being able to customize the depth buffer calculations would be one of the most important features in my opinion

imgine you could do depth buffer calculations for several lights at once and could store them in several buffers /texture units

thus i could create pointlight shadowmaps into 6 directions on the flow without passing the same geometry multiple times

these tessellation shaders sound pretty expensive for more

imagine storing a triangle in the gfx memory
so you want to tessellate it how would you do it?

splitting into several smaller triangles should be nearly impossible to do

the only thing i could thing about is applying a depth map onto the surface

if you look at it from the front it won t matter but if you look at it from the side you had to do a per pixel depthtest where you get the triangle plane position (+- the depthmap * scale)*normal
that find out which depthbuffer value to test against

+ you need to do normal recalculations to send the proper per pixel normals to the pixelshader + other things that need to be solved

would be a nice feature but do we really need it already??

shouldn t we unify modelformats and editing tools output at first before we go on with further pushing towards reality?

didn t microsoft and a few companies plane to release a standard modelling format last year?

Quote:Original post by Sages
We have tesselation code now, the issue is that it isn't supported in hardware. With hardware tesselators you'd be able to take a 1,200 polygon character model and render it with a displacement map. The hardware would then tesselate the model and mold it based on the displacement map to create a perfectly smooth mesh. However, since most games aren't geometry limited these days, I don't see consumer level hardware tesselators in the near future. *sigh*

Is this an offline rendering technique now? So the model would essentially be tesselated to the point that there is one vertex for every pixel on the display? I suppose it would be more likely an estimated tesselation level since the screen space position isn't known before the vertex shader. Sounds interesting. Does anyone have any references for information on this topic?

Quote:Original post by EvilDecl81
FYI: The Geometry Shader is not a tessellator, using it in such a way would be... abusive.

I was under the impression that 'geometry shader' was what they were calling the programmable tesselation stage. Maybe I'm wrong about that - what do you think 'geometry shader' refers to?