Hey guys,
I am rendering this animation, using IR and LC. My shaders, lights and IR+LC have enough samples, the adaptive is 1-8 with 0.01 threshold...
I am lighting this mainly with and blurred low-res HDR in the dome light for the diffuse, and the high res non-blurried version in another dome light just for spec and reflections.
Every other pass seams very good and clean, but I get this weird acne in some parts of my scene!
If I switch the primary engine from IR to BF this acne disappear, and some fine grain noise is introduced every where, of course the main issues is higher render time then the past setup...
The other way I alleviate more this acne is lowering the adaptive dmc threshold from 0.01 to 0.05 of course at a cost of higher render time.
Another way that I find is not use the dome light with the blurried low res image. By using the original HDR to contribute for the diffuse as well, I get rid of this acne, but my raw GI pass got very splotch and not acceptable!
I know that I can alleviate this noise with denoise in Nuke, but I thought would be nice to illustrate this case here, so maybe the math code responsible to cause this could be investigated and improved in later versions.
Best regards,
Antonio Neto.
I am rendering this animation, using IR and LC. My shaders, lights and IR+LC have enough samples, the adaptive is 1-8 with 0.01 threshold...
I am lighting this mainly with and blurred low-res HDR in the dome light for the diffuse, and the high res non-blurried version in another dome light just for spec and reflections.
Every other pass seams very good and clean, but I get this weird acne in some parts of my scene!
If I switch the primary engine from IR to BF this acne disappear, and some fine grain noise is introduced every where, of course the main issues is higher render time then the past setup...
The other way I alleviate more this acne is lowering the adaptive dmc threshold from 0.01 to 0.05 of course at a cost of higher render time.
Another way that I find is not use the dome light with the blurried low res image. By using the original HDR to contribute for the diffuse as well, I get rid of this acne, but my raw GI pass got very splotch and not acceptable!
I know that I can alleviate this noise with denoise in Nuke, but I thought would be nice to illustrate this case here, so maybe the math code responsible to cause this could be investigated and improved in later versions.
Best regards,
Antonio Neto.
Comment