Deep Shadows

I’ve read several places about Pixar’s Deep Shadow technology. It sound basically like a shadow buffer, but with many benefits related to aliasing, volumetrics, blurring, motion blurring and small geometry like particles and hair. There’s some more info here:

http://features.cgsociety.org/story_custom.php?story_id=96
http://www.animationmagazine.net/article.php?article_id=526

We’ll I sort of understand WHAT it is. But I haven’t been able to find any technical documentation about HOW it works. Perhaps its too closely guarded and secret for any outsiders to know. If anyone does know more about it, or has experience using it, please share.

A more robust shadow algorithm (without raytracing) would be a great addition to the standard render engine.

have you been there ?
http://graphics.pixar.com/
and more particularly this paper ?
http://graphics.pixar.com/DeepShadows/paper.pdf

I still havn’t read this so I can’t say you if it’s what you search or not…

This is sort of embarrassing. I’ve been looking all over for something like this, and now that I’ve posted saying I can’t find it, I found it.

http://graphics.stanford.edu/papers/deepshadows/

So it looks pretty impressive, with no real trade offs - nothing lost, only gains. And being as there are some algorithm descriptions in the paper, maybe this could be integrated into Blender’s renderer?

the only trade offs are they are more memory expensive and need a little longer to create the maps.

Bob