Chaos Corona Forum
Chaos Corona for 3ds Max => [Max] Bug Reporting => [Max] Resolved Bugs => Topic started by: maru on 2014-03-14, 15:44:42
-
This probably isn't a bug but I'd like to know why this happens and if there is any way to get rid of it.
Again, my cave scene. I know this is a nightmarish gi situation but we had to stress test the new build, right? :) The cave object is single-sided thanks to the latest bugfix. I am placing a portal in the hole. It will never be exactly capping the geometry because I am using displacement. Here are my experiments.
(on the last pic, environment is disabled, only light mtl is used)
-
Can you upload me the exact scene?
-
couldn't find the uploader thingy, hope it works, quickly cleaned, I have to go now :D
-
Those are self-intersections/random GI noise, not related to the portals.
-
Ok, but I thought portals should fix/help in this case. These fireflies don't go away.
-
portals can help with overall amount of noise, but not with outliers created by some other way (like selfintersections in this case). Plus the GI is probably playing role, maybe decreasing MSI would help.
-
Actually, it was just self-intersections. The random GI noise was caused by them as well, since they were basically a tiny windows all around the mesh.
I did following things to fix the scene:
1, Created large box around the cave, and cut the opening using boolean. - This got rid of selfintersections
2, I placed the portal outside of the opening. Portals in Corona are not lights like in other renderers, they are actually just targets for sampling to be focused into. When you placed the portal inside, that meant not only Enviro, but also the end of the cave opening would be sampled additionally, so sending too many sample to the edge of the opening caused rest of the noise. Portals should either exactly fit the opening, or be slightly outside. Never inside.
Here's the fixed scene... should render fairly quickly :) I hope you use Max 2014. :D
BTW, since this is scene with quite complex direct lighting (strong light, small opening), you have done exact opposite in regards to sampling than what you should have done. In these situations, it often pays off to increase sampling of GI, as well as light samples... yet you have done the exact opposite :D
-
When you placed the portal inside, that meant not only Enviro, but also the end of the cave opening would be sampled additionally, so sending too many sample to the edge of the opening caused rest of the noise. Portals should either exactly fit the opening, or be slightly outside. Never inside.
This makes sense.
BTW, since this is scene with quite complex direct lighting (strong light, small opening), you have done exact opposite in regards to sampling than what you should have done. In these situations, it often pays off to increase sampling of GI, as well as light samples... yet you have done the exact opposite :D
It's just a quick scene to test RC3. I know I should have used more GI samples, but again thanks for clarification.