Page 1 of 1
question about caustics.
Posted: Mon May 22, 2017 8:47 am
by choo-chee
no matter what I try, interiors with light hitting a shiny surface (like, sun hitting a glass table) make terrible caustics - tons of noise on ceiling. I tried disabling all the indirect flags, yet still I get it. what can be done other then render it for 100's of hours and hope for SL that is non-reachable ?
Re: question about caustics.
Posted: Mon May 22, 2017 2:07 pm
by luis.hijarrubia
What engine are you using?
Re: question about caustics.
Posted: Mon May 22, 2017 2:33 pm
by choo-chee
maxwell 3 (latest build, not beta or something), production quality.
Re: question about caustics.
Posted: Mon May 22, 2017 3:54 pm
by luis.hijarrubia
Have you tried setting reflection and refraction caustics to none under Illumination & caustics section?
Re: question about caustics.
Posted: Mon May 22, 2017 5:00 pm
by Mihai
I think it's rather direct refl/refr caustics you need to disable, not indirect.
Re: question about caustics.
Posted: Tue May 23, 2017 8:03 am
by choo-chee
well that's my render settings and still tons of caustics noise....

Re: question about caustics.
Posted: Tue May 23, 2017 1:19 pm
by Mihai
I remember when those settings were added, there was a mention of exactly what created "direct reflected caustics" vs just simply indirect lighting. I'm not completely sure anymore but I think it had to do with the roughness setting most of all. The rougher the material, the more indirect lighting it created and less of it was considered caustics. Would be interesting to make some simple tests of this....
Re: question about caustics.
Posted: Wed May 24, 2017 4:36 pm
by choo-chee
well I've tried to render with and without and no difference...
Re: question about caustics.
Posted: Thu May 25, 2017 10:06 am
by PA3K
Mihai is right,
there was discussion about it in the past. As I remember (not sure), there was something between 95-98% roughness border. Less was considered as caustic and more was just global illumination (diffuse colour bleeding).
Pa3k