungeons. So I've started making some - here's a screenshot of one of them:
We hope to start a larger beta - maybe even a public one - soon. Perhaps you'll join then?
For more about the game, check out our Twitter, or if you prefer our IndieDB page.
@Blackjack It's really amazing what you're able to create with MED - respect!
@Locoweed HeroScapeLikeGame: Love that style of camera and could imagine the result of an addicting game!
kID Beta seems to be much like resident evil. I really love the idea and can see the possibilities - go on!
@rayp Will your Simple File Manipulator be a commercial project. GUI seems pretty solid. Would be great to see it in action! =)
@EVI Looks very nice. Any chance to see any created dungeon from file?
@Error014 Where is my demo!? Really - Would like to play this game. You've posted some screens and videos already. This is definitly a candidate for the market - go on!
HellThunder, Sure, but it's just the base structure. Models and textures are not yet decided and placeholders are used until I get the editor to generate everything. There's still a lot to be done.
Thanks for the feedback, guys! Super happy that you all seem to like Dungeon Deities!
Quote:
hey error! How many beers should I buy to take part in the beta?
Quote:
Where is my demo!? Really - Would like to play this game. You've posted some screens and videos already. This is definitly a candidate for the market - go on!
Hah, if only you had registered for our special One-Time-Newsletter, then you'd be playing the Closed Alpha right now! But in seriousness, if you want to test the game, let me know, and I'll see what I can do
EVO is actually playing, and he's creating some really cool dungeons, like this one:
And only a few minutes ago did I die in a devious place entitled "Forbidden Ruins" he created. If only I had taken the "forbidden" seriously
We've recently updated our media page with a few new screenshots! I hope you guys like them! There's videos, too!
By now the british units (models and most of the AI) are completed and I implemented the A*-Algorithm for pathfinding. Now I can work on the remaining german units and their AI.
These pictures motivate me to continue with the Gamestudio. Shaders are too complicated for me and I have already started to change my games in other engines. Your pictures look good. What shaders are in these images? /////////////////////////////////////////////////////////////////////
For almost every model in the game I am using a normal mapping shader (the standard one from the gamestudio library) and a multiterrain-shader with normal mapping for the terrain.
Dungeon Deities uses subtle color correction, normal mapping on the non-mirroring tiles (everything thats not water in the shot above), a reflection shader for those (water in the shot above), and a custom written shaderbit that adds the vignetting to the outside of the walls (thats here technically part of the normal mapping-shader).
However, for the longest time, we went fully without shaders. This is how it looked like. The lighting is "just" a combination of dynamic lighting and a customly adapted static lighting layer (made using the pixel_-instructions on a 2D-plane put ontop of the tiles)
CocaCola, I haven't used any actual shaders for the level yet. At the moment, I've kept things simple. The walls, ceilings, floors and models all use a basic material effect.
Thank you very much! So I see, it's possible to make good loocking games without shader skills(especially without shadow/ lights). The bad example for a game without shadows is Starwars Knights of the Old Republick 1. Your Games are so different and good Examples for me.
@Jerome8911, cool graphics style, will it be a rts with building elements (I mean in the sense of placing small structures like obstacles or bombs or putting infantry in structures)?
Originally Posted By: CocaCola
Thank you very much! So I see, it's possible to make good loocking games without shader skills(especially without shadow/ lights). The bad example for a game without shadows is Starwars Knights of the Old Republick 1. Your Games are so different and good Examples for me.
, try shadowmapping shader (by Slin?) on this forum or iirc you can get one from Sivan's map builder too. Or get shade-c evo but its slower.
I am already working on a "cover"-system for the soldiers, like placing the infantry in structures (e.g. the trenches or buildings)
I thought alot about base-building elements, like placing barbed-wire etc., but at this time I decided to add no base-building elements at all. Maybe later I will give it a shot.
A small puzzle arcade prototype. Created within two nights during my vacation abroad. It also owns the capability to load levels from simple txt-files.
One of the things that I've always thought would be an interesting challenge are procedurally generated worlds.
So I went and coded that for Dungeon DeitiesPsst... if you click that link, you can register and start playing!!
It's not quite ready yet, but should be, soon. It generates the overworld from a combination of noise maps, then divides the world into sections, defines the path the player should take throughout these sections, places dungeons and obstacles, and finally fills each of the rooms with decoration items (such as trees and grass) fitting the environment of the room, and also (though not yet) enemies to roughly match the set difficulty of this room.
Pretty neat!
Also... If you want to play Dungeon Deities right now, you can visit the website and register right away! For a limited time, we've opened the registration, and just keep silent about it, to make it as easy as possible for people to register, but still making sure we do not get too many users while there's still bugs in the game. Can't wait to see what you think of the game!
Hi Error, i just play Dungeon Deities and although i already played the first alpha version some years ago and i know a lot about this game, i'm very pleasantly surprised. The game is perfectly polished. I see many new nice effekt in game, game mehanic is a great, i like the camera moving and shaking, story is interesting, music and sounds is nice, levels looks really nice, i like the combat and logical tasks,... simply we have here one really great game and i'm very happy about this
Hi Error, i just play Dungeon Deities and although i already played the first alpha version some years ago and i know a lot about this game, i'm very pleasantly surprised. The game is perfectly polished. I see many new nice effekt in game, game mehanic is a great, i like the camera moving and shaking, story is interesting, music and sounds is nice, levels looks really nice, i like the combat and logical tasks,... simply we have here one really great game and i'm very happy about this
Thank you! That is great to hear, and very motivating, indeed! Very glad you like it! I'll look forward to the amazing dungeons you're surely going to create!
Quote:
Hi Error, I cant login, my username is cola and i have a mail from your game, but no password. ... Is ist a 3dgs browser game?
Oh no, that's not good! After confirming your E-Mail adress, you should have gotten another one with your data. Perhaps check the Spam-folder? I've also sent you a PM with more detailed instructions, including some of what to do if the Mail is not in the Spam folder!
If anyone else has that problem, please let me know, and I'll take care of it!
(Oh and, btw, it's a mix of browser game, where you do everything from building dungeons and cursing people, and then there's the downloadable client - made in Acknex - where you traverse Overworld and Dungeons! It's a pretty unique mix!)
@txesmi: Nice looking desert and rock formations/ cliffs! That's not a regular (terrain) mesh, right? I can't really tell from the screenshots, can you show a wireframe pic?
Haha that's funny , I will see if I can add something Lotter inspired in my game too. Is there a collection of Lotter textures/models somewhere? (would be funny if every gs3d game has easter eggs like this xD)
you can pick anything you want from there and use it at your leisure. But be warned, some scripts and shaders may induce seizures if you are epileptic.
you can pick anything you want from there and use it at your leisure. But be warned, some scripts and shaders may induce seizures if you are epileptic.
Nightmoon Saga : The End Of Never Dungeon Crawler RPG
I been working on different parts of the game features as I slowly design out what final gameplay mechanics I will keep. Also been trying to improve the battle system visuals and setting up new themes for the labyrinth dungeons.
@EVO :Nice 2D art its remember me of the good old Rpg Times ! i want to see more from this
its weekend ! and time for Indie work
The navigator npc the right hand of the main character (Pathfinder) was now finished by me, if you have suggestions for improvement! let me know I'll take care of it
3run is right, the lines mark the positions where logical "units" (basically "rooms") are connected to each other. Sometimes, you can only move to the next area there, other times, the entire edge shared by those rooms are passable. You know, mixing things up
We have thought about adding such a faction control system, and it's still something we'd like to implement, but we expect that it might still take a while until there are enough players for both factions to be well-balanced for such a direct conflict to be fun. Until then, we stick with an indirect version of it, which is easier to balance without feeling cheesy.
The level is generated on the client. However, the seed for generating the map is sent from the server. In this way, we generate the same world for all players, but we don't have to send the entire level. We plan on regulary changing this seed (say, once a month), so that this place changes regulary. Hopefully, this encourages players to talk about what they find!
To be perfectly exact, there's actually two seeds: One for the rough world geometry (mountains, valleys, water, that kinda stuff). And one that determines placement of objects, such as trees, rocks, etc., and - perhaps more interestingly - that defines the order of these "rooms" that you traverse the world in. It also affects the placement of dungeons and enemies (therefore, also the difficulty of enemies).
We decided for this approach so that ALL players would see the same world, roughly, but the two teams would traverse and experience it differently.
Reconnoiter, BlackJack, 3run : Thanks for your interest, and nice work on your 3D art Blackjack.
Error014 : Progress on DD is looking great. I'm looking forward to the next beta update. Keep up the great work.
NIGHTMOON SAGA : The End Of Never Dungeon Crawler RPG - New Design Update Recently I have been unhappy with my current visuals, HUD and Battle system design. It was just too common and didn't have much uniqueness to it. I decided to completely redesign the overall mechanics and game visuals to get back to my original vision. (Meant to be a tribute to Arcana SNES). Also updated the dungeon textures to be more colorful so that the levels blend in better with the 2D artwork. After reworking the foundation of gameplay, story and designing a more unique and user friendly battle system, I am finally happy with the new design and can now comfortably continue progress and start getting into some more serious development on the project.
New Title Screen : Meet Vhan and Rue, the two Protagonist of Nightmoon Saga.
New Dungeon HUD Design : This is the dungeon HUD menu. Next I'll be working on the players menu system and map menu.
New Battle System Design : The battle system is now more simplified, however it is also now a real-time battle system using an ATB approach. (Action turn-based)
Hi error! Im working on a sequel to my game Swordlord!
I am working on a shader, and I wanted to make something that had a bunch of effects on it: diffuse,normal,specular,SSS,rim light, so I thought a crystal thing would look nice. I will be making more test models
Hi error! Im working on a sequel to my game Swordlord!
Oh, nice! Looks like you're upgrading the visuals quite a bit! When will we hear more?
Quote:
Any new screens for Dungeon Deities?
I can do you one better: Here's the current draft of a new Trailer! We'll have to cut it down to get it shorter -- especially at the beginning -- but overall, this is a good start, I think.
(Well, I say "start", but that's already taken a lot of work..)
Hey error! Honestly I dont know when, its all very freeform until I finish the framework stuff. I would like to get into some small vlogging/dev videos more
Thank you for the trailer!!!! I love it my dude!! The humor and overall vibe of this game is awesome. Are you putting your screens and trailers on reddit/steam/twitter and all that?
da ich schon mdl von blender aus exportieren kann will ich auch gleich weiter mit blender tools machen um die herstellung von modelen effizienter zu gestalten. am ende werde ich wohl alles in ein riesiges addon stecken mit dem man sehr schnell für 3dgs modele kreieren kann
da ich schon mdl von blender aus exportieren kann will ich auch gleich weiter mit blender tools machen um die herstellung von modelen effizienter zu gestalten. am ende werde ich wohl alles in ein riesiges addon stecken mit dem man sehr schnell für 3dgs modele kreieren kann
, did I understand it correctly that with your tool we can export Blender models with anims etc to .mdl with only ~1 step? (because that would be great )
yes with my exporter you can export a model in blender directly to MDL7 file with vertex weights =) but at time its a beta. to time have the export in world position from blender a bug. all other funktions works fine.
to use in blender 2.78(newest version): select in blender you objects(only Mesh) and not more then 1 Armature then open the .py script in a blender text editor and press run. then you come to the export panel. make all your settings and export it =)
Im still working on a single shader, called the Hero Shader, that has alot of effects in it, all based on multiple maps, multiple channels Its for Dx9 currently, pixel shader 3.
A big special thanks to Txesmi for help, advice, and snippets!
Nightmoon Saga : The End Of Never I've been working on my custom dungeon editor, level changing, dungeon theme ideas and the main menu system. Here's a video of some current progress. (Video has a small lag from recording gameplay.)
McAfee Web Adviser is now flagging the download page of the engine as "risky". Not sure why the page was added, but it's a process to get removed for McAfee.
It's great to see all the active work and activity. Surprising but great.
I'm Playing around with 3ds Max... I was going to do a video comic book 6-sec clips like a Vine but all on a webcomic page in panels. It's a bit too much for my system.
I can be doing environment modeling and stuff if people need. Or for cutscenes. No actors - I don't or animate humanoids. IT's B.Y.O.M/S Materials/Shaders if anyone is interested in game environments.
isn't that a humanoid-like animation above? wink It looks smooth.
I can see the confusion. Yes, it's humanoid. But it's not typical of a human model. Human bodies are modeled as one mesh - This is, in fact, a collection of hex cylinders that are both unattached and not welded into one object. They are simply held around a 3ds max biped. As for the animation - it's 3ds max built-in footsteps. I didn't do it, I just assigned it.
This model and animation is an experiment in art styles that are as they say "Simple and easy to produce". It's not even correct like the modeling being done on "Zeal". It's more the Rayman idea of simple and quick.
Added color correction curves to my image processing application
(image upload site scaled the image even though I checked the "don't scale" option )
Now the app has most of the features that I want: dither, blur, blur with threshold, sharpen, the curve stuff above and a somewhat advanced HDR filter including (local) level-, contrast- and saturation-correction.
Really happy with the outcome so far. The code, however, is all over the place. Ignoring the large interface framework, it's a 3000-lines-mess all in one file which makes navigating the code super annoying
Do you have an email for LocoweeD? Is he still around? He helped me learn quite a few things when i was 15-17 and really would love to reconnect.
If anyone is still connected with him, will you reach out to him to connect with me?
I miss those days.
Also, does anyone need a website for their game? I'm graduating for my bachelors in computer science in a few months and would love to make a website for the right game for my portfolio.
I've been working alot on some of the particle effects recently, so I made some screenshots of some of the effects:
upper left: the british mortars upper right: impact of a tank grenade lower left: impact of a mortar shell lower right: impact of a mortar shell (just a second earlier)
Most of the effects are a mixture of multiple effects. The impact of the mortar shell consists out of four different particle functions: One for the explosion flash, another for the dark smoke streaks, one for dirt swirled into the air and one for a circle of smoke moving outwards.
You can also see the ragdoll mechanics on the lower right picture, but it's not working properly yet. I find it difficult to use the more-dimensional joints of PhysX.
Still working on the AI of the british units. So no new units yet...
Playing around with experimental graphics stuff. Nothing serious here, I just messed around for fun and thought the result looked neat.
Done using a special ray-marching approach. It's actually an "improved" version of the black hole shader I made some time ago. This one is not realtime though (at least with the number of samples used in the above screenshot).
I've never been a big action-figure fan but I have to admit this one turned out pretty good I like the little bit of blood splatter. Gives it a nice detail withoug being exaggerated.
YRPG Toolkit - Player Interface Creation - 12/11/1702:47
Worked the last two days on a small tool for YRPG Toolkit: Fast Player Interface creation. Create your very own menus, cutscenes or minigames in real time and use them ingame. YouTube Video with more info here:
@3Run thanks for the kind words. I was unable to get work on the other forums and basically got beat by the more professional voice actors. It was a fun idea but it's hard to complete against people with real training and better equipment.
Been reading about machine learning lately. Asked myself, if it would be possible with 3DGS. Tried my luck and voila! It's not fast (obviously), but it turned out better than I expected! It's still "learning", but it can already detect most good-written digits:
I'll let it "learn" a bit more, then I'll upload it with the code so you can have fun with it as well. It was just planned as a prototype anyway. Currently this network has two hidden units and a 10 nodes output layer. I think this will be enough for this task. Also, thanks to MNIST for the handwritten digit database.
I've just noticed something... my first Superku prototype (the one from 2009 before the complete reboot in 2011) I used the infamous A4 warlock.mdl temporarily for the first enemy code that I wrote. I replaced the file but copied the entity over to new levels. Apparently... I'm still doing that, in a more or less new project, copying the same enemy over and over:
jumpman, your video is beautiful. I hate the gray look in games 2004, I have realy a problem with the 2004 grafics like in Battlefront2, I don't know why
Working on framework, but now im playing around with some cool looking basic stuffs:
Multitexture Terrain:3 images, with dynamic shadow, specular, normal mapping!
Soft Particles from Wiki!
G buffer Glass, with Specular! Using the soft particle tutorial from the Wiki, I was able to use the lines that sampled the depth buffer view, and make them sample the G_buffer scene. Then, after sampling the G_buffer scene with a poisson filter/taps, I was able to blur the image! This technique however is not very versatile, since objects will sample themselves, since they are drawn in the same frame Thats why if you have complex objects that are transparent, youll see flickers and distortion in overdraw areas, since the model is sampling itself and drawing on itself! It looks good from far away.
Experimental SSAO! Through much trial and error and alot of basic principles from online tutorials, Ive made my own SSAO filter using just a depth map buffer. Its very inefficient, but Im very proud of it, because I was able to apply the basics, without too much math or additional buffer views. I am very proud of this one, and will most likely offer it as an option to toggle on or off in a final game. There are alot of banding in flat planes, but other than that, the effect looks good to me!
for comparison without SSAO:
Bloody Dissolve! Objects can dissolve based on the alpha texture of one of their ent skins
If you revive a game project and find a man among your testers who was one of the first speedrunners, who has been holding at least 2 records for 20 years, and who is still an enthusiastic gamer, it's a great and scary thing. If you then hand out the trial version and just receive (almost) unrestricted praise, it warms your heart in places you didn't know existed. All the drudgery of you and your team might have not been in vain.
It's definitely great and unique. Where /When are the other 12 levels? *greedy eyes* (...) The introduction makes more sense now: "no holding hands, no constant pampering. An ego-thinker,". Okay,... good. It's gonna be incredibly stressful. Peaceful, harmless, only time-consuming puzzles... as more and more chaos breaks over the player with immortal and/or recurring monsters and far too little ammo and no unarmed/ammolous combat.... The only option is to speed/flight.... I'm loving it more and more.
@Kartoffel Did you use the tangent in order to solve the map coordinates (x=pan, y=tilt)?
Well the rendering is 'fake' 3d. The planet is rendered in a 2d post-process shader by using an equirectangular projection which maps the planet texture to a sphere in screen space. I think I'm using the generated normal of the "sphere" (it's not a 3d sphere but just some screen space math). The rotation of the different layers (planet, clouds) is done by simply adding an x offset to the coordinates. The other UI elements were drawn on top of it. It's all done in "processing" (that java library/ide thing), although the main aspect of it (planet, stars, lighting) is just a big glsl shader.
The assignment was to make some kind of "artistic" visualization of time. I just went for a dynamic planet (because it's rather easy to get something that looks good, haha) with a stylized clock around it. Here's a screenshot:
There's also some slight, fancy animations. But I didn't feel like recording a video. ( Also ignore that the day of the week is wrong )
Lots of new stuff added! Commentary video coming soon!
1. Translucency/Glass/Refraction!: using a second view, I am able to use this view as a transparency view for shaders! In the shader, I can distort the scene based on RGB values, on top of the normal map! I can also do frosted glass by sampling the scene multiple times in the shader using a poisson lookup! Objects that have flag1 will be rendered in the transparency view, but they will have only their backfaces drawn. This will allow you to have glass objects that you can see inside of!
2. Ocean shader: using the above refraction, I am able to make an ocean/water shader, that moves vertexes up and down, with 2 normal maps overlapping each other. It has normal mapping, specular, cubic environment mapping, soft edges (from the soft particle shader from the wiki), and it has a foam shader! I can specify where I want the foam based on the G channel of the one of the entskins! The ocean can also receive realtime shadows!
3. The intersection shader: seeing sample code online, I was able to make an intersection shader, for things like shields and stuff. Using the depth camera view and transparent objects, the shield can light up objects that are intersecting the shield!
4. Dynamic Lights!: nothing too special, just copied and pasted code from the wiki into my shaders, and with some fudging, I got it to work!
I've done creating a tile based level from .png image (my editor became useless, since you can simply draw stuff in Gimp or such). After creation I merge all tiles into one model (on the picture you can see 3 entities created, one for collision models, one for visual darker faces and the last one for bright faces (fake shadow/light from sun)). So far I do have problems with setting different skins, for merged models (I could only set U, V - in MED, so I need to have separate model for each texture.. that's kind a odd and I don't want to create everything dynamically). Thanks to MasterQ32 and TUST in general (to all contributors as well!). Next thing to add is pathfinding.
These shots are a part of my RPG Creation Toolkit aka Yume RPG Toolkit.
First screenshot is about the Environment Editor, which includes a tile based map editor, a terraineditor, entity- and eventeditor. Flexibility is keyfactor number one. So it's possible to import and use very own resources. I've also created a Map Format, which will be saved within no time by clicking on the disk button.
Second screenshot presents HOG, which creates the environmental atmosphere. Sun effects, Fireflies, Glowing Dust and much more can be add to a created map within some clicks. Of course those effects can be used dynamically for every created map.
Third screenshot is part of the Quantum Director. With this little tool it's possible to create "Particle Objects" aka Quantum Objects - in order to improve the atmosphere of the created game a quantum more.
The fourth screenshot is a ingame screenshot of what happens if somebody uses the named tools.
All these tools reached usable and stable state.
There are some more tools of my toolkit I didn't mentioned. Maybe I'll reopen my old thread if somebody is interested.
I have added the paint function and a console window.
Click to reveal..
Code:
function manipulate()
{
VECTOR temp;
var nowpaint_x[900];
var nowpaint_y[900];
var j,i=0,k,m;
while(1){
if(mouse_moving == 1&&mouse_left ){
if(i<=899){
nowpaint_x[i]=mouse_pos.x;
nowpaint_y[i]=mouse_pos.y;
i++;
}
//wait(1); //line is faster, slimer, longer and have more draws in ticks
}
k=i-1;
m=0;
draw_line(vector(nowpaint_x[0],nowpaint_y[0],0),vector(0,255,255),0);
for(j=1;j<=k;j++){
draw_line(vector(nowpaint_x[j],nowpaint_y[j],0),vector(0,255,255),100);
}
draw_line(vector(nowpaint_x[j],nowpaint_y[j],0),NULL,0);
wait(1);
}
}
konsole Window starts with a AsCII A8 logo and comment my key's
all updates and progress were packed into a video, as the list would be too long to post a life badge from the project after one year.
but here is a summary: STORY 100% GAMEPLAY 70% WORLD 85% NOT FIXED BUGS [OVER 9000]
@jumpman Did you already succeed in combining the shaders? in relation to the waterfall. if you succeeded well I would be prepared to pay you, if I may use something like that, assuming this is not a shade-c <
ve Found something similar and made a screenshot of
I do not speak much german but that looks awesome.
I have been following your game for a while now, great progress. And for it's atmosphere/genre it's a perfect game except for the animations. Too distracting and stiff, blows out all the effects/atmosphere. Hope you can find someone to go over that enemy/monster animations.
Working on combination of HellGhasts and 3runs ragdoll (v2...again thanks you 3run) with my bodypart gore script. Ragdoll will take care of Body deformations now. For example if u cut off both legs, ragdoll model has no legs too. Excuse the rough WIP blood splats...shadeC makes this really not easy in this case.
Its pretty funny to watch i have to say. ...still heavy wip I will make a post in Projects soon so u guys can see it in motion
edit:
all wip. Last frame shows the floor blood is getting better. Remade the texture. Problem is u cant use transparence here
Next simple update will be red glowing eyes of zombies
edit: Its possible to cut off more parts of dead Zombies. Weapons/movement causing force to bodyparts/ragdolls (physicX).
Last frame shows the floor blood is getting better. Remade the texture. Problem is u cant use transparence here
Next simple update will be red glowing eyes of zombies
edit: Its possible to cut off more parts of dead Zombies. Weapons/movement causing force to bodyparts/ragdolls (physicX).
, yeah encountered the same problem with the transparency and gore with shade-c, still I like your results. Did you use seperate models for the legs/limbs etc?
Worked last days on a doom-style pumpaction. Wasnt that easy cause of the reloading ( even if i already had done this weapon once in my "oldschool fps" ). Bullet by bullet. It fires "real" shotgun bullets ( random impacts ). The shotgun has the flashlight mod applyed in this pic.
And a few minutes ago i tested it with the gore System...great ...iam happy. still heavy wip. Finally in next weeks i will make the vid. Some blood sprites too big right now, like i said, wip.
edit flashlight mod update (new model etc)
maybe i go back to first flashlight model and add some armor tape stuff
edit: Next mod: More bullets loadable/bigger "Clip"
MY CURRENT PROJECT UPDATES : Some updates on the current projects that I have in development.
Retro JRPG Prototype : A basic retro JRPG that I have been casually putting together. Created in the style of old-school games such as Dragon Warrior/Dragon Quest. Nothing too special at the moment, but there is a playable demo to allow a walkthrough of the small island, town and tower areas. The menu system and battle system are complete and all items/skills are unlocked for testing purposes. At the moment you can explore those areas, browse the menu and battle some enemies. I'll eventually get back to development on it. All resources used in the demo are either "Open Source" or "cc-by" license. Download the compiled prototype : Retro JRPG Demo
Evonet : Online Project Development Template : A project that is currently in development and there are a few details to finish before its ready. There is still a lot of beta testing to do, but here is a screenshot of the example template after user login. The template uses a secure connection and all features are up and running. I'll have more updates in a week or two. More information can be found at : Evonet Website
Nightmoon Saga : First-Person Dungeon Crawling RPG Development is currently on hold for nightmoon saga. I plan to turn this into an online game using my Evonet system. But for now, I need to finish and secure my evonet template before I can start back on it. Nightmoon Saga Website
3run : Default.c script is set and it's resolution settings are not correct for the project. I'll fix that with custom resolution settings in a new update.
Emre : The project uses my pixel collision system. Adding movement with mouse should be easy to add. I'll try to add that in the next update.
Time for me to also show some life signs This is an experimental split screen shooter. The foundation is similar to my old project Sky Titans. This project will be a multiplayer only project. There is still a lot missing, but I'm glad the split screen part worked. I played with three friends who had never played such a game so don't mind our bad skills
@EVO I just love your JRPG Prototype. It feels just unique and special next to the gameplay of a RPG Maker game.
@alibaba Seems to be pretty stable. I think it would be pretty funny to play this game with friends. Are you planning to create different roles or will it follow the old fashioned way like Quake and/or UT. Though - this isn't old fashioned at all. Good work!
@ChrstphFr Is this a real photography or ingame graphic? Wheter it's the one or the other way - looks cool.
I am in the middle of the bug fixing of my JRPG Toolkit.
The demo sequence of the video was created only with my toolkit's Frontend. The Core Engine interprets all saved Event Scripts and Game Objects which were created with the toolkit right after starting the game.
So I didn't write a single line of code for this preview.
What you can see are two different maps and changing environment effects. I've also tested some toolkit variables and conditions.
Playing around with shadeC evo and level design...
Just playing around...testing a bit...leveldesign is always hard and iam pretty lazy and have no special idea in mind. Thats why i go sleeping now i guess.
Thank you. I liked your lake scene too, pretty detailed.
Yes this is cause iam playing around with light settings right now. Guess i will setup a night scene with fog and glowing lights. Maybe some kind of red light district.
@Emre: You should turn down the shininess/ specular on some objects IMO, esp. on the stones near the water. Additionally, those stones lack some sort of transition or blending into the sand. You could blend for example into a sand texture based on the stone's y-normal, preferably not with a direct blend (that looks like an alpha blend) but with some sort of irregular dirt mask as a multiplicator for the blend value. Have a look at the following references I took some time ago:
Damn it's too hard. i know this is not exactly what you said but all i can do is this for now. Still working on it. God i wish i was a good developer with good english. Sometimes i can't understand programing language. it can be difficult to find the right method. I feel myself trying to describe an elephant in a dark room. I mean... like these guys.
Much better! However, sand is not transparent and this is why the transition looks a little strange or game-y. You could sample another rough texture now and multiply your blend factor variable by it (one channel) or probably better, subtract it, like this:
Code:
blendFac = what you currently have, maybe times;
float warpFac = tex2D(warpSampler, InTex).x;
blendFac = saturate((blendFac-warpFac)*10); // try different values here
float4 color = lerp(sand,stone,blendFac);
@Emre May be something like a single entity of a bunch of rocks scattered on the Z=0 plane that has their center XY offsets saved in the secondary UV coord set. Each copy of the entity has its own rotation matrix so each rock can be rotated around each own center but all together on each entity. A scalation of the down vertices in reference to its height expands the mesh and can look half-buried. If you use the same texture, orientation and size of the ground texture into the down part of the rock entity, it will look fully merged.
*edited
A XY offset of the down part in reference to its height could emulate an orientated accumulation like in slopes or the effect of wind.
When the gradient is 0:rock <> 1:ground, multiplying the ground color by the gradient could be a good looking effect.
@Superku: That did the trick! Thank you! There are still some problems but i think I will overcome them.
@txesmi; Your method exceed my abilities a bit. But maybe I can learn something new, while trying to do it. So i will try your method. Thank you so much!
Working on my topdown project again. Leveldesign. Right now iam working on the diner park place. Still heavy wip. Now placing some details.
This "faded" entry took me the last hours
No ShadeC. Only HeelX's SSAO and some Slin movement blur pp's used. Guess ill remake the pre-rendered Zombie-Sprites soon cause the level is starting to look really cool for my taste. For the multitextured Terrain i used a seed map to create grass and plants.
i couldn't implement tsexmi's method yet. Because the shape of the some rock models are not suitable for that method. But i think the environment is better than the old one now.
Worked this week all nights ( after working in real life ) on my town map for my topdown arcade project. Iam very happy for now. Still working on the basic elements of the map. Soon ill add more details like dirt on streets etc...the real building fun then Town will be quiet big.
First Zombie is 100% ready and features stuff like eating corpses ( 3 test eaters in the shot ). Will post some stuff in my topdown Project Topic in near future.
Player 100% ready and two weapons, machinegun and shotgun.
First HUD objects like cash kills and health working. The heart is beating. If health gets lower heart will beat faster.
Some buildings can be entered now. Then the rooftop smoothly blends out / in.
Of course the camera isnt this high ingame, i flyed up for the shot. Complete game runs really smooth. Fps are limited to 60 and all functions (npc+player Controller, pathfinding etc) running with 0.2-0.4max Stresstest with around 50 Zombies (on screen) went well too.
No kinematic or character controllers used, only rigid bodies and some tricks for gravity! Still undone, but basics are already working and I'm pretty happy with results! Going to share this, when it will be finished.
Any chance you can finish your code until June 29th and make it open source?
Hey guys!
You can watch it right here. You can download it here.
Basics are already working, BUT To be done yet:
Quote:
-crawling -some tweaks
About tweaks, f.e. right now gravity is turned off when player is on ground surface (any kind of surface, doesn't matter - terrain, level block or another rigid body), but you may want to create f.e. seesaw or similar effect, when controller can affect other bodies with his weight. To archive that you can make gravity trace ignore all other rigid bodies by their group (and customize few other lines of code which are effecting disabling jumping). I'm not sure if I'll add this, cause this probably depends on what kind of movement you want to archive.
I've added some tweaks to it and it works better (for my taste). Tomorrow I'll add crawling and then will upload newer version (probably in a separate thread).
OK, two early (!) screenshots of a project I am working on: 'The old Millerntor Zollhaus' (customs examination/clearance house) in 1806. This is short before the new Zollhaus was build, the one you can still see today in St. Pauli. I intend to create an alive situation when the player approaches the custom control scenario and customs officer are going to stop and check the player.
THX for the info, Quad, and correct adresses. Usualy I link to my pictures the way you wrote. I dont know what was going on and why the enhanced cloudflare/bmi parts were put into the link to square seven. I did not add this adresses to my link, just used copy and paste from an older link of my pics from the forum very late in the night.
Maybe square seven uses cloudflare or it was added when I enterd the adress here on this forum, dont know. In my browser both the adresses are showing my sreenshots corecctly btw. Uhh, I hate this hazzle dazzle :-)).
Anyway, thank you both very much for your info and help.
Playing around random charts of unique colors that are built by the mix of themself, so they can work as develpment tree of a game. Chart building function is flexible enough so charts can hold arbitrary categories, members by category and ingredient count. They are also balanced by production time so its last member total cost is rationalized.
Just a small "how to play" which was created for the community of my project -> Battle System and Class System of YRPG Toolkit. #mostcomfortableengineever (=
Merged meshes rendered with up to eight lights per room and procedurally backed shadows over two bitmaps, one channel each. It distributes the second uv coords over the needed count of pairs of shadowmaps so it mantains a uniform pixel rate in the whole maze.
I merge the meshes with the MasterQ32's DynamicModel module of TUST. I create the secondary uv set in the 3d editor, export it to MED ans save the buffers with a script in custom files. Then I modify the buffers procedurally, model by model, just before merging them up. The process is pretty simple but the truth is that I worked on both uv sets a lot in order to make them fit good with the pre-backed AO, normalmaps and shadowmaps resolution. I have to admit that I am amazed with the result. It deserved the effort.
Worked all night this week on ragdoll, floor Impact, blood, but most times on applying Forces ( from gun and explosions etc ) to ragdoll bodys. https://youtu.be/YhwE9k2RAl8
The nice explosion optic (visual effect) was made by 3run and i used Hellghasts/3runs ragdoll script (v2 not public now) to start with.
Dont forget the Video above is from dev-state. In the Project thread there are some newer Videos. Next time ill make a new Video to Show new weapons, gore and ragdoll progress.
Bumping this thread again, now with fully working version of what I've posted above Youtube link Thanks to Superku and his awesome hitzone contribution!
@3run: Neat! I've once coded something like that, looked pretty much the same. Nowadays I'd probably go for animation blending but... who's gonna make those, oof!
@txesmi: The zoomed out view looks nice, always like me some procedural stuff!
A slightly noticeable but huge update. Animated shoreline and palms. Five layered tiles in four lod steps. Four channeled floodmap for pathfinding. An gc-out event system by downsampling. Fast tiles write/read (locking a 2kx2kx4x32b takes more than 20ms in my oldy machine!).
Since units are pixels, they are moved by finding the surrounding highest value in the floodmap balanced by a speedmap, so they are moved pixel by pixel and describe the fastest path to the nearest emitter in the floodmap. A* pathfindinng works similar.
That ist a smal town or labyrinth editor and a loader, i have make from weekend to yesturday. it can choose 3 kind of tileds with key_cur button and save/load (key_F8) I have 3 kinds of buildings; house1x1, bigHouse2x2 and street and a simply run script I have make a projektfile with 2 exe files to download. https://www.file-upload.net/download-13405948/T-editor_nov27.cd.zip.html it's the download button without the cloud-pic maby you must klick 2 times on the butten because advertisement
@Hellthunder Thank you! I love your magma shader! Ello is awesome! Will your magma be solid or liquid?
@CocaCola very cool! Will you make it randomly generated?
Here's the latest on what I've been working on,
PARTICLE SHADOWS
I've made emulated particles cast and receive shadows! Each emulated particle samples the depth render view for soft edges. There is a particle camera that only has emulated particles rendered (each of these particles is the child of the particle view genius). In this view the particles are shaded with a depth shader like the shadow camera view. Then all the relevant shaders combine the depth check of their normal shadow camera, with the particle camera. I've hard coded all particles to have a fixed transparency for their shadow.
The water shader can now receive particle shadows at the surface, and the translucent view (what you see through the water but also distorted by the water's normal map). So you can place particles or objects, and there will be a shadow at the water surface, as well as on the bottom of the water floor, since the transparency camera uses the same shadow calculation!
@jumpman your shaders skills are getting more and more impressive with each post!
I'm currently working on a floodfill pathfinding algorithm (thanks to Superku, in the past he poked me in the right direction with arrays) Here is a video on my youtube channel (click screenshot to redirect)
Edit2: newer version arrived Same, here is a video (click screenshot to redirect)
Edit: I got 1000 function calls (it's called recursion I guess?) to find a path on an empty 4096 tile map (64x64 - from upper left edge to lowest right edge), what do you think about this results? Is it really too slow for this kind of pathfinding, share your thoughts please.
@3run As a recursive function, it seems like it must be. I saw your function stats leap some ms on the video. Too bad. You will probably need to cut it in some frames or somehow enshort the process.
Do you use pointer offsets instead of array coordinates? It can speed up things a lot.
There is a way of speeding up short paths but unfortunatelly it is more expensive on long paths. It consists on a list of candidates, where you include new candidates as last member in the list and you take the first in the list, one by one, in order to analyze and add its neighborhood to the candidates list. This way the algorithm checks the cells sorted by the count of steps to the end, so it only runs until it finds the ending point, flooding up only the cells closer than the shortest path length. All this stuff can perfectly fit in a single function.
Another way of saving some cycles is to save the direction of the incoming neighbor into the cell struct, so with an offset array like the following:
Code:
int cellOffsets[7] = {1, ARRAY_WIDTH, -1, -ARRAY_WIDTH, 1, ARRAY_WIDTH, -1};
you can totally avoid the cell the flood came from. Same for the subsequent path-finding process. It is just a little bost but every cycle takes into account on hard processes.
I saw your function stats leap some ms on the video. Too bad.
Could this be caused by calling 'draw_text' for each tile? I use it for debugging purpose... Without it f.e. fnc ms goes down to 0.3 from ~8
I'm using something like this - array[tile_size * i + j]. Not sure if it's called an array coordinates (I don't know about offset either unfortunately, you code above is over of my dumb head :< ), or not, but this is the way I access the tiles that I need. I don't use any lists at all right now. To find the shortest path (just found out that it's called Lee algorithm), I simple check all surrounding nodes (tiles) starting from B to A (from finish to start) and increasing each node's (tile's) price by 1 (no G, H, F costs, only one price/cost per tile). So finish node has price of 0, then all surrounding nodes have price 1, then all surrounding nodes around those surrounding nodes have price of 2 etc, till start node is found. Walls and already checked nodes (with price != -1) are ignored. At the point when start node was found, I stop all flood_fill functions and trace back the path, by going from the highest price number to lowest (got this algorithm idea from one of those maze solving mouse robotic videos on youtube). This works pretty well, but it seems to be too slow... Also I got an advice from MasterQ32 to avoid using recursive functions and to take a look into dijkstra algorithm.. It's better to use lists for such stuff, but I will need to implement list usage by myself in Acknex, and that sucks. :<
Could this be caused by calling 'draw_text' for each tile? I use it for debugging purpose... Without it f.e. fnc ms goes down to 0.3 from ~8 grin
Sure those constant 7/8 ms is the drawing function, but I meant when you click to solve, it grows to more than 10 caused by the flood I guess, but I can be wrong.
You don't really need a list implementation. Two pointers and a linking member into the struct is enough.
Here goes a fast and untested implementation of the flood. I was too lazy to build the canvas xP
Code:
typedef struct CELL {
int depth; // <0 = wall, 0 = unvisited, >0 = visited
struct CELL *parent;
struct CELL *next;
} CELL;
CELL cells[4096];
BOOL flood(CELL *_cStart, CELL *_cEnd) {
int cellOffsets[4] = {1, 64, -1, -64};
CELL *_cListFirst = _cEnd; // candidate list
CELL *_cListLast = _cEnd; // at the beginning it only contains the ending point
_cEnd->depth = 1; // starting depth
_cEnd->next = NULL; // clean the linker
_cEnd->parent = NULL; // clean the parent
while(_cListFirst != NULL) {
CELL *_c = _cListFirst; // take the first member of the candidates list
_cListFirst = _c->next; // remove it from the list
int _newDepth = _c->depth + 1;
int *_off = cellOffsets;
int *_offLast = _off + 4;
for(; _off<_offLast; _off+=1) { // 4 directions loop
CELL *_cT = _c + *_off; // get the neighbor
if(_cT->depth != 0) // it is a wall or already visited
continue;
_cT->depth = _newDepth; // set the depth
_cT->parent = _c; // set the parent
_cT->next = NULL; // clean the linker
if (_cT == _cStart) // end when the starting point is reached
return TRUE;
// include into the candidates list
if(!_cListFirst) { // if there are no candidates in the list
_cListFirst = _cT;
_cListLast = _cT;
} else { // otherway, set as the last member in the list
_cListLast->next = _cT;
_cListLast = _cT;
}
}
}
return FALSE;
}
As you can see, it is pretty short and, let me praise, the path is solved by the chain built by the parent member.
but I meant when you click to solve, it grows to more than 10 caused by the flood I guess, but I can be wrong.
I solve by hitting space key, not by clicking. But I think I got what causes it to increase > 10... It's probably caused by the way, I change the skin on the clicked tile. I create bmap and fill it with color each time I click (yes, it's an awful way to do this, I changed it, so now I create 4 bmaps right at the start of the engine, then change tile skin via ent_setskin, instead of creating them each time at run-time).
Thank you for the code man! Going to get into it right now
they are no more than little squares facing the camera in a single entity, all procedurally created. I rotate the faces with the camera rotation matrix. Believe or not, it is much more expensive rendering the multi-layered and alpha-blended dust.
Working on my portfolio piece! Really trying to get hired somewhere for game art/development. As much as there is doom and gloom over this engine not being supported anymore, I still thank the team that made this engine and you forum members. Hard work is greater than what an engine is.
Here is my world with some final art in.
6 texture terrain shader with shadows, normal mapping, fog, specular and dithering. The water shader with reflection and refraction, foam. There will be a waterfall dropping into the lake, and fishes and plants. Currently there are only grass models and dead trees. Cant wait to add more.
@txesmi Looks impressing. I'm a man who loves moving pictures. Is there a way of presenting a small video.
@Jumpan I really like you working result. The details are talking. I would like a smoother terrain a bit more. But boulder models in front of the stretched terrain parts, which blend over with the terrain would improve the look even more.
Really nice water and reflection!
And here my piece of work. Created with my Toolkit which was created with the C# Wrapper for the Toolkit and Lite-C for the Gamecore.
The gameplay was created with 3D Gamestudio, but also without writing a single line of code.
The target of this techdemo is to create a game which uses all features of YRPG Toolkit and complex combinations of Entity- and Event Management.
First Minutes of gameplay of the YRPG techdemo, which will be a real game.
This video shows the premiere of Battle Events. The event editor of YRPG Toolkit can be used to create battle events which will start for selected battles. Battle Events offer the possibilty of an interactive feeling during battles.
Presented game is still work in progress. (Click to open youtube video)
Needed some fast gui solution for a project so I implement imgui feature by feature as I need them. Stupid VAR taking the immediate out of an immediate gui. Annoying casting from and to var so direct pointers can’t be used for entity and other engine object variables.
I had watched the videos 2 days ago, some feedback: Good video, like this game dev commentary. The hammer and the bridge are looking awesome! And a hero shader indeed. I must say I'm not a huge mega fan of the terrain itself though, the rock and yellowish grass textures look a bit low res. Additionally, the UV looks a bit wonky on some steep surfaces. Tri-planar mapping could help a lot (at least for the rocks) but with normal maps and blending this might become too costly. (At 5:42min for example it's fine though, maybe just ignore my nitpicking.) Overall the scene is looking nice though, when you look at it from some spot of the level. There's a bit much movement on the water btw. The fish are cute.
I think I can get and see the style you are going for but maybe some stylization could help a lot in various ways - wait, you mentioned this is for your portfolio. Are you working on an actual game with that scene (like the Swordlord remake/ successor) or is it a demo scene? In case you want to create more content (make a game) I'd think the current style/ approach would take way too much time. I'd probably skip making normal maps then completely and try to come up with a somewhat unique art style which allows to pump out assets more quickly. For example simple stylized textures which you use as a base texture only and then apply various dynamic materials (like you are already doing to some degree on let's say your hammer) instead. On terrain or other things you could setup different local material properties for different areas and do most of your art with code instead.
I can increase the resolution on the textures, I shrunk them in size and format to make them as small as possible.
I should have tried tri-planer mapping when I was first making the terrain shader, Im not sure Im able to change it now It would have solved all my stretching problems (by looking at some tutorials on google)
This is a demo scene, no longer a game. I was working on Swordlord 2 back when I had a job, and I would have made it in gamestudio. But after my team was laid off, I had to switch gears, I didnt have enough portfolio work, and taking however long to finish a full game wouldnt be wise in my opinion. So I am now making a demo scene in gamestudio, so that whomever wants to, they can run the EXE and walk around, looking at the models/textures/animation and programming I did.
The hero shader is the pinnacle of my work, as it has a lot of features in it, however it has a huge production time for an asset, so I use it only for showpiece models, like the hammer tree and bridge. The rest are simple shaders with diffuse lighting and shadow mapping, no normal map shading. So content definitiely will take time to make if everything used the Hero shader. The hero shader is meant for characters, which is what this all was originally for, as I love character modelling and animation. The Hero shader has normal, shadow, diffuse, specular, glass, jiggle vertices, rim light, fake sub surface, cubic environment mapping.
Honestly, all of your suggestions are things I would have done if I was a little more proactive in the beginning. But now I just have to get this done as quick as I can. Thank you so much for taking the time to watch/listen and give feedback. Youre a gem in this community.
@Ayumi I like the weapon models and your UI. Maybe the red color doesn't fit for the information panel. My first feeling was something like: 'Oh - a warning or error?' Is this screen from a game you are working right now? The structure in the background looks interesting too. I want to see more! (=
New Feature of my 3DGS Project:
Volumetric Fog: An effect of the HOG Tool which can be used within an Environment Effect. The user decides how the effect will look in the end. Setting up the color, the density, the spread area of the fog and its height is possible.
@HellThunder Thank s for the hint and words. I think about the red Color. Yes, I am currently working on the game...many many years (start 2009), but start new last year (Rework Code and Level). The weapons and characters created by my brownie.
In my last Level, in the beginnung, i actually wanted to set fog, but the fog of gamestudio is not really appropriate. That's why the level is without fog (Kind of desert fog). I think your generator is interesting for this reason. I like the adjustable style.
A little offtopic: whats wrong with the forum? Am I the only one getting this weird changes in style?
The same here. Especially on my phone it looks all different.
To keep it a little ontopic
Implemented buildings. Like in the old Codename Panzers games you can see the inside of the houses when units are inside. The level is (as always) WIP thus still very empty.
Just for fun i worked on a Resident Evil style engine. The Backgrounds are pre rendered with 3d models on top. Inventory features ( with boxes to drop items ) 41 Slots. Blending in some anims missing now, just a prototype. The Pixel look is a Postprocessing pixelshader ( can be deactivated via Settings.ini file and isnt that strong...the captured Video pixels it too ). All taken items and the inventory are saved in ini files too ( own save functions ). If the gun isnt used its holstered on the back of Player. Inventory works in all graphic modes window and Fullscreen.
Very impressive tech demo. And also the artwork (except for the "background" of the inventory) looks very nice.
On minute 1:05 when the shotgun is shot, it seems to me that only the metal facing to the gun is lit by the muzzle flash. Do you store these information in an additional texture? EDIT: just reread your post. The lighting obviously comes from the underlying 3d models.
But more important I like the overall idea of a retro Resident Evil style game. It still catches the claustrophobic feeling of a zombie infested mansion best.
Thank you for your nice feedback. Really appreciate feedback from serious users. edit: The lightning only affects 3d models ( foreground like Zombies, Player, gun etc ). All other scene stuff is just a Image file. edit end
For pre rendering background i used ShadeC evo btw.
For now i dont plan to use it / make a game with it. Just wanted to see if i could do such type of script, and have to say, iam quiet happy with it so far Greets
edit: Invisible models handle the scene. Cam switching is done via regions set in WED. The shadeC-pre-render-script saves camera pos and angle values from scene in game. This way new made level scenes can be implemented very quickly. Finally a new room / cam view can be set via a single command in script
Code
init_scene (#)
While the invisible models make it very easy to have same collision etc like in the pre rendered scene map. Every scene (set with init_scene) takes care of taken objects like health ammo etc automaticly and saves it. The inventory also takes care of saving "by itself". For now saved files are not crypted, so yes, they can be manipulated very easy. You decide if items can be combined with each other. Also you decide if items of same type like shotgun ammo in the techdemo can be combined to one inventory Slot. If you hover them with cursor a amount box will show up. But i have to mention that adding new items to the existing script takes some lines to copy and paste. But its pretty fast done and well explained by the three example items. Theres not much limitation for moving 3d models, cause like the original resident evil, it renders very very fast. The script also features SuperKu's latest pathfinding with pathes set in WED. Navigating NPC's in the pre rendered scene is "easy" this way. All in all this text sounds like an advertisement Greets
Small update. A letter / info system. Important infos can be taken. Those will be archived in the letter archive shown in the video. U may have noticed i used a new font. Looks better then the old one btw.
A letter-info can have 1 or N pages with text and a unique background image. Once predefined, they can be load with a single command. For now 27 info - slots are ready. Maybe ill expand those slots to around 50 like the inventory slots.
U may have noticed i used a new font. Looks better then the old one btw.
Very True. The new font looks way better... but I think a font like from a typewriter would suite the letters even better. Maybe these are useful: Serifenlos Mit Serifen Edit: all of them are free fonts, so no intellectual property infringements.
Do you read the text for the letters out of a file or are they "hardcoded"?
It seems to me quite a coincidence but I was doing almost the same damn thing just these days:
It is not a letter but a tutorial, showing two arbitrary images and text. At first all of the text was hardcoded, but now I read all strings out of an extra file. Took me quite some time but I hope this will make localization easier, once my project is that far.
Originally Posted by rayp
For now i dont plan to use it / make a game with it. Just wanted to see if i could do such type of script, and have to say, iam quiet happy with it so far
For not planning on making a full game you seem to work a lot on that project
but I think a font like from a typewriter would suite the letters even better
Even more Resident Evil style . I thought about reading from files but then i hardcored them. In my case i dont see a reason for reading them from files for now.
Quote
I was doing almost the same damn thing just these days:
I see yes very similiar work we were doing. Nice looking btw.
Quote
For not planning on making a full game you seem to work a lot on that project
Most times i broke my neck on Content / Level creation. Iam no model maker. I have a feeling for building maps with props. But making the Content / Level is not really mine. If i look to some Levels and graphics in Unity...a tear runs down my face. Its a shame 3dgs is that far away from now a days Standards. And me to lazy to start from scratch again...nah. BUT i still love 3dgs and to work with. Best light weight prototype System ever! Think of how awesome 3dgs could be...with more users...a asset store...but thats a dream. I never complained or posted much in the "Is 3dgs dying" topics...so those lines above i had to let out now And for me it never dies, cause its still a mighty and fast tool for developing game software of ALL sort of styles. But this is getting off topic yet
Greets
edit: btw...the forum post time stamps arent correct somehow.
Hi all! These holidays I have been working in an editor for my little 2d world. I just finished the basics for the tilemap editor, which can create 'brushes' (tiles setups). I have to say that the more time consuming part has been the UI, which I modified to manage everything with globally allocated structs allocated at start. The time for the world editor itself has arrived.
Hi all! One of the unsolved things of my little project was the way to manage a large count of units without slowering the render process to its stop. I found a way that makes use of some views and render targets to draw up to 16K units by draw call. It works pretty well. Each unit walks along its own path which is computed by the CPU on demand of an event manager as output of the rendering process when it reach the path end. They do not collide each other and will not, it would break the whole process. Here goes a demo with 4K units running on a 256x256 grid. I think some hundreds will be enough for my project.
I am very happy that my desktop program now works without any problems. Development still in progress though. Please check this video.
I started making this program for fun. But I think it's become a program I can share with people.
Could you guys help me choose a name for the program please? I wanted to draw attention to the ability of this program to change the desktop background according to the mood, so i was choose "Moody Desktop" but you guys know my english is not good. i realized that "moody" is meaning; "gloominess" or something.
Emre, It looks awesome! I'm no good at giving names (you can see that from the following 3 names), maybe something like 'Сozy desktop', 'Desktop mood', 'Wallpaper animator'.
Thank you, 3run! I would prefer "Cozy Desktop". I've never heard a word "cozy". Thank you for letting me know. i searched google translation, it says: comfortable, pleasant, warm, homelike. This word seems very suitable for my program.
Never tested this on another PC. I am not that familar with dll compiling. Could be, that I messed something up in the linker and referenced some dependencies that are not there.
i hope not. if this is the case, my rebuilt dll will not work on another pc either. Because i just simply build the project after changing the project settings based on my system. (change sdk paths and delete post build command events.)
Ah ok, I have an old Windows XP PC here, I will test it with that later this week and try to limit dependecies / bzw include them into the project when licensing allows redistribution.
If you want to distribute your DLL to other people, it might be a good idea to change Properties / C/C++ / Code Generation / Runtime Library from Multi-threaded DLL (/MD) to Multi-threaded (/MT). Otherwise your users won't be able to use the DLL without installing the VC++ Runtime Redistributable before - one of the funny ideas by Microsoft to make life more interesting.
I installed Microsoft Visual Studio 2010 do nothing. The example is runing. So, I think problem is Multi-threaded (/MT) and some lib support and publish dll.
BTW; i rebuild dll with "Multi-threaded (/MT)" as 20BN mention it. Also build it with release configuration, not debug configuration. Test it another pc and it works without problem!
Under Lite-C in my repo, there is now the updated DLL compiled with MT. Works on my other test PC. The VS Project is still a mess. I should clean that up and optimize the code. Perhaps lern to use some modern tools to compile DLLS
You can download imgui your self and look at the demo included in the source code. Most of the features are available through the DLL. I am working at the moment on missing features. I will recreate the imgui demo in Lite-C, so I don't miss a feature, but that will take some time, because this is only a hobby. And sorry for the different notation of the methods but I dispise Camelcase notation very much and used undescore notation for the imgui methods.
You can download imgui your self and look at the demo included in the source code. Most of the features are available through the DLL. I am working at the moment on missing features. I will recreate the imgui demo in Lite-C, so I don't miss a feature, but that will take some time, because this is only a hobby. And sorry for the different notation of the methods but I dispise Camelcase notation very much and used undescore notation for the imgui methods.
You did a great favor and we have no right to expect more from you right now. Take your time. I really appreciate your generosity.
To other friends who want examples/samples: imgui is not well documented but it comes with some good example projects. As JenGs mention it; "Most of the features are available through the DLL." Just look at the "imgui_methods.h" All functions are there.
Nevertheless, I have prepared a more comprehensive example/sample for you guys. i'm not a good programmer and as i said before; imgui is not well documented. So i just looked at the example project and ask to google for some methods. I hope that will be useful. Here it is:
Code
//
#include <acknex.h>
#include <default.c>
#define PRAGMA_PATH "%EXE_DIR%\templates\images"
#define PRAGMA_PATH "%EXE_DIR%\samples"
#define PRAGMA_PATH "imgui_header"
#include "imgui.h"
//sky_color.red for slider
int skyred=128;
int skygreen=128;
int skyblue=128;
//for checkbox
int draw_test=0;
char* listbox_items[9];
char* combobox_items[4];
void synchronize()
{
//listbox items
listbox_items[0]="Apple";
listbox_items[1]="Banana";
listbox_items[2]="Cherry";
listbox_items[3]="Kiwi";
listbox_items[4]="Mango";
listbox_items[5]="Orange";
listbox_items[6]="Pineapple";
listbox_items[7]="Strawberry";
listbox_items[8]="Watermelon";
//combobox items
combobox_items[0]="Car";
combobox_items[1]="Bus";
combobox_items[2]="Plane";
combobox_items[3]="Ship";
while(1)
{
//if checkbox true show this text
if(draw_test==1)
{
draw_text("CheckBox is TRUE",screen_size.x-300,10,COLOR_GREEN);
}
sky_color.red=skyred;
sky_color.green=skygreen;
sky_color.blue=skyblue;
wait(1);
}
}
BMAP* b1="rock.tga";
BMAP* b2="sand.tga";
BMAP* b3="detail.tga";
BMAP* b4="display.pcx";
// show-hide handle for widnwow 1
static BOOL showWindow1 = true;
STRING* temp_str="";
STRING* temp_str2="";
void window_func()
{
//!!!CREATE FIRST WINDOW!!!
#ifndef create_first_window
#define create_first_window
if(showWindow1)//for VISIBILITY!
{
//Set position and size
imgui_set_next_window_pos(1, 50, ImGuiCond_Once);
imgui_set_next_window_size(700, 400, ImGuiCond_Once);
//Create window --->(name, handle,flags)
imgui_begin("Pictures Window", &showWindow1, ImGuiWindowFlags_NoResize | ImGuiWindowFlags_NoMove | ImGuiWindowFlags_NoCollapse );
//Add image
imgui_image(b1);
//Next image will be on same line
imgui_same_line();
//add image (Same line)
imgui_image(b2);
//Add image
//it will not be on same line because we didn't use "imgui_same_line();" again
imgui_image(b3);
//Add text
imgui_text("Click to picture below");
//Add button image
if(imgui_button_img("Button Image",b4, 200, 200))
{
//event
printf("Test Message");
}
// Always finish with this function (imgui_begin...other codes...imgui_end)
imgui_end();
}
#endif
//!!!CREATE SECOND WINDOW!!!
#ifndef create_second_window
#define create_second_window
//Set position and size
imgui_set_next_window_pos(screen_size.x-400, 50, ImGuiCond_Once);
imgui_set_next_window_size(400, 600, ImGuiCond_Once);
//Create window
imgui_begin("Test Window", NULL, NULL);
//Add sliders ---> (Name, variable,min,max)
imgui_slider_int("Sky Red", skyred, 0, 255);
imgui_slider_int("Sky Green", skygreen, 0, 255);
imgui_slider_int("Sky Blue", skyblue, 0, 255);
//Add button for change first windows visibility
if(imgui_button("Hide/Show Pictures window"))
{
//button event
if(showWindow1==true)
{
showWindow1=false;
return;
}
if(showWindow1==false)
{
showWindow1=true;
}
}
//Add checkbox (Name, variable)
imgui_checkbox("Test",&draw_test);
#ifndef listbox_sample
#define listbox_sample
static int listbox_item_current = 1;
//Set listbox specific width
imgui_push_item_width(90);
//CREATE LISTBOX
imgui_listbox("##listbox1",&listbox_item_current, listbox_items, 9, 9);
//Event for listbox
int i=0;
for(i=0;i<9;i++)
{
if(listbox_item_current==i)
{
//Draw selected word
str_cpy(temp_str,"Listbox: ");
str_cat(temp_str,listbox_items[i]);
str_cat(temp_str," is selected");
draw_text(temp_str,10,10,COLOR_GREEN);
}
}
#endif
#ifndef combobox_sample
#define combobox_sample
//Create handle and set current item for Combobox
static const char* item_current = "Bus";
//CREATE COMBOBOX
if (imgui_begin_combo("Combo", item_current, ImGuiComboFlags_HeightSmall))
{
//Event for combobox
int n=0;
for ( n = 0; n <4; n++)
{
BOOL is_selected = (item_current == combobox_items[n]);
if (imgui_selectable(combobox_items[n], is_selected,0))
{
//Print selected item
item_current = combobox_items[n];
str_cpy(temp_str2,combobox_items[n]);
str_cat(temp_str2," is selected");
printf(_chr(temp_str2));
}
}
//END
imgui_end_combo();
}
#endif
//Child+Column
#ifndef Child_Window_Sample
#define Child_Window_Sample
imgui_text("\r\nChild and Column");
//CREATE CHILD
imgui_begin_child("Child1", vector(300,35,0), 1,ImGuiWindowFlags_NoScrollWithMouse);
//CREATE 3 COLUMNS
imgui_columns(3, "COLUMN", true);
//Set column width
set_column_width(0, 100);
set_column_width(1, 100);
set_column_width(2, 100);
//Colors (for buttons)
static float red_color[4] = { 0,0,100,100 };
static float green_color[4] = { 0,100,0,100 };
static float blue_color[4] = { 100,0,0,100 };
//Color begin
imgui_push_style_color(ImGuiCol_Button, red_color);
//Create button
imgui_button("Red Button");
//Color end
imgui_pop_style_color(1);
//Next Column and button
imgui_next_column();
imgui_push_style_color(ImGuiCol_Button, green_color);
imgui_button("Green Button");
imgui_pop_style_color(1);
//Next Column and button
imgui_next_column();
imgui_push_style_color(ImGuiCol_Button, blue_color);
imgui_button("Blue Button");
imgui_pop_style_color(1);
//End Child
imgui_end_child();
#endif
//PLOT LINES
static float arr[7] = { 0.6, 0.1, 1.0, 0.5, 0.92, 0.1, 0.2 };
imgui_plotlines("Frame Times", arr, 7);
//!!End of second window!!!
imgui_end();
#endif
}
void loop()
{
imgui_start_imode();
window_func();
imgui_end_imode();
}
void main()
{
mouse_sync = 0;
mouse_pointer = 2;
mouse_mode = 4;
video_switch(10, 32, 2);
wait(1);
level_load("");
bmap_preload(b1);
bmap_preload(b2);
bmap_preload(b3);
bmap_preload(b4);
synchronize();
imgui_init(0);
on_d3d_lost = imgui_reset;
on_scanmessage = custom_scan_message;
on_frame = loop;
}
Thanks again, JenGs!
______ Edit: Some of my methods may not be properly.
Hi everyone, just wanted to show my A8 gamestudio environment work, The Valley. I made all the textures, models, animations, shaders (with help from Txesmi and Superku) , level design and music. I spent a long time on the framework, and only a few months putting the level together. Its mainly for my art portfolio, I'm actively looking for anything game art related. Take a look and let me know what you think.
There is alot going on under the hood, and I dont want to make a big wall of text but ill list some things that I've achieved through trial and error and help from this forum.
- trees and foliage have vertex sway through shader - object dithering to simulate a brute force fade away on distance to camera, so that objects in the solid pass can stipple away when far away - soft shadows that fade in the distance - Transparent/translucent water with distortion, foam mapping, wave strength, planar reflections - The Hero shader: Diffuse shading, normal mapping, specular (specular tightness as well), Rim lighting, fake subsurface scattering, baked AO which intensifies in shadow, translucency and self translucency, blurry/sharp transulency, jiggle vertices using spring particles (hair, clothing, not a full cloth simulation though). The hero shader can also do Colored self shadow on translucent parts! - terrain shader with up to 8 textures, each texture can have a normal map for normal map shading, and a specular using the normal maps Alpha - Emulated particles that can cast shadows (hard) and receive shadows - Post processing, pixelation, broken glass, and even my own brute force AO (looks kinda terrible, but was fun)
Just wanted to thank all of you active developers here that helped me, and helped keep this forum alive. Thanks for taking care of me for this long, and I'm always amazed at the great work you all make here.
@txesmi Your 2D progress with 3D Gamestudio is amazing. Do you plan to create a game with your reached progress? @Jumpman. Great progress.The light and shadows play seems flawlessly. The models seem great. For more details of your terrain you could use separated models or terrains for mountain gradients. Helped me a lot. Personally I think some particle fog would be a great improvement for the whole scenario. As I said. Great progress so far - both of you. =)
And here a small video of my JRPG Engine, created with 3D Gamestudio. This techdemo was created to test functionalities and eliminate bugs. No single line of code was used. There a still some ugly memory issues. But time will fix it. (;
A new feature is Emres ocean shader, which can be used within the c# tool now
@Jumpman That looks much better than last screenshots. I specially liked the water-work and the golden parts of the logo. For a portfolio completion, I would miss some LOD work as the essential piece of level design it is. I totally agree with HellThunder about the ambient particles.
@HellThunder Thank you. Yes, there is a game always behind Some sort of production/city manager. But for now, I am working on an editor that would let me test different setups easily.
@Jumpman: I like the stylized graphics of your showcase and your shaders compliment this style perfectly. Edit: How do you blend 8 textures for your terrain? Do you use multiple passes?
@HellThunder: As a fan of the RPG Maker I like your project. How much „content“ (models, textures, sounds ,...) will you ship with the YRPG Toolkit?
@Jerome8911 Gamestudio by default can only send about 4 internal entity textures to be read by the shader. However Dx9 has the ability to read more than 4 textures/samplers. Superku let me know of a material event function that will allow you the ability to send more internal entity textures to the shader, as long as you keep the same names within the shader:
Code
#include <d3d9.h> //include this at the beginning of your script to use the LPD3DXEFFECT function from the d3d graphics library
...
function mtl_event_terrain() //----------------------- the terrains material event
{
LPD3DXEFFECT _fx = mtl->d3deffect; /// use this new function to grab the correct skin
BMAP* doot5 = ent_getskin(me,5); ///get the bmap pointer of the entity's 5th skin
//------------------------------ent skin 5!
if(doot5!=NULL) //------------- If the entity's 5th skin is not NULL
{
_fx->SetTexture("entTex5",doot5->d3dtex); // force the texture sampler to be the bmap of the entitys 5th skin!
}
}
Now in your shader, as long as you use the same exact name, you can sample this 5th texture!
@Jumpman impressive Video man! And iam intrested in this shader mod too. Great to use more Skins on Terrain, really! Maybe u could upload your modified shader to the forum ? Sure many people would like it.
Just wanted to showcase some of the progress made with my RTS with WW1 setting.
Finally I came up with a name for the game: Tactics of World War One. Pretty straightforward, I know. But I guess people should know instantly what my game is about.
And some new screenshots of the first map:
Every feedback is highly recommended! How do you like the title and the game logo? Do you like the look of the vegetation? etc...
Looks very nice. The broken world in the title as allegory The vegetation looks very nice, a bit scanty for my taste but nice. It has an overall stylish looking. I agree with the darkness. I think I would encrease the ambient light and reduce the shadow contrast a bit.
Looks nice. You've done a good job on the game logo and name but the wall and soldiers are to dark from behind. You should prevent to duplicate the Vegetation in one line.
Some progress on my RTS "Tactics of World War One".
Another picture from the first level. Nothing groundbreaking new to see here. Amongst other things a new grass shader for "tilting" the grass away from the camera to make it look denser. I pulled this technique out of a GDC talk about vegetation. The effect is subtle yet quite effective.
This picture shows the third level (or maybe fourth, depends if I add an additional mission prior to this one) in the game. The level is set in a partially destroyed forest with several trenchlines the player needs to fight through. This time I did not hide the HUD so you can get a better impression of the gameplay.
As always: Still WIP but feedback is highly appreciated!
Iam working on my fps DeadTaste again. Meanwhile story takes place in creepy space ship. Zombies are modified and called infected now.
Working on level design for days and atm iam trying to add the creepy feeling/atmo via sounds, flickering lights etc. Also trying to get the fps started somehow. Player arrive and so on. Light can be dynamically changed via single var. Like alert mode. Also theres a flashlight player can use. Trying to aim the Doom3 feel somehow. edit:
Working on level design for days and atm iam trying to add the creepy feeling/atmo via sounds, flickering lights etc.
Although its just a single screenshot, I think you make a really good job in creating a dense and creepy athmosphere by the lighting of the level. Not because it is just simply dark, but because the space between the light in the foreground and the door in the background is almost completely in darkness. And I guess that is the way player has to take. I would love to see it in motion with flickering lights.
And you definitely need an option for gamma correction in the final project I could not see anything at first.
Thank you for your feedback. Its pretty dark yes. But on my devices ( TV, screen and phone ) it seams ok. Ill add something to control the gamma. I also noticed it heavy depends on the view angle to the monitor. This is a taken photo from Smartphone (almost to bright): Same scene but another view angle (almost to dark):
But of course a gamma correction in some kind wont hurt.
as always I am working on my strategy game. The last few days I reworked the code for terrain transformation to fit the trenches into the terrain (as it is a game set during World War One trenches are a rather important art and gameplay asset)
So a lot of games (like Company of Heroes) adjust only the z-value of the terrain-mesh vertices so the trenches fit in. In order to make this technique look good you need a very dense terrain mesh (like Company of Heroes ). Since I did not want to make the terrain mesh super dense (for various reasons, mostly for an easier workflow) I used a different technique and modified not only the z-value but also the x- and y-value of terrain-mesh vertices that are close to the trench model.
I implemented this some time ago but the result was very unsatisfying. The texture is stretched close to trenches and gaps in the terrain mesh appear at various places (red circles):
I recently came back to this and reworked the code. It is still the full vector-field displacement technique but I added some code for borderline cases (the places where the gaps appeared) and additionaly recalculated the uv-mesh after the deformation. The code for recalculating the uv-mesh
Code
float level_extent = 2 * 11400;
ENTITY* terrain;
...
CONTACT* c;
int i = ent_status(terrain,0); // number of vertices
double multiplier = 1 / (level_extent);
for (; i>0; i--)
{
c = ent_getvertex(terrain,NULL,i);
c.v.u1 = (c.v.x + level_extent * 0.5) * multiplier;
c.v.v1 = 1 - (c.v.z + level_extent * 0.5) * multiplier;
ent_setvertex(terrain,c,i);
}
It took me some time to figure this out because the documentation for ent_getvertex, ent_setvertex lacks some information. Especially on the various uv-mesh deformations.
But the reworked code makes the terrain look way more clean:
Still not perfect in every single aspect but I am really happy with the outcome. Next I want to improve the terrain shader. I still think the normal maps don't "stand out" enough and dynamic lights doesn't affect the terrain at all by now.
As always: Feedback is very welcomed, wether it is about the terrain or any other aspect of the game!
Ok, here are 3 screenshots of the tools I am working on:
The first one shows the concept of my world editor tool : A 3 level bmp tile array developed for sketching, hightmapping and texture painting for large open worlds. Special feature seamless painting over borders of neighbourerd bmp tiles via half left and right tile parts (border is then right in the middle. Also features texture painting depending of the heightmap. i.e. painting a valley texture at the position where the heigthmap is 'low'= valley, and other textures where the heightmap is heigh. Quick changing of tile arrays with F1-F3. Hope the pic here is explainable.
Screen two and three show my new translated old Neomill7 endless tile world engine. It is now in Lite-C instead of c-script. Tile treadmill engine (9 tiles) now uses 8192 quants tiles. Demo has absolutely border seamless terrain textures, I used Corvitextu to make the Terrain skin seamles as well as the detail skin. Terrain tile texture is 512x512 only but looks very good because of the detail skin, set to 128 pixels. Simple world editor script is included with selector for model to be created where the players pos. is ( will then be saved in a text file and used when the player reentries the tile he was on). An Action ID selector is also included which determines the Action which shall be used with the created model or sprite. The players facing will also be taken for the model or sprite, can be changed manually by editing the text .dat file in Notepad.
I will release a 3D Demo world in the next weeks when an optimised and tested new engine is ready.
Ok, here are 4 more early screenshots from my lite-c open world desert treadmill scenario. Loading/Unloading routines of hmps and MDLs now work very reliable. Borders of hmps are are not visible.
Desert size 10x10 Tiles. For testing. Tile size 8192 quants. Pics of big (Ayers rock like :-)) montain shows seamless positioning of 2 tiles which had been created with mirroring meshes. By mirroring neighboured tiles you can create big mountains or deep valleys .
Model and Action selector from positionig script work well.
Note: this is a open world engine still in early development status. There is still a lot of stuff to be optimised.
My goal is an engine which offers an open world game engine for everybody with at least the graphical quality of morrowind. And in lite-c for A8 Gamestudio.
Problem with unwanted camera.pan and .tilt changes (because of mouse movent) during remove and reload routines solved. Player looks in the same dirction after entering a new tile.
Sun angle change tested in realtime. Looks nice, but needs probably some tricky level design to avoid visible seams.
Sun light changing tested. Looks also nice.
Still all hmp tiles and MDLs are loading correctly.
Concept for monkey lsland style dialog system (editor) started.
Next time I will move to 'projects with screenshots', yes here is the wrong place I know. Project needs its own thread anyway.
I am working on an interesting project for a long while. This is a GUI Builder system. The problem has been preoccupying me for some time, I can't find a tool that could use to design the GUI of my programs. Obviously in GUI.
Some years back I decided, I am going to make such a tool for me. After many pauses and restart from the scratch I achieved, I have a working Proof of Concept system. Naturally in Lite-C. My older projects may posseses more functionality, but this one joins all my studies and results in one project. What can do this program now?
It is able to load from a file and set the basic configuration of the screen and window. It is able to load from a file GUI elements like panels and buttons. It is able to place them on a workbench where I can modify them. Now only their positions.
What will can do in the future? Just few examples... Will save the workbench content, and loads in later. Will operate the elements automatically, as can. Will use templates. Will be able to handle projects, subprojects that contain a GUI layout, appearance.
Finally there will be an operating environment eg. a DLL or an include library, that will work in my program and will do all the tedious work with the GUI instead of me.
Interesting thing, while I worked on the design and implementation, as a side product, fell out a file format that contains now my configuration and project data. I looks like the JSON format, but differs from that more or less to fulfill my requirements.
a small update on my progress. I finally added the possibility to compose your own army prior to some missions.
This gameplay mechanic can be found in a lot of similar real-time tactics games and it provides the player with some more tactical decisions.
The text to the right changes depending on wether the mouse touches an unit portrait or not. If not, a general description of the mission objectives is displayed.
Working for month on a smooth 3rd person controller and game stuff now for a silent hill style horror game (yes i did not work on my fps shame on me!). Player feats jumping, climbing, hanging and climbing, wade, swim, dive and run. Melee weapons like baseball bats and hand guns like Magnum.44. The animations used for player are from mixamo, some modified. With all stuff like anm blending etc. Player can activate a flashlight (on right shoulder). He can find and use a gasmask and some kind of nightvision with egoview feat. Also i finished some different types of ai already. one is able to jump and walk the ceiling like in the Suffering. After weeks of pure scripting with place holders and setting up shadec evo, i finally started last night with my first visual work. In motion it looks a lot better cause of moving fog and rain spalshes on street etc. For me personal, my best work yet. Iam very happy with the look. Spent a lot time on details like footstep sounds on different surfaces, bloody feets on blood leaks etc. When player aims with a gun the upper body can be moved smooth with mouse while hes still animated. A 3d crosshair helps finding the target. Camera uses obstacle avoidance and moves itself closer to player in narrow situations. In buildings the perspective is a bit different cause the cam is close to player.
For now i made a streets that follows a hill side and a lake (right). It will lead to a gas station. Crows and some other critters already "alive". Took some time to get shadec evo working fully correct but now i love the look. A bit like hl2 for my taste. The gore system already looks nice here and there, used my ragdoll-force-system from DeadTaste fps (like most of the tech in techdemo room).
Also i finished a nice inventory system like in Resident Evil. New items can be add very fast. All stuff like weapon health usage and so on is done in the inventory. Items can be droped in special lockers. Item system uses "weight". When no more room / too much weight, player cant carry more and must drop items in lockers for example. All stuff like inventory system works screen res independent.
Enemys will be some creapy obsessed/demon/zombie monsters. Normal demons fly/hover. They vomit all the time green slime. If you smash their head the vomit will spray out of the neck wound Detection of enemys works with a special instrument called the ghostmeter. A modified multimeter
Also stuff like security camera and a computer system is working. Basic stuff like dialpad closed doors etc. I really like the computers in game. The os - look is like win3.11 in them
For the story i have no good ideas only a direction. Not good at story starting.
I will make a project post for sure cause i guess this will be my final work / Showcase entry...but i know ill never switch engine anyway.
It seems to me creating a GUI builder is a hard task, but not impossible. Now my little pet project is able to load and render project files that describe a gui or a part of gui. Here is few examples.
After start the program.
A project loaded and rendered on the canvas.
Added a button to a selected panel.
Here is the source of the loaded project:
Code
{ two_panels:
content = "panel"
{ panel:
type = "native_panel"
name = "gui_2"
pos_x = "710"
pos_y = "0"
size_x = "356"
size_y = "324"
bmap = "images/panel.png"
layer = "2"
}
{ panel:
type = "native_panel"
name = "gui_4"
pos_x = "350"
pos_y = "350"
bmap = "images/panel.png"
layer = "2"
{ button:
type = "sgb_button"
name = "button_1"
pos_x = "130"
pos_y = "44"
bmap = "images/base_btn_1_bg.png"
layer = "3"
function = "bubu_function"
{ text:
type = "caption"
text = "BUBU"
font = "arial"
size = "12"
}
}
{ button:
type = "sgb_button"
name = "button_2"
pos_x = "211"
pos_y = "20"
bmap = "images/btn_for_panel.png"
layer = "3"
}
}
}
I implemented an object pooling system for my framework. I currently create 100 projectiles at level start and deactivate them. When an AI or player or whatever wants to shoot a projectile, the function will check to see if a projectile is not activated, and if not, activate it, move its position to the spawn position, and do a single line trace every other frame. Groups of 50 projectiles will do a trace within 1 frame, and while a projectile is not doing a trace, it will do a lerp from its last position to the halfway of its current and forward position. I saw Doom 2016 dev talk about fast moving projectiles not needing to trace every single frame, so I figured I give it a shot. So far projectiles look smooth enough.
If the trace hits nothing, the projectile will move to the new position. If it hits something, the projectile can make an impact effect, and deactivate itself. The name of the game now is making sure things arent created and deleted constantly, and reducing the amount of while loops on individual objects outside of AIs.
@rayp Your third-person game looks really promising. Even the quality looks rather good. Personally, I would change the life-points and endurance-points UI - I just call it like that. Maybe for the life-points a fullscreen 2D effect that will be redder after losing life-points. This also offers a kind of uncertainness of the real lasted life-points. Maybe for the endurance, a breath sound combined with an animation of a drained player model would be great. Yeah much work - I know. Just suggestions.
But it really looks great so far. Maybe you think about it and sell a framework for creating Third Person horror games with 3D GS?
About your first-person game - I like the camera view of the first shot. But I can see a small line between the tiles. But I can imagine this as a dramatic sequence.
About the second screenshot. Just nice light- and normal map effects!
@Aku_Aku I know how hard it can be to create a UI Builder - created something like this with javascript and c# and acknex wrapper. Looks good so far. Think about using the lite-C UI Framework for the UI of your tool. Just for the UX of your tool. Would be great to see a video about - how to build a ui - and - results of the tool.
@jumpan Just a cool approach. I will definitely remember that. Thanks!
@Truth Congratulations! I watched your project for a long time. Nice to see you finished it!
@HellThunder Just some news of my Acknex Project - I don't like the name 3D Gamestudio. Acknex just sounds more epic For every image a small video link was prepared - so by clicking on the link you will be redirected to youtube.
A small main menu test with a high poly model. I used the mesh reducer for <64 polygons. I also added some panel effects that the developer can use while creating the main menu.
Chain Creation for Entities. Create one Entity and use it multiple times - even within the same sequence. A small UI and example introduction.
I added the standard cell shading effect to the editor. It's possible to add it to a single entity or multiple entities at once. Acknex is really fast with processing shaders - selecting 100 objects at once and adding a shader is no problem.
It is also possible to connect entities now within the editor.
So Acknex is still alive with great projects - thank you for this effort, guys!
So I've gotten my level flow doing what I want it to do, which is completely randomly generated every time it's loaded. However, I've been stuck for a few months with not having a story to progress development any further. I want do have the player walk around and collect different objects, and depending on what is picked up, will uncover a story of some sort, to explain why the player is where the are and uncover the backstory. Unfortunately I haven't been able to think of anything. In some of the rooms, there are camera's that appear to be recording to give the player the idea their being watched, but they don't actually do anything which i thought was an interesting concept. Any ideas what I should do with the game? [img]https://ibb.co/CV64Kmx[/img] [img]https://ibb.co/0m4Skcc[/img] [img]https://ibb.co/xHLZkyZ[/img] [img]https://ibb.co/KsXh5tf[/img]
Thank you hellthunder for this nice feedback! I did not do much since then on all projects...did a lot of 3d printing again instead working on some replicas
But iam pretty sure ill return to game programming soon. You guys all know, once started u cant leave it behind.
About selling third person stuff...maybe ill release it for free some day around here in the forum.
Still working for month on my fps now. The movement is really smooth yet. This time i made all the script work in a large debug-level with all the stuff in one world, before i went on the "design" part like building levels.
Movement features running, sliding, swimming, wade, diving, climbing, jumping(jump from climb to climb surface), crawling, get shrunk(and climb / jump), hang on moving objects, move with moving objects, push objects. Now iam working for days on the first level. Player starts inside a gigantic monster A lot is done already, enough to fill a project-thread for its own. A lot of funny stuff...like glowing slime that will jump on humans head making them to zombies, like in hl2. The first level, inside the monster, is also a movement tutorial for the player. I love the diving btw.
Still WIP. Some organic objects pulsating. Those tentacles will attack player when close (pushing him away when hit). The intro and first map really feels like a game this time I put a lot of work into small things, like the tentacles smoking, fish-creatures under water, flying "orbs" on camview, smooth head movement etc. Its only around 50ents yet cause i started rebuilding the inside-creature-level from scratch with better models today. The debug-creature-map was a lot bigger already. I merged all my knowledge and (frozen) scripts into this project. This will be my showcase entry and some kind of "techdemo". It uses the ragdoll stuff from my zombie fps, some may remember the video of the explosion and shotgun, pushing the zombie with ragdolls around (2018). Script allows merging of models at runtime for extra detailed maps (thanks again to 3run and TUST). For main-movement AI uses a modified version of SuperKu's v2 pathfinding (most times area-based).
Project runs almost any time in 60fps. when looking at the screens and the scripts the game/story seams to be really cheesy and crazy. Focus is on arcade shooting and a bit about movement itself. After all the frozen projects, died ideas, maybe this will be my final work.
Now iam thinking how to highlight the climb surfaces. Guess ill use lights along the way i want the player to climb. Seams to work nice. In this case players starts in the left area. Must jump and climb the left "block" to the position of the screenshot. Then perform a big jump to the right "block", climbing this one up to the top. Those are the first climb-jumps, so theres water below the jump-area, player cant hurt himself when falling off.
Mine? Really? Will make a video soon to show the progress. Its close to finish, the ai/ragdoll/guns/explosions-stuff. Never got this far before, and it never looked better in motion Zombie Ai is pretty much like the Dead rising zombies. Headshots are supported. Zombie entity uses random head, skin and cloth textures, also random models. Females and Males, animations are shared and GPU animated.
Just working on some more blood spray effects and having fun with the grenade launcher.
Well i could cry. When playing a few minutes...the game randomly crashes. Means the engine just freeze without error message. Wow...great! First i thought its because the nexus gone over 500. But i reduced to 400 to test. Still, game crashes randomly after a few minutes.
Puhhh...iam pretty sure i just found the problem...at least i hope so...edit: still not sure. After around 5 hours of bug-tracking iam almost sure i found one problem...a damaged path. For now i think pathfinding crashed and froze engine because of that broken path set in wed. I made a function. This function spawned zombies in the air, and killed them, then they fall to ground with ragdoll. It run for ever without crashing. Then i realized that it seams only to happen on one place of the map ...so after all this horror tonight, i guess i found the freeze-reason now. edit: YES! Played some minutes...like a charm, no more crash. Have to finish this map now, cause i want you guys to play this ...and publishing works too!
Btw. 500mb nexus max is pretty low for a level, as it seams this is the limit for each map ( you can consume more memory, but for the map 500 seams to be the limit ). To be honest i never understood that nexus-thing.
edit: Another crash removed For now i removed the random texture-usage of the zombies. As it seams iam using ent_cloneskin wrong. After a time engine throw up random errors like wrong wav/ogg file ending in a ugly vbC++ runtime crash message. It felt like a memory leak problem so i removed the ent_cloneskin stuff and the crash was gone...thanks god.
Have to read more about ent_cloneskin. Guess i cant use it with my morphed - entity?
What a horror if random crashes appear only after minutes of playing. I dont understand why this not happened when i tested the zombie-spawning/morphing for minutes, cause of that broken path bug i found before.
But iam happy enough ive found the reason for the "random" crashes this time. I will try to make a work around, if not zombies will only use one texture. Its bad yes, but at least the game is running nice. This time i killed dozens of zombies, with full-cheated weapons, and no more crash...for now.
All one handed weapons can picked up twice. For example double Glock, double Uzi, double Mp5, done with one MDL file, VMASK and vector/script magic. Merged all(!) anims in one container and normed all bone structures of all zombies. This way i saved again a lot of resources to use. I added many of new zombie types. Added gun updates like silencer and stuff like ammo-block for the M79 grenate launcher so it can take more than one bullet at once.
One new NPC is the SWAT zombie. You have to remove his armor first before u can kill him. You better have enough ammo or explosives before messing around with those guys
Working on players safehouse...it was a lot of work to get this asset game ready look and run smooth. But now iam happy (windows will follow soon). I liked Trevors safehouse in GTA a lot so this was my inspiration. The model designer did a great job with groups of the model so i could rebuild it. To get around the 64k polycount limit, remade the textures etc. Now it takes almost nothing in the nexus.
wip Outside
inside
It feels really comfortable Maybe it will be driveable (on start of the current script i implemented a mod of the knights on wheels car physic script maybe to use later...just to have it, like in duke nukem 4ever fex)
Update Almost done...working from time to time on the safehouse since last we. Its crazy how much time every step takes and how much work everything is. You have to like it, if not it wont work All shadow and light stuff now looks correct. Used some tricks for example if player enters safehouse the inside gets visible while the outside becomes invisible. Inside windows close to finish too, was more work then expected. Also the lights are now created / removed, if player is outside he wont see anything from the light inside, like it was before.
Thanks again for your feedback and interest. Really appreciate.
Iam now working since beginning of this year on the project. It was born out of the ashes of all frozen projects. Like thomas edison i say "i did not fail...i know 1000 ways how not to programm a 3d fps" Feel like i must upload now...afraid to have a accident before or something...
This is the first time it looks and feels great and renders fast. Almost all the time 50fps constant. Especially the ai movement gore and gunpplay looks and feels pro.
Finished first mission. A guy comes to your trailer you help him clean his house. Then he becomes a gun seller. Its open world more or less. Helper npc are able to attack enemy npcs dynamicly. Now i have to polish the mission. Then i go on building the map. Cash u get from corpses etc
Since my last post i was working almost every weekend on the fps project. Over a year now. Enough to say for a new stand alone thread. With that much code it becomes more and more a challenge. Meanwhile it became a small arcade zombie shooter
Episode1 will be done "soon" and ready for download (free of course).
Dive / Swim area wip, almost done. iam happy with look and feel. Simply animated transparent water with swimming (pushable) objects, and bubble particles. The electric panel in the background opens a pipes gate in the water area.
Most work left is polish here and there, finishing levels etc. Greets
Played around with Duke Nukem forever and managed to extract and convert models and animations Just for fun. A dn3d "remake" would be nice (even if there was one impressive on youtube already). Pigcop and Assault Trooper looking groovy in acknex (all anims). Just seen i have to mirror them. I read people complaining about the "bad anims" of forever...what i just converted looked pretty cool in motion i have to say (and while i played it, years ago, it looked ok for me too). It was not easy to convert the used model/anim-data into acknex-MDL(2010?)...wonder what Gearbox (if iam correct atm) would say about a free hobby dn3d remake.
Besides i rewrote most of my fps scripts, like player controller. New melee-weapon system.