There is a new skin tutorial in the tutorials section on skin by Iuvenis_Scriptor.  It looks wonderful, and I wanted to follow it.  On page 3 he has a zip file to download, and I can't seem to get it.  Can someone test the link?  I've contacted the maker who believes the link is alright, so I just want to know if I'm doing something wrong.

Page 3 is here:

http://www.renderosity.com/mod/tutorial/index.php?tutorial_id=2233&page=3

The link is here:

www.hsmespanol.com/SkinFresnel.zip

Thank you, I really wanted to use this tutorial.

Indi.

Nope, get a page not found error message.

:)

Attached Link: http://www.hsmespanol.com/SkinFresnel.zip

Same here.

Quote -  The hyperlink is wrong. Try this.

You're amazing.  Thank you!

Indi.

No, the difference is not in the visible hyperlink text, but in the URL that the hyperlink points to.

My hyperlink points to the .zip file, whereas yours still points to the (non-existent) .mt5 file.

(edit) You fixed it - YAY!!

I accidentally deleted the previous post of mine, but that explains alot!  I thought altering the text automatically altered the link.  Anyway, I've made another attempt to fix it.  Anyone fancy testing it for me?

Yep, it works great now!

Now off to make the shader, thanks so much!

Finally!  Indigone, if you don't mind sharing, I'd love to see the render(s) you make with the shader!  Also, if you think there's anything that could stand to be explained more clearly, please don't hesitate to let me know.

Thanks everyone for your help!

I'd be glad to.  May take a couple days, but I'm very happy to see how this goes!  Thanks for the tutorial.

Indi.

dphoadley@Iuvenis_Scriptor
I don't use the generation 4 figures from Daz.  My figures of choice are my remapped versions of Posette, Dork, and Judy.  Will your material work with them, and/or how much adjustment will be needed say for the transmapped eyebrows to work with V3 head textures?
Yours truly,
David P. Hoadley

No adjustment is necessary aside from the maps themselves. As long as the eyebrows (and all other facial/body hair) are separately transmapped onto completely (or very nearly) hairless skin textures, the shader will work on any figure.  For a V3 head, for example, all you'd need is a browless face map and an eyebrow transmap, both compatible with V3's UV layout.

Quote - If I remember right, I set ambient occlusion to 1.000
DPH

Are you referring to the AO Strength parameter in the AO node? That parameter doesn't do anything at all. The strength is always 1, even if you put a 0 there.

Quote - Also, since this is an old MaskEdit V3 texture, I dispensed with the anti gamma part of the tutorial.
DPH

What does that mean? You didn't do anti-gamma on the way in or you didn't do gamma on the way out, or you didn't bother with any of the gamma steps?

And why would you state that the reason you ignored the need for gamma correction has something to do with it being an "old MaskEdit V3" texture? Because its old? Because its MaskEdit (whatever that means)? 

I'm trying to understand if, despite my having explained this for what seems like 10,000 times, you think you don't need to gamma correct. Because your render looks like you do.

Iuevenis_Scriptor:

I'm going to call you IS because your name is hard to spell. I have to look at it 3 times. I ... U ... E ...  What is that, Latin?

Anyway, I'm playing with your tutorial. You didn't specify what your Poser Display Unit is set to. That means we can't reproduce your shader. When giving screen shots of material settings, you must always tell us what your display units are. Without that, the displacement strength, ray bias on AO, and a couple other numbers are just guesswork.

IS:

While I'm still experimenting, I have a question. Did you know that Blinn is a specular with Fresnel effect built in?

All this time when you spoke of including my Fresnel nodes into your shader, I thought the reason was you were using real reflection and you needed to control that. That's the only time i use those umpteen nodes. For 99% of all specular effects, and particularly on skin, a single Blinn node, or maybe a Blinn powered by an Edge_Blend, does the job.

Thanks for checking it out, Bill!  Calling me "IS" is fine.  Aside from being a hobbyist CGI artist, I'm also an aspiring writer and linguist.  "Iuvenis Scriptor" means "Young Writer" in Latin.

I am aware of the utility of Blinn, and for the longest time I did use it.  I just decided on a Fresnel-powered Specular node for that extra bit of accuracy that you gave me the impression such a move would yield.  Given your most recent clarifications on the matter, however, I may yet put an addendum on the tutorial explaining that a Blinn node, with or without an Edge Blend, would be a reasonable substitute.

Thank you also for alerting me to the units issue.  I use default, which I believe is US Customary.  I'll add that info as soon as I finish my homework for my Spanish and French classes. Might it be a good idea to indicate my render settings as well?

Don't bother mentioning the Edge_Blend with Blinn - just mention Blinn. I saw face_off do that, but I never really understood why. I think it was a roundabout way of conserving energy, which we know is accomplished much better and more directly by subtracting the output of the specular node from the diffuse value. Although that still isn't correct, but its reasonable.

Oh and the Fresnel equation we were using is only for a microscopically smooth surface, such as water or glass. For micro bump, as found on skin and just about everything else, the amount of reflection does not follow that idealized curve. The Blinn algorithm is a direct attempt to capture the idea of the Fresnel effect for a micro-bumped surface. That is what the eccentricity parameter is for. The Specular node does not model this, and so it does not produce the nice "rim light" effect. It can't, even if you amplify it via the Fresnel equation. It just doesn't capture what happens at the edges of a micro-bumped surface.

So my first suggestion would be to dump the Fresnel nodes and just use a Blinn. Not only because its less nodes, but also its more accurate for the specular effect on human skin (and a lot of other things, too.)

Quote - I use default, which I believe is US Customary. 

Huh? There's no choice in Poser Display Unit for "US Customary". The choices are PNU, Inches, Feet, Meters, etc.

What is your unit set to? I don't know what the default is, and even then we shouldn't rely on you remember if/when you ever changed it. It's in preferences, interface.

Render settings is always a good thing. Many people know what to use themselves, but quite a few don't. After 20 people ask you why they get no displacement or no AO, you'll just come right out and paste "Enable Displacement and Ray-tracing and Shadows" as a matter of course.

Thanks again for your continued input.  The unit is Feet.

Oh dear, oh dear. I have really bad news for you.

I kept looking at your renders, and I kept thinking that I was seeing the opposite of the Fresnel effect - that the specular reflection was strongest pointing toward the camera and it was altogether too much and looked metallic.

Today I built your shader in matmatic, and I didn't bother downloading your set of nodes for the Fresnel effect. I used my TrueFresnel function, assuming that you built the same thing. Everything was looking fine and my renders did not look like yours or DPH's. My implementation of your shader, with my Fresnel equation in it, looked right.

So I remembered that you "reverse engineered" the equation from my glass shaders. And so I thought perhaps you might have made a mistake.

Boy did you ever.

You actually made two, but one is minor and the other is egregious.

The very last node of your version you do 1 - the fresnel effect. I have that part in my shader, too, but that's for the diffuse value, not the reflection value. The diffuse value is the complement of the reflection value.

But in your transcription, you used the complement of the reflection value for the specular strength. This resulted in the completely opposite effect. The specular is strongest pointing at the camera and weakest pointing away.

You should delete that tutorial and start over.

The other problem is that you transcribed one of the factors wrong. There is a term where I'm computing the sine of the angle of transmission and the correct factor is (1 / IOR) ^2, but you used 1 / IOR. This doesn't change the results very much unless you are using a high IOR. Because you're using 1.45 it didn't make a big difference. But the other problem, using the complement of the calculation is a big deal.

I loaded your Fresnel nodes, ran that through a gamma-correction (to see it more clearly) and plugged it into alt-diffuse. Yours is on the left.

I did the same with my Fresnel nodes, on the right.

Quite the opposite of what it's supposed to be.

This is directionally correct, but slightly inaccurate due to the incorrect factor, where you were supposed to use 1 / 1.45 ^ 2, but used 1 / 1.45 instead.

And also, I'm curious why you thought to use 1.45? Water is 1.33 and skin is 95% water.

I looked up the IOR of skin, and 1.45 was roughly the average of all the figures I found.

Wouldn't doing what you're suggesting put all the sheen and SSS on the edges of the figure instead of having it towards the center and fading out towards the edges?  I thought the purpose of Fresnel was to weaken specularity/reflection along the edges.

OK 1.45 seems reasonable - I was just curious because I never looked it up.

The Fresnel effect strengthens reflections along the edges.

Look at your watch. Hold it up at eye level. Face it towards you and concentrate on the reflections in the glass. Now turn it almost 90 degrees, so the glass faces sideways so you can just barely see the watch face. Observe the reflections - it becomes a near perfect mirror.

"What does that mean? You didn't do anti-gamma on the way in or you didn't do gamma on the way out, or you didn't bother with any of the gamma steps?

And why would you state that the reason you ignored the need for gamma correction has something to do with it being an "old MaskEdit V3" texture? Because its old? Because its MaskEdit (whatever that means)? "

I've been given to understand that V4 and M4 textures have gamma correction already built into the textures, so to install gamma correction for Poser 6, for which this tutorial was intended, Young Writer first had to counteract the gamma correction in the texture.  Since I was using an older V3 texture, such measures would have been for me counter productive, since I do gamma correction within my Poser Pro render settings.  Also adding the ambient occlusion node, since I set that up already with my lights.

I also made at least one other divergence from the tut, and that was to link my text image directly with the main diffuse slot, since othewise it was rendering way too pale and way too bright.  I set the diffuse color to white, but the value I reduced to something like 0.001.

Bagginsbill, I know that you believe differently, I find that white absorbes light more naturaly than black, but that by regulating the value of white I can manupulate this absorbtion better than by using shades of grey and black.  I also always do thesame with specular and ambience.

I cant't remember who ariginally touted the idea of setting the colors to white and adjusting the value, but I believe that it was either BluEcho of Victoria Lee.  Anyway, when using this method previously in Poser 5, my renders came out looking more natural than they had otherwise.
DPH

And, yes the sheen is supposed to be strongest on the edges, but only when the light source bounces the light at the right angle towards your eye. Which means that when the figure is front lit, the edges show no specular effect because the light bounces to the back, away from the camera.

But when the figure is back-lit, the specular reflected light bounces towards the camera and becomes very strong.

This is precisely what the Blinn node calculates.

On the left is with my VSS shader, which uses Blinn.

On the right is with my implementation of your tutorial, including the Fresnel bug. I had to make a couple minor changes because the hair is painted right on the texture and there is no hair mask for this texture set.

In the middle is your shader, but with the Fresnel complement bug corrected. Notice that your shader with correction and my shader look very similar, but your shader with the bug (#3) is way too shiny.

What about SSS?  How does it relate to incidence angle?

Oops - that last render was with wrong render settings. I was doing a fast bad-quality render just to check lighting levels. I'll do it again.

Quote - What about SSS?  How does it relate to incidence angle?

That's a good question and I'll talk about that in a minute. I'm doing another render and also will do one with rim lighting.

Notice I'm not suggesting that #2 is good (the one with correct Fresnel). Both #2 and #3 are wrong. #2 is too little, #3 is too much. That's because the Fresnel equation you used is for a smooth surface. Skin isn't like that. You want to use Blinn to get the right balance in all directions.

Compare the VSS shader to yours. Because you're not using Blinn, but a Specular node instead, you're not getting the sharp bright specular on the edges. Even with the correct Fresnel (in the middle) its not there, because the Specular node just can't generate that kind of response. The Specular node is a left-over from 1985 and should not have even been included in Poser, in my opinion.

Notice also that the colors in your shader have gone cold. That's because of how you did the SSS. You aligned the SSS with a specular node, which means it only happens mostly in line with the half-angle between the observer vector and the light vector. And also, you used a fastscatter node, which as face_off and I have written many times, is utterly useless.

Whereas, my shader is nice and warm in the front like it's supposed to be.

This is actually a combination of multiple effects. We talk about front-side SSS and back-side SSS.

Back-side SSS is when light passes straight through a translucent material, like the thin part of an ear. When lit from behind the ear seems to glow red. This is what fastscatter is supposed to do, but it does a terrible job. There's really no good way to do back-side SSS in Poser. It should only happen on thin parts, which means mostly the ears. For most renders, this is a detail that can easily be ignored without ruining the render.

Front-side SSS is when the light goes in, bounces around, and comes back out near the area it went in. The greatest part of this should happen on the shadow side of the "terminator". The terminator is the transition from direct illumination to indirect illumination. Face_off built his shaders to do this by incorporating knowledge of the position of your dominant light source. From that and the actual normals of the surface, he calculated where the surface was slightly pointing away from the light, and he added red there. This is very effective, but a pain in the ass, because you have to enter the coordinates or vector to your light every time you move the main light.

He wrote scripts to do this, and those are the "real skin xyz" products. They all rely on the user clicking a button to re-calculate where the front-side SSS should happen.

I never liked that because it's a pain in the ass to keep running the script every time you move a light and that doesn't work for animation at all.

So for years I've been trying to find workarounds. At one time I simply ran a Diffuse node into a Color_Ramp and where the Diffuse was low, it generated some extra red. Face_off liked that idea, and built it into the Daz V4 and M4 shaders. But that technique has a bunch of problems. One, it doesn't behave too great with certain IBL situations. Two, in very low light, it makes the figure seem to glow, because getting the balancing act between light and shadow cannot quite be based on an absolute value coming out of the Diffuse node. It would really require that we know more about the light intensity and direction, information we don't get in a Poser shader.

Another technique i tried was to simply add red wherever there was little or no specular. This had other problems.

My most recent trick is all a subjective hack, but it seems to work pretty well. It involves a combination of measuring the measuring the diffuse value, inverting that, running it through a little more wave-shaping math, and then multiplying that back with the diffuse value. People seem to like it. Also, this technique has very easily adjustable characteristics, so when it isn't quite right, the user can make a change to a single parameter and make it work.

You can see it best in the abs on his belly, his left side, and also on his left thigh. It's also quite clear on the bridge of his nose. People notice my noses look really good - it's because of the SSS happening at the terminator.

Slight highjack here, but regarding the ear translucence issue... what would happen if you added very slight transparency to the outer edge of the ears (say, a creamy-grey)? Anything?

Then they'd be transparent and you'd see what's behind them.

Now that I got that out of my system, would you mind sharing your front-side SSS technique sometime this weekend?  Hopefully I'll find the time to tinker with it between the two first-draft papers I have to prepare.  In the meantime, here's the result of using a Blinn for the sheen and an Edge Blend in place of the Fresnel value for the SSS (outer black, inner white).  As expected, the skin came out looking dry and bland to me, but maybe better front-side SSS will help that.

Thank you as always for your continued support and patience, Bill.  I may not be where you are yet, but I owe alot of whatever I have thus far acheived to you.

Quote - Then they'd be transparent and you'd see what's behind them.

Like lights? or not?

I guess I'm just thinking that a cream colour would still be sufficiently opaque that you're not likely to see actual objects, but it was just a guess. I've never built a transparency map, so I figured it would be quicker to ask than to try to figure out how to set one up (though I'll be doing that soon enough for eyebrows etc for the character I'm working on).

I've experimented with that node, and that node ALONE, for at least 100 hours.

You haven't even got started yet.

Besides the color, there are three parameters on that node. They need to be adjusted simultaneously in order to get all the right effects. I have reduced this to a single parameter for VSS users: Shine. What you've seen in previous renders was my VSS Shine = .25 - the default.

Here is Shine = 1.

Quote -
In the meantime, here's the result of using a Blinn for the sheen and an Edge Blend in place of the Fresnel value for the SSS (outer black, inner white).  As expected, the skin came out looking dry and bland to me, but maybe better front-side SSS will help that.

Hm. You may think that, Iuvenis... but to be honest, I think this looks more realistic than the example image on the final page of your tutorial. Of course, the only way to really compare would be identical lighting and pose.

Quote - And here is Shine = 2.

To me, this is really cool, because he looks wet, rather than looking plastic like most shiny characters usually do.

Hi IS,

Ok, I managed to build the shader just like you said up to the AO/GC part, so I'm not ready to post renders.  I did it with V4 and her simple face map.

It's pretty easy to follow, and from my standpoint, is a great learning experience.  I'm glad BB is here to help us understand it too.

It's fascinating how you build the SSS, sheen, and displacement map right from the color map.

A couple of non-technical things about the tutorial, quite picky actually, but did have me guessing some. 

I liked the way that when I followed the tut, word for word, my nodes were named exactly the same as your nodes, even after copying in the massive stack of fresnel math nodes.  In your shader there is no Blender_3.  Maybe it was deleted?  Your Blender_4 became my Blender_3 and once I got to the point where I was hooking everything up, it started making my eyes spin.

On the page where we're plugging in all the nodes (page 6 perhaps?), you say to plug sheen into the alt spec channel.  I think that was blender_2, but I'm not positive.  Probably could be corrected.

And on the same page, you refer to the displacement map as blender_6 and I'm pretty sure you meant your blender_4, which was my blender_3.  (LOL)  At that point my head was spinning, so I did a quick render and went to bed.  It came out ok, but definitely needed the gamma correction.

So with all that said, let me ask a couple of questions.   I don't have any textures with SSS maps, but I hear they exist.  So I'll let Poser build that one, but if I have sheen (specular?) maps and displacement maps that I like, should I plug those in instead, and blend them in as you have done?  I'll do some comparisons when I get this done.

I'm looking forward to working with GC and AO part, because I'd really like to know how to put that on a texture that I otherwise already like.

Thanks for building this, any chance to learn something new from someone is wonderful.  And thank you BB for taking a look at it too.

Indi.

Edit:  Um... somehow I had only read page 1 before I typed this long post.  I see there may be enormous changes anyway, so I'll be patient, hehe, sorry!

If you have an SSS map, you should use it. If you have a specular map, you should use it.

 very important >>> If you have a bump map, you should use it.

Frankly, I'm not too big a fan of using the color map for bump. There is no reason to correlate color with bump, and many good reasons not to. This shader, for example, makes a dimple wherever the color map has a slightly darker color. Bump is really important and all the good textures sets come with a bump map that is NOT a derivative of the color map.

If you're using a texture set where the author, being lazy and/or naive, has actually made the bump map by converting the color map in any way, then you should not use that. You'd be better off with just a turbulence node. This is what I supply in my VSS shader when the figure doesn't have a bump map.