skin shader a la Alsurface pleaseeee

[cpnichols]

Keep in mind that the Emily data has extremely high res displacement, so we get the micro facet from the actual geometry. The map is 16k and down to 10 microns. That is the whole point of why we used phong. Instead of faking it with a BRDF, we get the real microfacet as well as it's anisotropy.

We wrote a post about it here: http://www.wikihuman.org/index.php/s...kin-roughness/

For those that don't have that, they will need some sort of microfacet simulating BRDF, such as Cook-Torrance, GGX, Beckman, GTR, etc...

The Single Scatter Map is a bit complicated and still a matter of debate amoug Digital Human League members as to its best use. The best "layman" way to describe it is the first layer of skin that does not actually get penetrated.

Lastly, yes, we have some level of hair, but it is not exactly correct, so we didn't include it with the data.

[Paul Oblomov]

Now, when u have vrscans - just slice a bit of Emily

[sharktacos]

Keep in mind that the Emily data has extremely high res displacement, so we get the micro facet from the actual geometry. The map is 16k and down to 10 microns. That is the whole point of why we used phong. Instead of faking it with a BRDF, we get the real microfacet as well as it's anisotropy.

We wrote a post about it here: http://www.wikihuman.org/index.php/s...kin-roughness/

For those that don't have that, they will need some sort of microfacet simulating BRDF, such as Cook-Torrance, GGX, Beckman, GTR, etc...

Yes, I found the idea of differentiating between meso and micro tremendously helpful. For me the question then becomes how to use that understanding in a context where we do not have access to scan data and are needing to create things from scratch. How can we take the observations from this, and find a way to apply them in a practical workflow, say for creating a fantasy creature... or even a cartoon character? In other words, while the goal of the Digital Human league was to present artist-free observations about how best to emulate humans, the question I ask myself as an educator is how to translate those observations into tools for artists.

In that context, I still think it's really helpful to create those meso and micro level displacements. Even with a GGX or even with two lobe spec that micro map really makes a difference. I've also grown to appreciate the math behind the bump in Vray, which with the delta allows me to get incredibly fine detail that in other renderers I could only get with displacement. What I was not able to generate with that approach (creating a micro-level displacement map by hand) was both roughness and anisotropy. For that I continue to use traditional approaches.

Speaking of which, you know what could be quite useful is if you guys could generate maps, based on the meso/micro details that map out what the actual roughness values look like spread across the face. Same for mapping out the anisotropic direction. I have an idea as an artist what those both would look like (T-zone and all), but it would be helpful to see maps actually built from scan data.

It would also be interesting to hear the debate around single scatter that you mention. I often learn just as much from listening to multiple views as I do from hearing one finished and agreed upon perspective! Overall, let me say tho that the work you guys have done, both in sharing your thought processes, and in sharing your data is greatly appreciated and insightful. Let's hope its a long conversation!

[Bigguns]

Also Vlado, can you add the translucency and backlight option in our Alsurface please ?

[vlado]

Can you tell me why do you need them?

Best regards,
Vlado

[sharktacos]

Vlado, it would also be great if in addition to diffusion and directional sss, you also had empirical sss. My understanding is that empirical is better than cubic, but faster than both diffuse and directional.

Why? The other sss modes are better quality, but often it is desirable to have fast sss (say if one is not doing a photoreal human but a cartoon character). In that case the other things the ALmtl does are still desirable... like double-lobe ggx spec with microfacet Fresnel.

[sharktacos]

Vlado, I'm noticing that the ALsurface shader does not show up in any render passes/render elements (using Maya 2015). Has that not been implemented yet?

[vlado]

Vlado, I'm noticing that the ALsurface shader does not show up in any render passes/render elements (using Maya 2015). Has that not been implemented yet?

It's on my "to do" list.

Best regards,
Vlado

[bigbossfr]

It's on my "to do" list.

Best regards,
Vlado

I want to use this shader in production. When do you think we can have render elements please ?

[vlado]

They are almost done, I just need to test the code and push it to Github...

Best regards,
Vlado

[bigbossfr]

Oh, it was fast

[JeffG2]

Vlado,

I see a vrayalmtl for Max 14, 16 & 17.

What happened to 15?

[jstrob]

Can you tell me why do you need them?

Best regards,
Vlado

found that on Alsurface doc: BACKLIGHT
The backlight layer represents light that is diffusely transmitted from the reverse side of the object and it useful as a cheap subsurface effect for thin ojects such as cloth, paper and leaves. It sits underneath both specular layers and so its intensity is scaled according to the Fresnel transmission of both the specular lobes.

and about translucency I wonder if, Bigguns, you want to have something to be able to see hair root through the skin?

[vlado]

I know what it does - the result is the same as the translucency effect of the VRay2SidedMtl material. I was asking why Bigguns needed it...

Best regards,
Vlado

[vlado]

What happened to 15?

It's the same plugin as for Max 14.

Best regards,
Vlado