lol... Thanks bro.

I've actually been thinking about this. Here's my ideas in "Guitar language". Hopefully they make sense to everyone else...

- set up x amount of render targets (2 - 4)
- switch the camera's render target to the next target every frame, and cycle through after x.
- Send the targets to the shader starting with the most recent to the last target.

This will avoid using bmap_blit because the camera renders directly, so there will be no slowdown in getting the images. I can take them directly from the camera view. There are no problems here...


The problem comes when actually blurring the images. How should it be done? I could easily mix the images at different alpha's, but this would probably be too visible between "slides". This would be the most FPS friendly

[New Idea!]However, I could also store the difference in camera angles and pass those to the shader. This way I'd know which direction to blur and how far. However, mixing all these images and blurring them would KILL FPS, and I don't have the fastest computer.

Let me know your opinions, and I'll see what I can find time for


xXxGuitar511



..And in case your wondering how I found this? I go through all the [unread] posts in the forum about 3x a day... lol. Well, at least the sections that appeal to me.