I tried not the same scenes, but comparable ones in unity, 3dgs, and esenthel, each made due to my concept world.

yes, unity is definitely fast and easy, if you pay $1500 for it! the free is not okay for large scenes as it does not offer lod, dynamic shadows, and shaders for nice water, no model editor. and it "cheats" a bit, it uses static lightmap shadows for vegetation, and offers only billboard vegetation lod, which is sometimes not enough quality. by the way, its programming is really good, and the editor is okay only you have some 3d art experience. but it falls out a bit from the comparison, because of the high price of the pro.

in esenthel, you get for free or for $100 if in a commercial project: fast and really high quality pssm shadows, gpu bone animation, streaming terrain with lod like in UDK for really huge-huge worlds, more advanced model lod (several models and a billboard in the end, can be created in its model editor), grass is also handled fine but not optimized as in unity just clipped (but leaf normals can be set in its model editor which is very important for plants). it offers DX10+ rendering what I found much faster than the DX9c, and you can compile as x64 or x86 in Visual Studio, but no instancing, so a bit slower than unity, but beside higher quality. it requires more programming knowledge because it uses C++ or its own scritp language which is a bit more lightweight but still not as easy as lite-c or unity-script. importing 3d assets is just smooth supporting many formats. the world editor is also fine, you can choose or define new object classes and properties, it features a built in navmesh pathfinder. for open world project it is definitely a good choice.

in case of 3dgs you need the commercial licence for $200 to really get something, the free is okay for toying similarly to unity free. imo its scripting is the easiest, probably the best 3d engine to start with, but its editors are outdated, importing 3d assets can be problematic, you cannot make a nice terrain in MED or WED alone, you can't see anything in real time. my original concept was to make my own game editor to test my game AI on design time, so it was not a drawback for me. its outer lod system is fine, but you get no other optimizations like terrain lod or instncing, only in the pro which is over $800. it is possible to get good performance and acceptable quality if you use pssm shadows (its quality is not the best at all), but it can be a good solution to combine it with static lightmap. it still features vertex animation, what I found good if lod is applied strongly, the cpu bone animation is slow, the gpu is pro only. so this engine is a big compromise, you have to do a lot of scripting to achieve things what you get in other engines by default, but good for a one person hobbyist development.

recently I'm playing with UDK, but cannot say anything yet. it's complex as hell.

That resume well the strenght of engines.
Yep UT3 is big, but have lot of advanced graphic features could it be textures morphing shaders, clouds etc ...
Im' not sure you'll be able to create any game with it easyly ?
I hope to see some new thread about UT3 by you some day
What i don't like is the fees you pay per game once you reach enought money and all life as your game is sold.

That's why i avoid it and because im'm not sure to be able to make some enought big and next gen game needing UT3 !

I have alwayes had some doubts about the supposed graphic quality of game engine
Can people really distinguish a Unity game from, let's say an Esentahl one ?
I dont think so
In my opinion the best engine is the one which allow you to get the same results in the shortest time
Developing time depend of course also on the famniliarity you have with the engine but , aside from that, there are big diffrenece

From my experience
Unity supplies the best editor
Leadwerks the best programming style

Leadwerks will also supplies in next version some betters tools and workflow and new scipt and programming features.

You forget 3D creation time that depends on the workflow , enought tools and enought and graphic features.
Someone wanting to create some big medieval world will need big terrain tools, if he can have cloud tools, full screen filters etc ... already existing it will only big boost in speed, caus he won't have to ask or wait for a programmer to ask for that features.

but don't forget, only compare an apple with an apple, a pear with a pear...

and sometimes it can be seen which engine is used.

i am already using essenthel for my first big game project and i can tell
you its got the goods for rendering and a not to hard workflow but
i am stil on the previouse version and there is certainly a bit of work to
be done for the editors mainly better camera controll .. objects needs handles
to be moved on axises in a better easier way bones aswell and a bunch
of small litle changes ..also it realy needs a good api reference/help manual

lots of example code though...

the integrated c++ code edit is a welcome
i like it and hope essenthel addresses all these small things in
version 2 ..

hi wjbender, good to hear you are working on a game, what kind is it? and can you extend the esenthel recast pathfinder with dynamic obstacle avoidance?

after checking the functionalities I need, I plan to start a project in esenthel, but only in the current free version, after I finish my current editor release (I'm just adding graphic assets, but apparently I caught from my son a vomiting disease that could cause some delay...), and reorganize my pathfinder scripts to be more modular and portable to c++.
try to write some feature requests to esenthel, I think he want to make a much better editor, as I think of his new licensing concept.

Hi sivan , yes i see you have been a busy bee

I see the coder made some changes to the editors mainly for the new version.
And on the subject of pathfinding i am yet to see what he exposed to us coders
for pathfinding alltogether but if you look you will find the 2d edit mode in the path section
of the editor where left and right mouse buttons may specify whether a tile area is marked
as walkable and non walkable although it is said in the manual that
build in pathfinding does not make use of this data set using 2d mode and can only
be used for custom purposes by accesing it inside the game using custom programming
functions....
If this could be accessed and used in code by activating deactivating tiles
as walkable or non walkable then we can make a system
for tile shaped dynamic avoidance like buildings or obstacles
in a rts game i believe..

My game is a combination of rpg and shooter taking place on earth/or some planet and heaven(a scientificly advanced secret planet) so quests will be devided between these world/locations
fighting for both sides ... Mainly my first level wil be desert based.... But enough on that

essenthel surely has a bit of a learning curve , i am also working on the free version
for now , whether i stay with essenthel or head to gamestudio for this project
will depend on essenthels developement so at this stage i am giving it a shot
if it doesnt meet my needs i shall jump to gamestudio perhaps..

sounds interesting! it is hard to decide which engine is better for a certain project, the best to make some tests in both, which is a bit time consuming in the beginning, but will be much faster to complete the final game.
now I approximately know 3dgs limits, I'll have to test esetnhel with my pathfinder by writing custom classes for it (probably independent of the terrain areas because of their streaming in/out).
I found that 2d system, unfortunately it is limited, and was used only in the past, after navmesh it was abandoned (and I need to handle a few building levels not only one).

Mm i see yes , your pathfinders code
to handle streaming would be very cool.

I think full source is needed to have a look at the navmeshing
but okay i believe more detour/recast features could be integrated when
he decides to update the engine itself i read that on the forum some features
were planned on the roadmap just not sure if its true..

I also noticed the non navmeshed pathfinder may work from a image file
where colored pixels can represent non walkable areas
maby this would be more usefull by image minipulation and building
a array of images of your map by pixel representation..