13 ms·
Having worked at Pixar, I think Catmull's wage-fixing involvement should absolutely disqualify him from receiving the Turing Award. And really, Catmull-Clark s
by tayistay 7y ago
Having worked at Pixar, I think Catmull's wage-fixing involvement should absolutely disqualify him from receiving the Turing Award.
And really, Catmull-Clark subdivision surfaces aren't that amazing. Z-buffering is obvious. Those were the days of low hanging fruit in graphics.
- micmil 7y agoIt's rather interesting that his Wikipedia page no mention of this, nor does the one for Pixar.
- pcwalton 7y ago> And really, Catmull-Clark subdivision surfaces aren't that amazing. And the fact that Catmull-Clark is based on recursive refinement makes it really bad for GPU evaluation (or parallel evaluation in general). It's completely incompatible with hardware tessellation, for example. If I had to pick an especially non-obvious, elegant, and useful 1980s graphics technique, Pineda rasterization would come to mind.
- tayistay 7y agoSubdivision surfaces aren't so bad on the GPU. Since each limit vertex is a linear combination of control vertices, you can precompute the stencils to evaluate limit vertices. Those stencils can all be evaluated in parallel in a compute kernel. Generally, the base mesh topology isn't changing in an animation, so this works fairly well. For Catmull-Clark, the regular parts of the mesh (quads with valence 4 vertices) are just b-splines and thus can be evaluated using the hardware tessellator.