wraggster
January 5th, 2006, 19:03
bluescrn (http://forum.gbadev.org/viewtopic.php?t=7983) posted this:
"The early version was discussed here (it's based on the display capture/dual pass demos discussed there):
http://forum.gbadev.org/viewtopic.php?t=7958
But as I've just uploaded an updated demo - which looks nicer and now runs at 60fps (although it's still C, and there's a few optimizations left) - I thought it deserved a thread of its own...
The demo (with source) is here: www.bluescrn.net/dsbloom.zip (It's still based on those previous examples, and using the ship model/rendering code from them)
And my hopeless attempt to take a screenshot: www.bluescrn.net/dsbloom.jpg
Anyone else got interesting ideas for postprocessing a capture of the 3D? - already suggested was antialiasing - you could do supersampled 2xAA at 30fps by rendering a buffer twice the size of the screen in 2 passes, then resampling it down with the CPU.
I suppose the capture system could be used for very constrained render-to-texture effects (shadows/reflection maps?) But in practice it would probably be better to have a software renderer for those things, to avoid the huge VRAM/framerate costs that could be involved?"
"The early version was discussed here (it's based on the display capture/dual pass demos discussed there):
http://forum.gbadev.org/viewtopic.php?t=7958
But as I've just uploaded an updated demo - which looks nicer and now runs at 60fps (although it's still C, and there's a few optimizations left) - I thought it deserved a thread of its own...
The demo (with source) is here: www.bluescrn.net/dsbloom.zip (It's still based on those previous examples, and using the ship model/rendering code from them)
And my hopeless attempt to take a screenshot: www.bluescrn.net/dsbloom.jpg
Anyone else got interesting ideas for postprocessing a capture of the 3D? - already suggested was antialiasing - you could do supersampled 2xAA at 30fps by rendering a buffer twice the size of the screen in 2 passes, then resampling it down with the CPU.
I suppose the capture system could be used for very constrained render-to-texture effects (shadows/reflection maps?) But in practice it would probably be better to have a software renderer for those things, to avoid the huge VRAM/framerate costs that could be involved?"