I'm wondering if anyone has some advice on a good workflow for creating a point cloud in SOPs for use with SSS. I have pretty extreme geometry that is deforming and has changing topology each frame. I need to crank up the multi-subsurface quality in the shader to a very high number to avoid artifacts, and it doesnt seem very efficient, (its about 5 million SSS points when I visualize them with a file sop).
What I believe is happening is that scattered points for the SSS point cloud are jittering each frame since the topology is changing, just as they would if I used a scatter sop. For this reason I want to create my point cloud in SOPs so I can have more control.
I have created a sort of rest position by using a timeshift on the deforming polygons and scattering points on the final frame, then culling them throughout the animation. that solved the jittering issue. I have also given what I see are the necessary attributes “S” and “O”, both vectors.
I've generated these attribs by using an xyz distance vop to extract the primitive UV's, the “S” attribute. And the prim number that the points are scattered on, the “O” attribute. One thing I dont fully understand is how the second value on the O vector is being calculated. The first value appears to just be the prim number that the points are scattered on.
Finally I have a shader that executes a pcwrite to write out the point cloud file. Alas when I plug this SOP-made point cloud into my SSS shader, I get an error saying invalid point cloud.
I hope I've been clear enough, I'm unable to post the file. If anyone has any advice for a workflow for generating an SSS point cloud in SOPs, I would love hear it. Thanks
SSS pc generation workflow
3161 4 3- Loudsubs
- Member
- 37 posts
- Joined: Oct. 2009
- Offline
- garretthoyos
- Member
- 43 posts
- Joined: Aug. 2012
- Offline
- Loudsubs
- Member
- 37 posts
- Joined: Oct. 2009
- Offline
- garretthoyos
- Member
- 43 posts
- Joined: Aug. 2012
- Offline
- sl0throp
- Member
- 258 posts
- Joined:
- Offline
During the shader development boot camp over at side fx Corine from Method went over all subsurface scattering methods.
You do not need a per frame point cloud for sss even if the geometry is deforming, this has to do with how Mantra uses the data. I know this seems counter intuitive but it is a single bake solution. There are shaders that she created that were distributed…not sure where they are.
You do not need a per frame point cloud for sss even if the geometry is deforming, this has to do with how Mantra uses the data. I know this seems counter intuitive but it is a single bake solution. There are shaders that she created that were distributed…not sure where they are.
-
- Quick Links