Attached Link: Full Size Gallery Image

So I finally got PP2014 and can once again run Poser on my computer!

Im straight back into skin shaders and I've been staring at this one too long!  Let me know (in as honest and brutal way possible) what is wrong with this image, from a realistic skin standpoint!

This is a custom single SSS node based shader that splits up into NON SSS, SSS, Direct Spec and Indirect Spec.

It looks good, but in this light even 2009 VSS without scatter looks good. What is going on here that requires effort? Did you do the blemishes procedurally?

The over-the-top eye reflection is distracting.

Yeah I am not sure what happened to my scene file with the latest eye shader I created, but this one is terrible.  I'll spend more time on it once I get the skin back up to scratch.

What would be a better lighting setup for feedback bagginsbill? Single light in a black room from the side?

I am trying to get the softer looking low scale SSS to render without losing the texture detail of straight diffuse while still looking "realistic" or close to.

Thankyou for the response!

Her fingernails look cyanotic. I'd check her hemoglobin levels, do a drug serum analysis, ask about a family history of heart abnormalities, and check for asthma or other respiratory issues. ;)

In other words, she looks real enough to consider inquiring about this particular abnormality as though she were an actual medical patient. Nice job.

Here is another shot with 1 light, no scene.  Dont you hate it when you forget to change background type to black!

Would like to see how you're handling your specular. I'm seeing good 'dry' skin going on, but I agree that her fingernails look too bluish. Otherwise, I'm liking what I see.

You need to lose the artistic pose and dramtic lighhting. A neutral pose with fairly flat lighting would be more suitable for initial assessment.

Here is a zero'd pose with a single inf. light at 90 degrees.  Im not sure what you meant by flat lighting?

Can you describe or link an example of the type of lighting you mean?

I increased the specular from about 40% to 100%. I actually didnt realise it was so low!  Normally I would have it at less than 100%.

--EDIT--

WHOA!  I just noticed the smooth poly bugging out the shadow accross the chest!  Dammit

Oh, I was meaning that I'm looking for a 'dryer' looking specularity than I'm getting with EZSkin. So I am interested in what I'm seeing here.

I looked closer at the image you uploaded and it looks to me like your bump or displacement is rather high. If you look closely the mesh lines are coming through your texture in the render.

The specular here is a combination of 2 microfacet nodes one coloured by a HSV modified texture map.

Thank you! I will mess around with this bit of nodes. I'm tired of my people looking like they've been working out hard or else oiled up for the beach.

attached is half displacement and smooth poly's off

This is going to be one of them threads that is ripe for knowledge pickings!

Much better render there, Templar!

I've implimented your specular node set up on the face of my current texture creation and am doing a test render. How did you adjust the specular for the lips?

Pumeco, yes, I'm hoping to learn more with this thread! I love threads like these where we learn shader tricks!

Actually the specular on the lips is the same as the skin. On my particular texture set (DAZ Elite Alana), the specular map includes much brighter specular value for the lips. If you find the lips too dry, increase the specular values of both microfacet nodes equally. Probably double

I'm actually using the component node instead of a specular map and I'm getting major shinny.  I've done some tweeks and I'm rendering, but will have to head for bed. It's 4AM here on the west coast. Have to hope this discussion continues and evolves.

attached - same scene with light moved to 45° x 45°

I also slightly reduced the specular

LatexLuv I never stop playing with materials :P

It's very hard to share specular setups involving maps - nobody (and I mean no two vendors) can seem to agree on the interpretation. Thus, we have to futz with levels after applying every texture set involving a specular map.

Then you have the additional complication of forgetting to set gamma = 1 (or forgetting to mention to others that you did) on the specular map.

If you're using the red channel setup (as many do from copying my S+B shader or using EZSkin) you also have to accept that this is gamma = 2.2 data, and will be altogether different intensity as well.

What you have to learn is not a certain level - pay no attention to tg's particular values. Rather, you must understand what the level multiplication looks like when too high or too low and be ready to adjust. You can't treat the shader as a black box with simple inputs you drop in from random vendors.

The thing you really want to discuss is

1. What is the correct roughness for the specular node you've chosen (ks_microfacet). Is that node going to behave under the conditions you intend to light the model? (Hint: If you're doing a no-lights render as I frequently do now, the ks_microfacet does NOTHING.)

2. Does the real material (in this case, skin) exhibit dual roughness as do pearls and require two nodes?

3. Does the real material exhibit colored specular reflections (as do metals)? If you think skin does, then you would use the colored input on the specular. Otherwise, it would be white.

The thing I'd most point out during the experiments is this: It is very common to add a string of nodes and think you've improved things, when they're actually doing almost nothing. I have seen many complicated shader trees that plug into black. Much discussion around the purpose and results of such a shader tree made me laugh as it was doing nothing at all.

Quote - LOL, thanks Moriador :P

 

Here is another shot with 1 light, no scene.  Dont you hate it when you forget to change background type to black!

This was the post with the white background, side lit hand over face.

I like it. A lot. It does look a little dry to me (i.e. I think the speculars are doing pretty much nothing) but then you have some who seek that, so there you go. The scatter looks good to me.

I may try some experiments and demonstrate a trick or two for specular analysis.

I'll be back.

scene with light moved to 45° x 45° consists of only one poser light, and no diffuse surfaces other than ground and girl?

 in re: bluish-greenish fingernails, I recall your skin shader from last year also had that cast on the skin, before you corrected it.  perhaps fingernails shader was from old shader.

however, it looks very good.  lower lip also bluish near corners.  in PP2014, does blue still scatter more than other colours?

I have wanted to bring up a discussion like this for months. There must be a way to do specular with the scatter setup that does not make my poser people look like they've got a grease problem or like I said above, have been oiled up to go to the beach. I guess what I'm getting at is the 'Octane' look, although that's a little too dry in my opinion.

I don't know if this will add anything to your nodal ideas, but something I started to take seriously regards Specular is something I think of as the "Tarnishing Effect"

You're all sat at a computer, the space bar gets a lot of use, it might have been dull to start with, but do you have a shiny patch where your finger has worn it over time?

The specularity of the material on your spacebar has not changed, but the finish has.  In other words it's the loss of those ultra-tiny bumps that makes it look shiny.  The "specular" of the material looks to have changed, but it hasn't, it still the same material as the untouched keys that still look dull, like new.

The only difference is, the ones that look dull still have those ultra-tiny bumps on them, and I think you might need to differentiate between a "Finish" and a Specular to get your skin absolutely spot on because that's what happens in real life.

I suppose you could say that Specular is the result of a Finish.

To play with the "finish" would involve creating ultra-tiny displacement and changing the spacing between the bumps to adjust the effect of Specularity.  So, maybe set a nice tight specularity and leave it while you play with the finish itself.

Haven't tried it in Poser but it's a very real thing, I used to play around with it a lot in Bryce many years ago, but I was limited to bump and it didn't really work that well.

That is a good point. I am not a fan of how bump nodes are handled in EZSkin. And since I don't know of a way to edit the shader set ups in EZSkin before applying it (which I don't think you can), I will still go back and use BB's handy dandy VSS prop with experimental skin setups. The bump problem is contributing to the too shiny look I'm seeing.

Quote - scene with light moved to 45° x 45° consists of only one poser light, and no diffuse surfaces other than ground and girl?

 in re: bluish-greenish fingernails, I recall your skin shader from last year also had that cast on the skin, before you corrected it.  perhaps fingernails shader was from old shader.

however, it looks very good.  lower lip also bluish near corners.  in PP2014, does blue still scatter more than other colours?

That image has a black sphere encompassing the scene to catch and contain any stray light rays, but as its black, it should not effect the scene.

With regards to the fingernails, it seems that I lost the latest version of that scene somehow, including the really fantastic eye shader and better nails.  To be honest, i have not actually done any work on the nails in these images, they just use the skin shader with sharper specular and less SSS.

I am unsure if that is fixed, but from my experience there seems to be little to no difference in PP2014

Quote - I have wanted to bring up a discussion like this for months. There must be a way to do specular with the scatter setup that does not make my poser people look like they've got a grease problem or like I said above, have been oiled up to go to the beach. I guess what I'm getting at is the 'Octane' look, although that's a little too dry in my opinion.

This is the reason that I keep at my own skin shaders!  When I look at my skin, compared to what I can create in Poser (or use in Poser), its the specular that always look nothing like my skin.  One day!

Please note the too shiny lines of light on her arms, hip and leg.

I think it looks really good but slightly waxy :-)

I don't know if it's an easy thing to do in Poser, but if you could mix an ulta-fine displacement with the settings you've shown here, I think you'd see what ultra-fine displacement does for specular.  The only problem would be for localised detail, unless maybe there is a way to control the density distribution of a noise shader by a map.

Without some way to control that, what effectively is the finish map, would be uniform.

Here I have taken my specular map, inverted it, the used it to control the max of a noise node, finally the noise and bump are adjusted to -05 <> 0.5 and then compared against each other and the max between the two is used for bump mapping.

Here is the shader (actually the complete skin shader) with the "finish" added and a comparison to without the noise and with the noise

The color ramp node is closed, can't see the settings. What are your Poser units? I use Inches like BB.

The color node has the top color at 200,200,200 and the others at completely black. Unfortunately my laptop ran out of power, so I can't screenshot for you.

I use actual "poser units" or the default setting. I don't think it has a real world representation

templar, I really like your new shader.  poser units correspond to real-world units in inches; see posts by baggins, geep et al..  

what are your plans for this new shader? e.g. do you desire it to be included in PP2016 release (no date yet)?

I have been trying snarly's pose-2-lux implementation for OS X.  shaders are rather simplified by using only diffuse, specular, transparent, ambient and translucent (classic poser 4 channels).  should you desire to render in luxrender, I would enjoy seeing it.  let it run for a few hours.

**
@Templar**
The one on the right looks less waxy although you didn't say which is which, and the math nodes just go right over my head right now.

If the one on the right is the one we're seeing in the nodes, I'd say there's some success even on the first try, but if it's the one on the left then maybe the bumps aren't the correct tightness together or the correct height.

Basically, that's what I'm getting at here; the rougher the finish something has, the more diffused or softer the specular highlight becomes.  Think of frosted glass compared to shiny glass, the difference is the frosted glass looks dull because of the ultra-tiny bumps, but polish it up and you get normal glass again.  In other words, it's a matter of playing with nodes until you are able to alter the hardness of the specular through playing with bump but without adjusting the specualr (for the most part, anyway).

I used to test it on a sphere and cube with just one light in the scene, no distractions.

Pumeco - the micro bump you're describing is too micro to do with Poser bump, and it is why the various specular nodes have various ways of describing microscopic roughness, which they then model in a statistical sense.

Blinn, Anisotropic, and KS_Microfacet all do this.

If we were to, instead, model with explicit microfacets, each pixel would have to be sampled around 1000 times.

For actual reflections, the softness parameter also does this, although in a rather hamfisted way.

Ah, cheers Bill, so that's what the microfacet thing is; I did wonder but didn't get very far!
What does the "KS" part actuallty stand for?

Quote -
Ah, cheers Bill, so that's what the microfacet thing is; I did wonder but didn't get very far!
What does the "KS" part actuallty stand for?

I don't know. I don't recall SM telling us. I assume those are letters denoting the authors of that BRDF model. If I had to guess, it would be:

K = Csaba Kelemen

S = László Szirmay-Kalos

http://sirkan.iit.bme.hu/~szirmay/scook.pdf

So this thread caused me to investigate the ks_microfacet node and the "Subsurface Skin" node. Since I learned quite a few things in the first few hours of investigation that I didn't know before, I'm going to assume I'll learn some more and so I won't talk about those yet. Otherwise I may end up contradicting myself.

But - meanwhile, I did some demos using my well understood good-old Blinn node and the Reflect node. I've been working with these a long while and I can show good results that are similar to the kind of thing you're used to seeing with better renderers like LuxRender.

I want to point out that if you're really after realism, you need to increase your concern for reflections, rather than endlessly tweaking the specular node(s). The greater part of accurate speculars comes not from lights, but from reaction to geometry.

I have a little setup here to show you.

Now keep in mind that Specular, Blinn, ks_microfacet, anisotropic - all of these react to the point light and NOTHING else.

Meanwhile the Reflect node reacts to the glow box and all the other props lit by that box.

Here, we're under the point light.

The light is between the rows, so the back row shows us speculars that are close to the normals, and the front shows speculars that are way off the normals (shallow angle of incidence).

This render shows only Blinn nodes. Left to right they go from being very sharp and bright (low roughness) to spread out and dark (high roughness).

This is traditional Poser shading. This is what most people deal with and stop.

It's like a whole new image. Do whatever you can to make the shaders react to lights as best you can (tweak the speculars) but if you ignore the reflections, you've already failed before you started. Lights are less and less important as you get closer to your goal of realism. Therefore the shader parts that react to those lights are less and less important as well.

In such cases, with zero lights in the scene, your carefully crafted specular nodes DO ABSOLUTELY NOTHING. Utter waste of time.

In this image, the Reflect node is doing it all.

I still include Blinn in my shaders, because I already worked out how to set them long ago, and it's nothing for me to have them set up every time because I use matmatic.

If I was assembling high-quality shaders by hand, I'd just skip the specular nodes altogether.

Look at the reflected shape of the glowbox. No Poser light (which has no dimension) will ever produce that shape.

Go here

http://gl.ict.usc.edu/Data/HighResProbes/

and get the HDR for "Dining room of the Ennis-Brown House, Los Angeles, California ".

Render with that a bit and tell me how Blinn or ks_microfacet could do anything but mess up your results (if you were to try adding point lights) or actually do nothing at all (if you left all lighting to the HDR).

Cheers Bill, and that PDF just scared the crap out of me :-D

I took one look at that and thought, bloody hell, I wish we'd been taught that stuff in school instead of pointless things like history.  I never did get the logic behind that, I reckon I'd be a nifty programmer these days if my youth wasn't wasted like that.

Grrrrrrrrr!!!

**
**

PS: Sorry, didn't see that new stuff when I posted, looks real nice, will try to play!

The Ennis HDR is such high contrast that I still get grief in Poser with it.

I have seen Poser improve its blurred reflections but they're still a little grainy when dealing with Ennis-type lighting. Look at the right-most black pawn - the blurred window is ugly-grainy.

Then we get the bogus diffuse part, that just drives me up a wall. I had pretty high samples and IC here but not even close to good enough on the white pawns.

Puts hand in the air

Mr Baggins, sir, on that fourth image you said you added "matching relfections".
What do you mean by "matching", is it a mix in the nodes matched to something else?

IMO SM did give us an alternative by providing the add-on structure though - Reality to LuxRender, and Octane plug-in both use this so it is possible to set up a scene in Poser and use an unbiased render engine relatively easily.

Having said that I'd also like to see more work done on Firefly; as usual it'll come down to development time and priorities. More speed and accuracy with IDL would be nice ...

Quote -
Puts hand in the air

Mr Baggins, sir, on that fourth image you said you added "matching relfections".
What do you mean by "matching", is it a mix in the nodes matched to something else?

I mean that I did many experiments that compared

1. A point light revealed via Blinn

2. A small glowing ball revealed by Reflect

And I matched the intensity and size of the resulting highlights so that I could just dial in a certain amount of roughness and both nodes would react in lock step and produce pretty much the same highlight whether I used a ball or a point light.

Got it, cheers Bill.

Got my student robes on, BB! I messed around with node setups last night from the early discussions of scatter over at RDNA, using VSS to apply them. I found that I liked the look of Blackhearted's Anastasia shader the most in my tests. That shader uses reflect and fresnel edge blend that luckily isn't hard on the render time.

P.S. I just got Poser 10 3 days ago. But did you know that in the advanced material room you can now double click o the top of a node and change the name of it? I know that's such a little thing but for me it's huge. I have held onto a little python script for renaming nodes. Now I can just double click on the node and rename it. Wow!

Quote - **
@Templar**
The one on the right looks less waxy although you didn't say which is which, and the math nodes just go right over my head right now.

If the one on the right is the one we're seeing in the nodes, I'd say there's some success even on the first try, but if it's the one on the left then maybe the bumps aren't the correct tightness together or the correct height.

Basically, that's what I'm getting at here; the rougher the finish something has, the more diffused or softer the specular highlight becomes.  Think of frosted glass compared to shiny glass, the difference is the frosted glass looks dull because of the ultra-tiny bumps, but polish it up and you get normal glass again.  In other words, it's a matter of playing with nodes until you are able to alter the hardness of the specular through playing with bump but without adjusting the specualr (for the most part, anyway).

I used to test it on a sphere and cube with just one light in the scene, no distractions.

The image on the left is without "finish" and the image on the right is with.

I am really liking the look of the shader with this noise added to it, you are correct that it looks less waxy and more natural.  I am just tweaking some nodes now to try and improve specular, which was dulled significantly by the noise.

Baggins, I am unsure how feasible using the reflect node for skin based specularity is, I just gave it a quick try, and you need some serious render settings to get anything approaching solid specular.   And I think most people are not going to go down the route of lightless rendering, at least for now.

Thanks for the renders of different blinn and reflect levels.  They are enlightning!

I am sure the KS_Microfacet node works with infinite lights also