Page 1 of 1
Subsurface scattering changes between path / bidir engines
Posted: Wed Sep 02, 2020 6:26 pm
by universalsd
I have created a demo scene where the SSS changes drastically just by switching from bidirectional to path tracing. I have tried various scales for the object and it was all the same. Obviously the path traced one looks correct. What am I doing wrong with bidirectional? I have attached my scene
Edit:This is the correct comparison, with bidirectional on the left and path tracing on the right
Re: Subsurface scattering changes between path / bidir engines
Posted: Wed Sep 02, 2020 6:44 pm
by Sharlybg
From my POV it is a normal behaviour. Bidir is a more advanced engine capable to bring more light data to the final image compared to simple path. It is obvious for complicated simulation.
Re: Subsurface scattering changes between path / bidir engines
Posted: Wed Sep 02, 2020 7:51 pm
by universalsd
unfortunately i had an error in my post, in the first image the labels are reversed. Here is how they correctly are. You can see a problem with bidirectional across the board. Subsurface scattering has a "darkening" effect
Re: Subsurface scattering changes between path / bidir engines
Posted: Sat Sep 05, 2020 2:09 pm
by Dade
I may know where the problem is, I'm checking.
Re: Subsurface scattering changes between path / bidir engines
Posted: Sat Sep 05, 2020 3:21 pm
by Dade
Blender Suzanne is an "open" object (there are holes and overlaps) so it can really mess volume computations. I replace the Suzanne with a cube a done some test.
This is path:
This is BiDir:
The light tracing contribution in hybrid rendering (something plain path tracing doesn't render at all) is very significative:
so I'm not sure if you are just observing a normal difference in what path tracing can not render or there is a real problem (aside from the Blender Suzanne problem).