I have a shader that BB made that is for stockings, but this presupposes you are using it over V4's legs. I am just looking for a sheer nylon material that I could apply to other clothing items, say like a jacket or blouse.

Just playing with the idea: apply the shader on whatever material you want to make nylon, replacing the skin colourMap with the ColourMap of the obj material. If it is already a smooth material, it should work. If not, just use a colour instead of the colourMap (SimpleColor() node).

I tried plugging the Tile set up that BB showed us on a mesh/net shader. I did not like the result. I'm afraid I must rely on experts like you to come along and figure it out. Or show me a screen shot of a shader I can build. Or, ask someone to separate a shader. BB gave away a shader that was water puddles on cement. I love the shader and would like to have just the cement/pavement part of it. I do not know how to separate out the puddles from that shader. Maybe you have an idea of what to do or can show me?

Actually, I've been playing with that Nylons shader of Bill's and have made it GC... I think. I'm just clearing my Poser drive of old stuff (going on to my external) so give me a few minutes and I'll see what I can come up with.

Expert? Not by a long shot... but thank you for the compliment. I base my stuff on a Sherlock Holmes quote - used to be obsessed with Holmes, ideal man, I thought - which went:
"What one can invent, another can discover."

Generally holds true for just about everything except matmatic-generated shaders, since matmatic is very efficient and creates shortcuts / alternative node sets to what you would think.

I'll see what you come up with! I haven't had any luck with anything in my Materials folder, BB's shaders or otherwise (and I have a huge Materials folder).

So, what you're after is the same material with thickness of like what stockings are made of? Moving-Of-Files is done, so I'll have a go now...

Yes, very much so. I have at least 3 shirts like of this type of material in my real closet here at home that I use as over shirts when I go out in the summer.

Hmmmm, interesting.

Right. Had some issues with the mesh that needed sorting out. I can't really divulge whose mesh this is, but I had to add a edge-loop to the part right under the bust (bra bit) to get rid of some artefact, apparently caused by Poser's smoothing thingie. So, I took the mesh into Blender, added the loop, exported back out to obj and imported into Poser and turned it back into a figure with Obj2Cr2. One incredibly brilliant tool!

So, loaded the tweaked shader on the slip-portion of the nightie... the rest has been GCed but nothing else was done to the rest of the garment:

So, is this nylon?? It kind-of looks like it, but would you be happy with it? Or is this image the result of applying a material node set designed for another type of mesh? Well, legs... you know.

Anyway, the tweak is up for grabs, if you want to have a good laugh. Oh, and the mesh probably needs a bit more detailing where it bends: you can see artefact (those vertical lines) developing there.

Sorry to hear you had so much mesh trouble! The shirt I'm using for my image has a problem with a bit of poke through at one elbow.  Oh well. What I see in your image is close. Needs to be more transparent though. But it could be better than the simple thing I'm using at the moment. I'd like to try it out. In the meantime, I was working on my lighting for the scene I'm planning. Thank you so much for your efforts!

I'll help, but not today. Big family gathering this weekend at my house.

Hope you have a great time with the family, BB!

Hey so I'm back and have time to do Poser stuff.

What are we trying to make here - a standalone homogenous sheer fabric, 100% procedural? Or does it need to use textures? Areas with variable transparency, such as a combination shader that does opaque lacy in some areas and sheer in others? Or is it like the leg/nylon shader I posted a long time ago where there is one mesh but the appearance of two surfaces near each other?

Well, what I originally wanted was a sheer, nylon like material to apply to a shirt or jacket. I actually own two or three shirts like this that I wear as an over shirt in the summer. But others may like the idea of being able to add lacy trancparencies to the shader for items like frilly underwear.

This is very interesting. Long ago, when I built the nylon/leg shader, I did it by eye, the way everybody does. It took a lot of guesswork and never quite met my expectations for realism.

This time, as I always do now, I approached it from the physics first.

I made a simplifying assumption that the fibers are simply tiny cylindrical tubes, arranged in a criss-cross pattern. I also assumed (without too much loss of generality) that the orientation of the fibers is strictly horizontal and vertical. That isn't quite true, because rotating the cloth produces variety in the angles formed between the cloth surface and the viewer. This behavior is called anisotropy and I could deal with it if I could write shaders in software, instead of using the nodes, which limit the sorts of math I can do. But I think that I'm getting 95% correct results regardless.

Do you want to see the physics or no? It's very simple trigonometry, but I don't want to spoil the fun with too much "gobbledygook", as some people are offended by my brain, apparently.

I'd like to see the math.. and it's not like the forums are suffering from too many topics lately.

Sorry, I was out to the store. I'm liking what I see so far BB!

Quote -

Do you want to see the physics or no? It's very simple trigonometry, but I don't want to spoil the fun with too much "gobbledygook", as some people are offended by my brain, apparently.

Yes please.  I'd like to see the physics.

Ready?

Let's consider cloth fibers that are 100% opaque. Let's also consider only the vertical fibers arranged in a cloth that is straight vertically, but curved horizontally. We can model these using a bunch of tubes arranged in a semi-circle, like this.

The gaps are largest when the "surface" faces the camera. As the surface curves away, the tubes appear closer together, decreasing the gaps.

On the far left and right, the tubes actually overlap and there are no gaps in those areas.

Now when seen from a normal distance, the fibers in a cloth are so tiny we can't even see them individually, the way we see these tubes. But the same thing is happening. The gaps between the fibers let us see what is behind. The size of the gaps, relative to the size of the tubes, changes as the surface angles away, until the point at which there are no gaps at all, and the cloth becomes opaque.

So what is the math of these gaps? 

It's pretty simple. I'll just say what it is, but if somebody asks, I can make a trigonometry diagram to show how I derived this formula.

Let's let D represent the "density" of the tubes, expressed as a fraction. For example, suppose the tubes are 1 millimeter in diameter, and spaced so there is a tube ever 3 millimeters. Then the density is 1/3. Suppose the diameter is 2 milimeters and there is a tube every 4 millimeters. Then the density is 2/4 or .5.

Now we also need to consider the angle between the normal of the "surface" comprising these tiny tubes and the observer. Let's call that angle alpha.

Ready?

The opacity of the surface is

D / cos(alpha)

That's pretty f'ing simple, right?

Except what is opacity? Opacity is the complement of transparency. And since we're dealing with Poser, you may think we need to convert this to transparency. Well, strangely, when you plug a node into the Poser Surface Transparency channel, the data is actually interpreted as opacity, where 0 is completely not opaque and 1 is completely opaque. That's why transparency maps are 1 (white) for opaque, and black (0) for transparent. They are not transparency maps at all. They are opacity maps. But I won't bother trying to make everybody start using the right word for those.

Anyway, the shader to produce this effect is unbelievably simple.

In Poser, the Edge_Blend node internally calculates and uses the cosine of the angle of incidence. If you set the parameters to white, black, and 1, then the cos(alpha) is actually what it generates as its output. How nice!

So then I use a Math:Div node to divide the desired density by the cos(alpha). In this case, I used a density of .5.

If you go back to the tubes you can see that this sure does look like the same thing, only with microscopic tubes.

It's going to take me a little while to make the next demo. If anybody has questions, go ahead and ask.

I'm kind of amazed. I know patterning will up the complexity. Fresnel, specular...  Especially if you set the fibers at a severe angle to each other...

I don't know if it's what you are looking for but I came up with a stockingish material here (see the last version) and I can share the mat with you if you like.

The gaps are always the same, no matter where the curve is. The opacity is exactly what the "density" of the fibers is. If they are 30%, then 30% of view through the cloth is blocked, and 70% is open.

So - what to do about both sets of fibers.

Observe how the horizontal ones interrupt the gaps from the vertical ones.

Consider one "cell" of the pattern of tubes. The horizontal tube covers part of that cell. The area it covers is exactly the density value, i.e. if the tube is 30% density, then 30% of the cell is covered by the tube. The remaining 70% (1 - .3) is covered by the vertical tube. Of that, the coverage is based on the opacity given in the first part.

D / cos(alpha)

So the combined effect is:

D + (1 - D) * D / cos(alpha)

Consider a 50% density, where D = .5. When the surface faces the camera, cos(alpha) is 1, so the opacity is:

.5 + (1 - .5) * .5 / 1

which reduces to

.75

This makes sense, right? If each tube is half of the cell spacing size, then the little window left over is exactly half of half, or .25. The transparency is the complement of opacity. Since the window is 1/4 the area, and the tubes cover 3/4 the area, the opacity is .75.

What if the density is .3? Then the central (lowest) opacity is:

.3 + (1 - .3) * .3
= .3 + .21
= .51

Now whenever we have formula in this shape:

k * a + (1-k) * b

we should recognize that this is a weighted sum of a and b, with the relative weights given by k.

Any time we need a weighted sum, we can use a Blender node. The Blender node implements the formula:

(1 - Blending) * Input_1 + Blending * Input_2

In this case, we want Blending = D, Input_2 = 1 (or white), and Input_1 = the original formula from part 1.

This arrangement is shown in the attached screen shot.

It's still missing specular effects, and can only deal with opaque fibers, but it's pretty good just as it is.

I did not know Poser could handle this kind of micro scale. I want to see how IDL can handle it.

Can Parmatic produce an "opaque" border or something with a pattern to frame this shader?

Nice work,

gotta get up in 5 hours

This is what I see as genius: the genius of simplification.

Simple, you say? Well, yeah, sure, but who would have thunk it up? You took 24 hand-placed shader nodes and turned it into it into 4... with maths.

Another tick for maths. This is incredible stuff, Bill... THANK you!

Oh, and I'm pretty sure I got it... that Add() node Value_1 drives the whole thing, right? .5 density is like 75% opacity, .3 density is like 51% opacity, which is 49% transparency? etc...

You could make that a PM: Opacity node - forcing people to think in terms of those transparency channels as opacity channels.

Strangely, while Poser should have but does not have an anisotropic opacity node, it does have an anisotropic specular node. It is aptly named Anisotropic and found in the Specular branch of nodes.

I add one like this.

I have a rim light coming in from the left and the anisotropic node goes a little wonky. I suspect the formula it implements isn't quite right. I have a partial solution, but I must continue to experiment to tone that down.

Quote - I did not know Poser could handle this kind of micro scale. I want to see how IDL can handle it.

Can Parmatic produce an "opaque" border or something with a pattern to frame this shader?

Nice work,

gotta get up in 5 hours

No parmatic doesn't do any shader effects. Did you mean matmatic? Then yes. I'll be getting to that in part 5.

I have to get to bed myself, so I'll be pausing for now. But just so you know what is coming, so far I've only dealt with opaque fibers. But what if the fibers themselves are partially transparent? Nylon fibers are not opaque. This won't look like nylon until we deal with that. Part 3 will be translucent horizontal fibers. Part 4 will be translucent vertical fibers.

We'll be up to around 20 nodes by then.

Then part 5 will be how to introduce borders and lace into the equation.

Quote - So far all I did was black cloth. To support colors, turn on the Diffuse_Value and hook it up to the Anisotropic node with a subtract node, to maintain the principle of Conservation of Energy.

Dumb question:
Anisotropic out -> Subtract(1,1) -> Diffuse_Color is CoE, right?
[Q] Why Math_Function Subtract() and not Color_Math Subtract? Is Specular info numeric only?

Quote - This is what I see as genius: the genius of simplification.

Simple, you say? Well, yeah, sure, but who would have thunk it up? You took 24 hand-placed shader nodes and turned it into it into 4... with maths.

Another tick for maths. This is incredible stuff, Bill... THANK you!

You're welcome. I knew you'd like this one.

Quote - Oh, and I'm pretty sure I got it... that Add() node Value_1 drives the whole thing, right? .5 density is like 75% opacity, .3 density is like 51% opacity, which is 49% transparency? etc...

Correct. I forgot to explain that since I needed the density in two places, I pulled it out into its own node, so you'd only have to set the value once.

Quote -
You could make that a PM: Opacity node - forcing people to think in terms of those transparency channels as opacity channels.

Yep - did so already. I only built part 1 and 2 by hand. Then I went to matmatic for the next parts, and of course there are PM: nodes.