A Quantized-Diffusion Model for Rendering SKIN and Translucent Materials

[Bigguns]

Hi vlado,

Would be REALLY cool if you can do a new skin shader (you can keep the old one though, but adding this one as a true skin shader) based on this paper, bottom I put the link ( and by the way, they will release the papers about the hair shader too)

here is the introduction: We present a new BSSRDF for rendering images of translucent materials. Previous diffusion BSSRDFs are limited by the accuracy of classical diffusion theory. We introduce a modified diffusion theory that is more accurate for highly absorbing materials and near the point of illumination. The new diffusion solution accurately decouples single and multiple scattering.


http://www.eugenedeon.com/qd.html

http://delivery.acm.org/10.1145/1970...18692c780995ab

I ask this, because right now the sss2 is great, but have some flaws,here they are.

1- when we put the sss a bit strong it get greenish in some area, and we CAN'T get out of it without diminishing the sss amount, and this must be tester with a character with TEXTURES on it. the more we put the scatter color red and saturated,the more get the effect.
2- It's not perfectly accurate, as explained and demonstrated in the paper.
3- it's SUPER slow with scatter GI turned on
4- The back scattering is not effective, it only work well when you have only light coming at the back, once you put gi or a direct light in the front, the effect is almost lost at 100%.. So we need to have control over the backscatter, like in Mentalray . I tried to make the sss radius of the ear different from the head, but the ear of course look waxy compared to the rest of the head..

So it's a great achievement compared to the older SSS, and it's better then the one of Mentalray for now ( but Arnold's one seem to be better ), but it's not a reason to not upgrad it or write a new one to the addition of the other. It took so long to Vray to get a decent skin shader.. it was what making me not using Vray for characters, and most peolpes and studio's.

Now it's cool to see things moving in the right direction,with the addition of a new Vray hair material

So I hope you will upgrad the skin shader (or at least upgrade the sss2 in the 4 weak points I listed below) soon, like this year please, don't let Arnold have a better skin shader for the next 2-3 years.. , I (and everybody) want Vray to be on the top of the latest tech availiable

All the best

[vlado]

What makes you think that this method will solve any of the issues you mentioned?

1. Follows from the physical laws governing sss. Since this method claims to be physically accurate, it can be expected to have the same issue.
2. Maybe, but it is still *quite* accurate.
3. This would be still an issue.
4. This follows from the physical laws governing sss. Since this method claims to be physically accurate, it can be expected to have the same issue. (Side note - you really really need the "single scatter" mode set to "raytraced" to get correct back scattering, especially in thin areas like ears).

With that said, there are some further improvements planned for our sss2 shader.

Best regards,
Vlado

[joconnell]

Vlado, is there any handy way to decouple the controls over intensity versus spread in this material? For example I'd often like a non realistic or stylized look on something - take for example a glacier type surface where I'd like to have a really strong effect on the thin sections but almost nothing on the main flat body of the object. At the minute the scale and radius controls are used to attempt to limit the "spread" of the effect or "depth" of the effect into the surface, but bringing up the multiplier of the effect quite correctly also brings up all the other less bright detail that's diffused out into the object. In some ways what'd be great is a second radius control - almost similar to the standard max light attenuation controls where you have a start and an end range. This'd allow you some simple controls so you could very quickly decide how intense the sss is, and how deep a depth the effect fades out over. At the minute trying to layer it with other shaders seems to give quite odd results since it's trying to mix an sss effect with a non sss surface for example. The other approach is non accurate admittedly, but it'd come in very handy if it were possible!

[Bigguns]

Great to hear that there are some further improvements planned for the sss2 shader

But about the greenish effect, in Arnold rendering,they don't get this, even if they put the sss effect very strong, so it means there is a way to avoid it and it render really fast,even with gi.

I already tried to put it from single to raytraced mode, and it gaved me the EXACT same result( in fact, any mode give me the exact same result), don't know if it's working in max? or do I need to tweak some number to make a difference? ..
and if I tried to put it to raytraced refractive, it simply crash max everytime... maybe it's because my sss2 is blended with a vraymat for the reflection? and I have a map in the scale of the sss2, it's weird..

[vlado]

But about the greenish effect, in Arnold rendering,they don't get this, even if they put the sss effect very strong, so it means there is a way to avoid it

<shrug>I have not used Arnold, so I am not in a position to comment. I'm pretty sure a similar effect would still be observable. There might be differences in how you are setting up the shaders and where you are plugging the textures.

and it render really fast,even with gi.

As with everything, this is purely a matter of (gi) settings.

I already tried to put it from single to raytraced mode, and it gaved me the EXACT same result( in fact, any mode give me the exact same result), don't know if it's working in max? or do I need to tweak some number to make a difference? ..
and if I tried to put it to raytraced refractive, it simply crash max everytime... maybe it's because my sss2 is blended with a vraymat for the reflection? and I have a map in the scale of the sss2, it's weird..

I guess it would be best to get me an example.

Best regards,
Vlado

[Bigguns]

Can I send you a scene by email ? if yes, give me your email and I'll send it so you can play with it, and you will see it's crashing if you try it with the last raytraced mode on.

some friends are on the beta of Arnold, and they tried to boost the sss in order to get some greenish effect but it do not occured.

[vlado]

You can email it to vlado@chaosgroup.com

As for the green effect, like I said, it might be down to wrong shader setup. Looking at your scene could help find out if this is the case.

Best regards,
Vlado

[Bigguns]

Ok I sended it

[Dariusz Makowski (Dadal)]

Any chance u could email it to me too dariusz1989@gmail.com im interested.

[Morbid Angel]

the "green" effect is also an issue for us. It is easily reproduced - put a checker into the subsurface color and set the scatter color to H:360, S:1 V:.5 and its visible. It is really visible with any colors which have a high saturation point in scatter color. If its not possible to control that color it becomes just difficult to achieve some of the effects.

[vlado]

Of course it is possible to control this - all you need to do is reduce the green component from your sub-surface color.

Best regards,
Vlado

[Morbid Angel]

but my sub-surface color is a black and white checker, would it make a difference if i reduce the green contribution on it?

[vlado]

Sorry I mean the scatter color, not the sub-surface one.

Best regards,
Vlado

[Morbid Angel]

Thanks for the reply. Here is a screen-shot of what I have. Since the scatter color is pure red, I cannot reduce the green in it. If I connect R > R, B>B, and leave out the G, then it renders yellow.
Any ideas?

[vlado]

Is your "single scatter" mode set to "simple"? I get a greenish tint in that case as well, which goes away if I switch to e.g. "raytraced solid" (see the attachment).

Single scatter set to raytraced:


Single scatter set to simple:


Best regards,
Vlado