More subsurface scattering

Started by Hannes, October 17, 2019, 04:45:20 AM

Previous topic - Next topic

Hannes

Playing with the new sss feature. Here I will post some of my tests.

Kevin F


Hannes

#2
Another iceberg. It's basically a gengon object I heavily displaced in 3ds Max.
I had to apply a simple non SSS shader to the part that's below the water, since the SSS shader looked like crystal or glass below the surface.
I added a little chromatic aberration to the underwater part of the image to make it look more... yes, underwater.

Dune

Wonderful. So what were your glass settings here? Several meters decay I presume? Any pf for color variation in the sss?


Hannes

Thanks, Ulco and Doc!
Volume density 250, decay distance 50.

Quote from: Dune on October 18, 2019, 05:13:00 AMAny pf for color variation in the sss?
Nope.

Dune

Thanks, Hannes. PF as color input is nice to try too.

Hannes


KyL




Hannes

Thanks to KyL for his tipps (and Matt of course).

Just a glass shader for the skin. But as a child layer of a simple surface layer for displacement (color unchecked). The texture went into the volume density function slot of the glass shader.

Dune

I tried that too (texture into glass' density shader), but that made it really dark. Probably also depends on the color input color.

Hannes

Yes, you have to find the right balance. In this case the texture alone was rather pinkish, so I made the color of the density, additionally to the image map, slightly yellowish, which resulted in a more natural look. It's a bit tricky, but it can be done.
I like the fact, that, at least in my opinion, there's no need to merge the glass shader with another one. As I already wrote, if you use it as child of a surface layer that doesn't have any color, but can be used for displacement for example, you can have the pure benefit of the glass/SSS shader.

Dune

Yes, the child thing is what I've been doing too (amongst more), but it's indeed fickle to find the right balance. Merge is probably heavier, so that's good. Some of the image map and procedural blemiges do come through, so it seems. And bump is a nogo, though maybe some more experiments will find that to work too.