bagginsbill posted at 3:11PM Fri, 28 June 2019 - #4355215

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.

PhysicalSurface isn't a requirement of mine. I just started out with it because I was curious about it.

Any way to do it with PoserSurface instead? Would it require a ton of node noodles?

bagginsbill posted at 3:13PM Fri, 28 June 2019 - #4355217

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.

Sorry, I made myself not that clear - I meant both maps but the most important to me is the color one.

I have another for the distance, which I'm not using yet. I'll show what it looks like here then.

One way to do veins is to multiply your original color map with blue, then blend the original with the blue-tinted version using a vein map like here. This is affecting the scatter as well but not just the deep scatter, also the near-surface scatter. In other words, you just draw the veins on your color map (subtlely) and all is well. I did it with nodes but you could do it as well in photoshop.

But in that case, it wouldn't look more intense with more translucency (like, seeing veins better in a back-lit ear), would it?

Edit: wait, I think I didn't process the idea well. Maybe it would.

If you literally want to use a scatter color map, you would want to use one of the pure scatter nodes. I don't have a scatter color map but I'll simulate one with that vein map and some nodes.

If you have La Femme, I could send you my face maps, the three of them. (Diffuse, scatter color, and scatter depth/whatever that's called)

I have LaFemme. Send me a link via PM if you have dropbox or something easy.

I tried using a color map with veins with just the PoserSurface to see if the veins appear as blue or black when backlit. They appear black. Now you have me wondering what is realistic? When white light is reflected bluish (back in the direction it came) does that mean that the red went on through and comes out the other side or does it mean that red would be completely attenuated and come out black on the other side? I don't know.

Sent you a google drive link!

bagginsbill posted at 3:35PM Fri, 28 June 2019 - #4355225

I tried using a color map with veins with just the PoserSurface to see if the veins appear as blue or black when backlit. They appear black. Now you have me wondering what is realistic? When white light is reflected bluish (back in the direction it came) does that mean that the red went on through and comes out the other side or does it mean that red would be completely attenuated and come out black on the other side? I don't know.

I uh--

If you don't know, who am I to even try.

Correction. I accidentally wired up the nodes wrong. The simple use of a pink-and-blue color map affects the translucence just fine.

bagginsbill posted at 3:47PM Fri, 28 June 2019 - #4355231

Correction. I accidentally wired up the nodes wrong. The simple use of a pink-and-blue color map affects the translucence just fine.

Oooh I see those subtle-looking veins in the earlobe showing up more with the backlight! How did you achieve that?

I'm testing the CustomScatter node here - I might have made a mistake with the translucency/depth/whatever map in the ears area. It's only allowing the border of the ears to get reddenned with the light.

Aaaand now it dawned on me that I should darken the veins in the translucency map too. Right? No?

Afrodite-Ohki posted at 4:24PM Fri, 28 June 2019 - #4355232

Oooh I see those subtle-looking veins in the earlobe showing up more with the backlight! How did you achieve that?

Like I said, that's just my "color" map plugged into color and all the other stuff same as before. It shows up both for front-lit and back-lit scatter.

It's the same shader as before on James, but now this veined color map is my color map.

bagginsbill posted at 4:31PM Fri, 28 June 2019 - #4355238

Afrodite-Ohki posted at 4:24PM Fri, 28 June 2019 - #4355232

Oooh I see those subtle-looking veins in the earlobe showing up more with the backlight! How did you achieve that?

Like I said, that's just my "color" map plugged into color and all the other stuff same as before. It shows up both for front-lit and back-lit scatter.

It's the same shader as before on James, but now this veined color map is my color map.

Oh ow, I could swear it was going differently for depth.

Hmm, I'm gonna test playing with my maps a little, will show in a moment.

And after going around and around I went back to the basics. How does this look? Looks pretty nice to me but I don't know what to look for to know if it's wrong.

(The textures have changed from those I sent you - I added the veins to the diffuse map at 5% opacity, and with the translucency map I softened the white parts to make them grayer, made the entire ear lighter as opposed to only its border, and added the veins to it too)

I am stuck investigating numerous unicorns. (Things I didn't think were possible) As none of the nodes are documented, I am forced to explore their behaviors.

I have learned some weird things about the scatter in SuperFly. Thin blue lines come out completely differently depending on very minor changes IN THE RED CHANNEL. This should not be possible.

This is La Femme's right ear. The shader is using a Tile node with carefully chosen colors. The light is the front light (behind the camera).

Nothing unexpected here.

Now comes the unicorn. I simple turned off the front light and turned on the back light. The ear should glow pretty uniformly from the back lighting,

!!! but the blue parts DO NOT glow !!!

I can only conclude this is a bug. It is crap like this that makes me stop using Poser for a while.

Your problem is that veins are blue - a certain shade, against other certain shades. In that case, they turn black. Other, different, demonstrations with just shades of red don't show this behavior.

Further digging - just using any solid color with Burley scattering that has little or no red in it will not scatter properly.

I give up for a while. Have to use something besides Burley. But Burley is the best skin shader. Damn it.

WTH.

And it's happening with any scatter? I was hoping not to have to use Cycles - I have comments of people claiming that it takes too long to render. (Not for me but... different computers and all.)

Well, in any case, seems like you've hit a wall - thank you very much for trying and for the lessons anyway!

bagginsbill,

Thanks for posting this. I completely forgot about the ScatterDist inputs. The Physical Surface Root looks like a good way to make a simple decent skin shader. Can't wait to try it out on some of my older Victoria 3 characters.

Thinking out loud here.

The Color/Diffuse map is usually the only map you use for any type of color. The others maps in the PBR metal-rough workflow (which the Physical Surface Node seems to be designed for) are grayscale maps (well, with the exception of the normal map which is blue, but it works similarly to a bump map).

So here's my thought, which I am going to experiment with. I suspect that what the R G and B scatter inputs are looking for is grayscale data, not color data. Like a bump or displacement map would.

You SHOULD, however, be able to paint in the individual channels in Photoshop ... paint the areas you want to be red in the RED channel of your image; the areas you want blue in the BLUE channel; and the areas green in the GREEN channel. It will probably look like a pretty nasty color image when you collapse the channels back into the RGB image, but if you use the RGB Separator node you might get what you are looking for in the Physical Surface node.

And maybe the values in the Physical Surface node might be the MAXIMUM distance for that SSS color, with white in the image producing the maximum value, and black producing no value.

Off to experiment.

OK. Not quite what I expected. LMAO

Here's what I did in Photoshop. I viewed the channels in an image. Filled each of the R, G and B channels with black, and then added white text to represent which channel layer the text was on. The large image shows the result when you view the image in RGB, but you'll see the text on each layer in the layers palette.

So then I applied the SimplePBR shaders to LaFemme (they use the physical surface node for SuperFly). Then I added the SeperateRGB cycles node. and piped this experimental image into it

But the result was not as expected. I had to bump the scatter distances way up in order to see the red and green (which were not red and green as I expected), and the blue channel didn't seem to be doing anything.

It is non expected result of the Separate RGB node. I made this map, Soft dots on BLACK, Alfa in the middle and WHITE on bottom.

Tja, the ALFA above shows as grey, but it is Alfa channel.

With the "R" of this map connected you get the following result : The "R" is removed.

With the "G" connected, the Green is removed.

With "B" connected the Blue is REMOVED. The node is doing the opposite as expected.

And as could be expected here the GREEN and BLUE added up.

Only the RED on ALFA and WHITE is left over. => This would normally be expected as to be the "R" channel.

And on the BLACK and on the WHITE, you see both GREEN and BLUE removed.

Comments please?

Oh, and my little test does not explain why the BLUE did not show in Deecey's test, so there is more going on.

I think, I am wrong. Pfft.

Yes, nothing wrong with the Separate RGB node. Results are as to be expected.

Yep, it was why Deecey did not get the blue to show, that got me off track. Back to the drawing board.

The result Deecey got is as expected as well. She used the skin diffuse map which contains mainly red and green and very little blue. The color input does modulate both diffuse reflection and scatter light, as BB already pointed out earlier.

This is giving me a headache and I'm feeling dumb. LMAO!

Not sure if this helps. The following demos are based on cycle tutorials I've read and some experimentation. Three lights, Red, Blue, Green where used to generate the source colour. Note the top two are reflected light on a plane. The last one is light passing through a cube with the camera of the other side

1. ScatterDist params modify a blur radius as well as transmission depth
   

2. Scatter Inputs expect values not colours - the green light radius is modified only in the top image, the bottom the radius of each colour are changed equally

3. Surface colour modifes the final output - Effect of a skin like colour in the diffuse parameter on transmitted light.

One other - effect of ScatterDist on transmission, only green channel comes through

And what of this would explain why it's rendering blue scatter as if black?

Nothing at all, I didn't take in that the thread had changed from physical surface node to Cycles part way through.

ironsoul posted at 2:47PM Sat, 29 June 2019 - #4355374

Nothing at all, I didn't take in that the thread had changed from physical surface node to Cycles part way through.

From what I understand, anything Superfly is doing that - no matter the node. I could be wrong though.

Hmm. if I repeat the test with the cycles surface still seeing blue so looks ok to me but I'm probably doing something stupid, I can only hold one idea in the head at any one time and discussions on realistic skin settings seem to require a mental juggling ability

This is the test set-up. Andy represents the camera

If I understand correctly, the blue is only turning black when it's against a fleshy color. I think?

Possibly related to scale Here is the test with BBs params (apart from use of pure blue for clarity

If I increase the scale

Which is still troublesome for me as I can't really afford making a milk-looking skin xD argh.

If I make a slight adjustment to radius eg change "1,1,1" to "1,1,1.00001" it works ok even with scale set to 0.01