SSS-/Mantra Shader questions

[FromHell]

Hello guys,

I have some question about the SSS shader.They are some things confusing me, for example:

1.where are all the epidermal and sub dermal layers?

2.How can i adjust the SSS intensity depending on the model scale?

Lets say my model is a sphere which is 10 units wide, so i need to enter a value of 10 in the SSS intensity to make the light go all the way through it?

I have watched through the "SSS Shader in H11" topic and obviously, it is possible to achieve good results with mantra, however the houdini help on this subject doesn't give enough information.

Edited January 30, 2013 by FromHell

[jim c]

I'm also curious about this. I've got a human model built to scale (about 1.7m tall) and I'd like to use the SSS shader on it. Based on advice from other sources, I've created three sets of texture maps, a diff map for the basic skin color, and then 2 other maps, one a much darker reddish one, I guess this would be the sub dermal map (I think the book* also referred to it as a back scatter layer?), and another lighter map for the epidermal (I think that;'s what Scott referred to it as).

I tried plugging the epiderm map into the Subsurface map, and the subderm into the Attenuation map. Is that correct?

I notice that when I do this, the Attenuation density needs to be pretty low to prevent a sort of glowing effect from happening (presumably the darker colored layers coming through?)? Does this mean that this value represent some sort of depth?

*ZBrush Digital Sculpting: Human Anatomy by Scott Spencer

[jim c]

In case it helps, here's what I've got for the character.

Backscatter:

sub dermal

base skin:

Any ideas on what the best approach is to using this with the SSS params? Or is there some other technique that's recommended?

[danylyon]

You can use more then one Physical SSS node and ADD them together.

If you use the same color as Diffuse and Attenuation, it will only show the scatter effect (no diffuse).

So you can make two or more layers of SSS. Just set them to a different Attenuation Density (Attenuation Density is basically Radius ^-1)

You might want to lower the overall intensity and add your own diffuse (and specular) after that.

hope this helps. Let me know if you need a hip.

[FromHell]

Hello Dany, thanks a lot for your reply, i would love to have a look at a hip file, it certainly would help.

[danylyon]

Hi

Ok, I had a go.

Beware though.. it's quite tricky to control.. and unfortunatly it doesn't look to good with indirect diffuse light. (SSS doesn't seem to pick that up). Also I only used single scattering.. I didn't quite find much of an added quality with multiple scattering. (Besides making it a lot slower)

If you add textures, be sure to make them as red and as dark as the default colors I set up. The actual skin color should almsot be white. (in this setup really)

Skin is very uneven.. start with quite a bit of bump.

I didn't set it up for a particular scale.. (I think it was way to slow).. just multiply the radiuses on the shader.

Let me know if you have any question..

No the model is not included :-P

skin_shader.hip

[FromHell]

Thanks for sharing.Could you please explain how do you know that Attenuation Density is radius ^-1 ?

[danylyon]

trial and error.. hehe :-)

[3dome]

sry for reviving this thread but @danylyon your example scene renders completely white for me. using H15. Could you tell me what I'm missing?

 

edit: nvm, got it. guess you plugged the normalized N accidentally into the spec/diff Cd instead of nN

Edited June 22, 2016 by 3dome
got it

[Cybernetix]

Hey @3dome same problem here in 15.5 but i everything looks wired correctly to me.. normalized N is going to nN..

can u take a screen shot or post fixed file?

 

thanks in advance

[3dome]

I'm at work right now so i cant upload anything. but the fix is quite simple.

The file is quite old, so first thing is to select the diffuse and spec node and change it to the latest version (2.0 for diffuse and 3.0 for spec). It then automatically connects nN in the wrong inputs, you have to fix that then.

Then, after the add2 (add spec and diff) put a conserve energy. Put down a layerpack and wire the _f of conserve energy into F. After layerpack, put a compute lighting and wire its 3 outputs into the corresponding inputs on surface_output.

Plug the add1 (the 2 SSS) in the Ce of layerpack.

(You also might want to divide the result of add1 by 2. That way the values add up to a max of 1.)

[Cybernetix]

Thanks very much!