Originally posted by Bobboau
bout a second, probly, seeing as the card is able to do it on the fly like that sudgests it isn't a realy expensive opperation
You could do it in software which would mean locking buffers all the time which wouldnt be good.
Or try and do it in the card using vertex buffers or something. I think this would be difficult and would only benifit cards that support whatever vshader version you use. Also it means if anyone whats to use a vertex shader for anything else the shader would be really complex.
Im not shader expert sadly but I hear that one of the problems with DX8 shaders is that they have to be fairly small.
Even if we updated to DX9 which I think would be a bad idea at this point (anything that might make the engine slower should be fully researched first) that would limit the effect to DX9 cards.
To me it sounds like a hell of a lot of hasstle trying to do it in card hardware. If ATI could have implemented it as a shader rather than a card 'feature' then they would have done that and it would also be possible on GF cards.
I cant remember where I heard this but I think its a dead feature that never really got picked up by the games industry. I hope I wrong because its obviously a brilliant idea.
I did lots of stuff on patches at uni, I might have a stab at it at some point.