I'm colored curious about the Physical Surface Root with some discussions lately, and I'm working on a character.

Now I want to try it.

How would one go about creating a skin shader, with SSS, having a diffuse map, a subdermal map and a Scatter Distance/Depth/Size/Whatever map with a Physical Surface root?

Not completely sure about the question, but open the PhysicalSurface root node:

In the lower part you will find the SSS connections, in this case called ScatterDistR_mm for red, ScatterDistG_mm for green, ScatterDistB_mm for blue. The SSS 'Scatter is calculated in mm, and you can hang your red, green and blue SSS maps on them. Just like FireFly, the lowest option Scatter_Group is to separate the SSS with conflicting SSS groups. => Remember the dark blue area's when you got conflicting SSS groups in FireFly?

Here a simple skin setup where everything is build from the Diffuse texture. For SSS add the SSS maps on the Scatter RGB connections.

This is what you see in the renders in the other tread.

Remember to use very high render settings to get "believable" SSS renders because this adds lots of noise.

Best regards, Tony

But then how would I plug an SSS color map and an SSS depth map?

As in this (by Saraubh Jethani):

How would I apply that Albedo + Scatter + Translucency maps together? How to tell Poser that I want this map for superficial skin color, this other map for the color that comes "from the inside", and that other map to modulate what bits I want to be more translucent and what bits should be less translucent?

And if there's so much noise, would I do better just using the regular Poser root with a CustomScatter node instead?

Normal AO calculations are already included in the SuperFly render engine so the only way to get more detail there is to "merge", (read that as enhance the detail) of the Diffuse texture. I would combine Cavity and Albeto map in a paint app.

Or? You could leave the Albeto alone, and use the Cavity map to control the Specular and Bump.

The Roughness map I see? Open the backdoor and give it flying lessons. LOL. (Might serve if paint app merged with the Cavity map?)

For Translucency and Scatter, that's going to take some experimenting. Many apps and content creators use different workflows here. You (almost) never know what you are gonna get. I am out of my comfort zone here.

I see 2 Normal maps. What can I say? Chose one of them. => Remember => ! ! ! ! WARNING! ! ! ! SuperFly uses the normal map, but FireFly flips the green channel of a Normal map compared to other apps.

My setup would look like

• Albeto map in Diffuse

-Cavity map to control Spec and Bump => Ply with the settings => Small changes can go a long way.

• Normal map in Normal Map

• SSS = experiment

On your remark about PoserRoot CustomScatter? => Less noise, but noise, and also requires higher render settings. There is no "cheap" trick for SSS.

Oh yeah, no, I just meant to use Albedo, Scatter and Translucency, I only used that picture cause that's the best one I found with those types of maps that I wanted.

I intend to use a specular map to - but hey I can easily figure out how to use those, LOL.

I think I'm just gonna go for Poser Root with a CustomScatter instead, and leave my Physical Surface experimentations for another kind of shader then.

Thank you!

Throwing away the roughness map when using a node, like the PhysicalSurface node, that is made for a metal-roughness workflow… and using a specular map instead even though the metal-roughness workflow does not even has a definition for it… good luck.

For a complex skin shader IMHO you better go with a custom shader. And even than its still not easy to come up with something you have in mind. Thats mostly due to the limitations of the Cycles SSS node. Personally i gave up on creating an advanced skin shader. i came to the conclusion that i would have to write a custom shader node for it which is not possible in Poser. Please proof me wrong on this :-)

I have been pursuing creating PBR SSS skin shader for Poser quite recently using cycles as my root. Relying on Blender as a guide, I came across a couple of youtube tutorials which may shed some light on how to assemble such nodes. I hope that the links may help.

https://www.youtube.com/watch?v=M_kl5UgPIhE&t=1796s

https://www.youtube.com/watch?v=48Ag0zLL9zA

Unfortunately, there is some node work which is impossible in the current version of Poser. It would be nice to see improvements in those areas. My understanding of noise using SSS. that should clear up if you use higher render samples which will definitely increase render time. No way around it ATM. As for roughness, that is not exclusive to metals. It is applicable to dielectric material and skin as well. I wouldn't shy away from testing it with character textures.

Well, I do enjoy some tinkering, so I'll give it a try - but the intention isn't to use the Cycles SSS node, but the Poser root with the regular CustomScatter node.

I find that Superfly renders the regular Scatter node pretty well.

Though mind you, I'm NOT a mathematician or physicist - I don't know the specifics of what I'm looking at. I just know what looks good to me when I look at it. So it's highly likely that I won't come up with a proper material, but I'm satisfied with a pretty material

hborre posted at 11:34AM Fri, 28 June 2019 - #4355142

I have been pursuing creating PBR SSS skin shader for Poser quite recently using cycles as my root. Relying on Blender as a guide, I came across a couple of youtube tutorials which may shed some light on how to assemble such nodes. I hope that the links may help.

https://www.youtube.com/watch?v=M_kl5UgPIhE&t=1796s

https://www.youtube.com/watch?v=48Ag0zLL9zA

Unfortunately, there is some node work which is impossible in the current version of Poser. It would be nice to see improvements in those areas. My understanding of noise using SSS. that should clear up if you use higher render samples which will definitely increase render time. No way around it ATM. As for roughness, that is not exclusive to metals. It is applicable to dielectric material and skin as well. I wouldn't shy away from testing it with character textures.

Thank you for these! I've been adapting Blender Cycles stuff for a while too.. But it was brought to my attention that some people won't use anything with a Cycles Root because it renders too slowly. It's the opposite for me, but what can I do, such is the life of a vendor, must work around things to make sure the customers are pleased

Honestly I haven't noticed what the roughness texture even does. It feels to me like just bump.

The custom scatter has one advantage. It does produce soft shadows on the skin (not sure if all of them do) whereas the cycles SSS gives you hard shadows like a hard surface does.

Tja, in the end when all is said and done, each rendered pixel gets a single RGB value. It does not care where that RGB value came from. So a lot of testing to get to the results desired.

Nagra_00_ posted at 12:23PM Fri, 28 June 2019 - #4355159

The custom scatter has one advantage. It does produce soft shadows on the skin (not sure if all of them do) whereas the cycles SSS gives you hard shadows like a hard surface does.

The Scatter skin node doesn't let you pick a scatter color tho - I wouldn't be able to use my subdermal maps.

Afrodite-Ohki posted at 6:30PM Fri, 28 June 2019 - #4355161

The Scatter skin node doesn't let you pick a scatter color tho - I wouldn't be able to use my subdermal maps.

Yep, and none of the SSS nodes gives you control over the amount of back scattering.

Nagra_00_ posted at 12:46PM Fri, 28 June 2019 - #4355162

Afrodite-Ohki posted at 6:30PM Fri, 28 June 2019 - #4355161

The Scatter skin node doesn't let you pick a scatter color tho - I wouldn't be able to use my subdermal maps.

Yep, and none of the SSS nodes gives you control over the amount of back scattering.

I can modulate the amount of translucency though by plugging a map into the Mean_Free_Path in CustomScatter, no?

Afrodite-Ohki posted at 7:01PM Fri, 28 June 2019 - #4355168

I can modulate the amount of translucency though by plugging a map into the Mean_Free_Path in CustomScatter, no?

Doing so mainly affects the forward scattering (scatter depth) and has little to no effect on back scattering. Its been quite a while since i played with it, so please don't nail me ;-)

I won't even respond to all the previous posts. Suffice to say they're wrong. I will give a tutorial on how to approach this. It will cover a lot more than the directly asked question. This is happening in real time so will take some time between posts to generate demo renders.

If you want to follow along with me exactly, I will use James - we all have him. Or - use whatever figure you want. If you're using a figure that has separate lips, nostrils, and so on you will suffer. I suggest either altering this head so it is all one material zone, or get ready to waste a lot of time.

Set the Construct to all black on diffuse and specular. We want no bounced light. Load your figure. Set all of its material zones to black on diffuse and specular. We want no bounced light.

Bounced light confuses you during analysis. Eliminate it.

With your Poser Display Units set to inches like me, create a Point Light and set its intensity and XYZ position like so

Duplicate that light to make Light 2. Move it like so (behind the figure's right ear)

Close your figure's eyes and mouth so they don't distract you.

Set the head material like so, move your camera a little to the side, and render like this. There should just be a sliver of light coming on the edge of the head on our left.

If the terminator (the line where light ends and shadow begins) is jaggy, subdivide the head once as I did with James.

This is the color (gray, level 143) I used on the diffuse "Color" of the PhysicalSurface node. You can try others but don't do black, or we'll get nothing later on.

First experiment. Let's change the ScatterDistR_mm to 5. We don't really know what this means yet. Let's render.

If you have progressive rendering on, you get to see the pixels cook. On the first visual update, this is what I see. Does this surprise you? It surprises me.

Let's stop a moment and consider what all that blue could possibly mean to us.

It seems that the scattering is taking away some of the red from the direct surface reflections, and sending it to come out somewhere else.

This is perhaps plausible, but where are my gray-level 143 diffuse reflections from the direct surface?!? I don't know. But they show up after a while.

Possible takeaway - some of these effects cannot be properly shown with only a few samples. Always cook plenty of samples.

I have a theory that if I also set the ScatterDistG_mm to 5, I will get yellow scatter about the same distance as I previously got red. Why? Because equal amounts of red and green appear to be yellow in our eyeballs.

Suppose we now set the red scatter to 10mm? What happens is that the red scatter goes twice as far as the green scatter. The result is we have yellow close to the entry point of light, and red further away from the entry point. We do NOT see more red everywhere, just more red at greater distances. The amount of scatter close to the entry point is still equal amounts of red and green. We are not changing the AMOUNT of scatter, but rather the ATTENUATION of the scattered light.

Oh god a full blown tutorial. Thank you! I'm excited. Will follow this along - let me save the scene I had and prepare a new one following your instructions.

So the scatter parameters of this are offering us a way to directly modulate the scatter ATTENUATION, based on distance from entry point, for each of the red, green, and blue color channels of the rendering engine. (SuperFly only has 3 channels - other renderers have more and can be more accurate about scatter. SuperFly is not physically correct - it's a severe simplification - because it has only 3 channels to play with. This is not serious problem in practice but it does limit what you can do.)

Now - if we want to plug SOMETHING into these parameters, to modulate them, that SOMETHING has to be a representation of the scatter ATTENUATION.

Who has an attenuation map for there figure? Anybody? Anybody?

Thought so. Nobody does.

Are we dead though? Hardly! We have nodes. We can do math. We can take OTHER maps and INTERPRET them until they speak the language of attenuation!!

Does this look familiar?

So I decide to mess around with the rgb scatter mm's until it sort-of looks like that girls ear.

Here we go.

you'd want to make your own custom compound node that did all the stuff you needed. It would look a little like my shader.

So now our gray-skinned but red-blooded head looks kind-of correct. Humans are not gray skinned though. Let's fix the diffuse color with - a diffuse color map. We'll keep the attenuation values as is for the moment. Let's see what we get.

OMG! There is WAY more scatter now, and it's way more red than before.

OMG! OMG! That was totally unexpected right? Or is it?

On this node, there is no place to modulate the AMOUNT or COLOR of the scatter - just the attenuation. It's not unreasonable to suppose that the programmer (Stefan) wired up the Color input to both the internal Scatter and the internal Diffuse.

I say again - the Color input is going to modulate both the diffuse reflections AND the scattered light. They are not isolated. They cannot be modulated INDEPENDENTLY.

If you plan to use a diffuse color MAP and a scatter color MAP, you cannot use this node. They are not separate inputs. They cannot separately be modulated by independent sources.

If that's what you're trying to do, use separate, independent color maps, one for scatter COLOR(amount) and one for diffuse COLOR(amount) then stop right now.

Just to confirm how this is wired up, I went back to no map, but used Color white. The total amount of scattered light is now extremely high.

Again, but now with color black. There is no scatter AT ALL. Scatter AMOUNT is 0.

Now does this mean we can't use anything to modulate scatter? No, we can modulate scatter ATTENUATION. And overall, the diffuse color map is useful to modulate the scatter AMOUNT. So in that sense, the color map is the scatter amount map. Is that wrong? Actually no it's not. It may not be what you need, but it's not entirely wrong.

Can you modulate the scatter ATTENUATION with some kind of reddish map? Sure - let's try. I'll reuse the color map.

The attenuation inputs are in separate RGB values. How could we break the color map into separate RGB values? OMG there's a node literally called SeparateRGB.

So - if you have a color (3-channel) attenuation map, or something LIKE one, you can plug it in that way. I don't know if that's really relevant though. I have never seen anybody demonstrate that skin appears mo'betta' if you avoid using the same attenuation on it everywhere. Prove to me you need it, and I'll believe it. So far I haven't seen it.

This is equal amounts hilarious and scary, the way you teach it - I don't know why, but it is, and I'm loving it

Alright. My Poser had bugged on me and I had to rebut, so only now I'm managing to get any render test here at all. I can't really prove anything to you - only that I've seen renders with subdermal maps modulating the scatter color and you can see underlying veins and such, very subtly, and things like the skin around eyes looking more purplish than reddish. And then people modulating the depth of the scatter with grayscale maps, to make things like fingers and ears look more translucent against a light - but then I'm not sure if Poser simply detects how thick a part is and uses that for the translucency.

I don't know things well. I'm just gonna test my maps for now, and show you In a minute when the renders cook.

Fair enough.

Meanwhile, just to show how far you can go with NOTHING but the color map, here I cheat and use it as a bump map and a specular map and a scatter attenuation map. This is not bad for a background figure.

Alright, a control render - using your color for the scatter.

With my map.

I can see the difference, but I'm not sure if I like it. The purple bits are showing up as gray instead. The veins I've painted are too subtle to be noticeable - I wonder if this should be a map less like usual subdermal maps for other engines, and more meant for Poser. I'll try to make those purples more saturated and bright, and the veins more obvious.

Maybe - or maybe you really can't use that map as a depth/attenuation parameter and get the effect you want. I think that map is more of a scatter AMOUNT, not distance.

I don't think it likes the purple - I can only see it where it's dark. It also seems to have made her ear less translucent where the purple is, or is that an optical illusion or something similar?

bagginsbill posted at 2:52PM Fri, 28 June 2019 - #4355206

Maybe - or maybe you really can't use that map as a depth/attenuation parameter and get the effect you want. I think that map is more of a scatter AMOUNT, not distance.

It was supposed to have been scatter COLOR.

I'll learn. Still testing things here.

ghostship2 posted at 2:57PM Fri, 28 June 2019 - #4355192

you'd want to make your own custom compound node that did all the stuff you needed. It would look a little like my shader.

Forgot to reply to this - your shader is amazing, but it's a jack-of-all-trades. I'd rather make a specific shader, as this will be a commercial product and I prefer to have absolutely no node there that isn't needed.

Unless you mean I'd have to build a surrogate for a node - in which case I don't have nearly enough mathy knowledge to try to make something like that.

Ok, I removed all the purple and left only the veins - I can see the veins but only in the ear and nowhere else.

This leads me to believe that I'm seeing proof of what BB said - this isn't really changing color much, more like it's changing the amount of scatter. So I can only see a difference when it's against a light making most everything be very translucent.

What I intended was that, on your body you have a thin layer of skin, and then it's translucent enough so that you slightly see the colors of what's underneath it, more so under a strong light. Can Poser truly not do this?

What happens if I switch to a PoserSurface root with a CustomScatter node instead? Does this give me more control? Or will it be the same thing in the end?

Here is my test render using your SSS proposal. In my test-render, with my test scene and lights, I think 4,6,6, is a bit on the strong side : 'too waxy".

You're trying to get diffuse-reflected (i.e. near-surface scattered) light to show those colors, but you're connecting them to the attenuation distance. Won't work - it can only alter the distance it takes to fade the deep scatter, as I demonstrated.

Instead you probably want to not use the Physical Surface, as the starting amount (ratio, aka color coefficient) of scattered light is not under your control and you have no place to plug this map in.

tonyvilters posted at 3:08PM Fri, 28 June 2019 - #4355213

Here is my test render using your SSS proposal. In my test-render, with my test scene and lights, I think 4,6,6, is a bit on the strong side : 'too waxy".

Yea but it's not JUST 4,6,6 it is 4,6,6 MULTIPLIED with your color map, which is brighter than my color map. When abused this way, brighter becomes "travels farther". You need to make lower ratios than 4,6,6 because your input is starting with bigger numbers than mine.

AO, in your original post you asked how to connect a scatter distance map. I showed you how, using SeparateRGB, but then what you plugged in is NOT a scatter distance map, it's a scatter color map.