So. I need to figure out what's changed in Superfly that I need to adapt for my future products, and what's bugging that I need to report and wait for them to fix.

Case in point: rendered in P11

Settings:

(rendered on CPU because P11 isn't compatible with my GPU)

P12 render to follow.

P12 render is cooking, and puzzling me to no end - from what I can see so far, some of the visual oddities I was getting consistently don't seem to have happened now. Huh.

Will post when the render is crisp enough - it's going on progressive refinement, not rendered on background. I'll stop it when it's not much noisy anymore.

Adaptive sampling works best with progressive refinement unchecked. And it REALLY does speed render times up.

ghostship2 posted at 4:58PM Fri, 25 December 2020 - #4408535

Adaptive sampling works best with progressive refinement unchecked. And it REALLY does speed render times up.

The problem with that is figuring out how many samples I need xD I'm bad at that. Right now I had to start my P12 render over because it finished rendering while still very grainy - with the same 60 samples that in P11 I could stop the render in less than midway and had the result above.

Edit: did they fix the background image bug? It seems they did, which makes me a happy camper - thank you to whoever was responsible for that fix!

Okay so I'm impatient so I stopped the render with it still grainy. P12:

Settings:

Huh - nope, background bug is still present. I guess the image I was using was just too similar to a plain gray so it looks decent in the render anyway.

Other thoughts:

-It seems like P12 is having a hard time understanding the sudden change in the gloss of the lips material. I hope that's a bug and not a feature - kinda need to be able to define different highlights for lipstick options.

-This skin was rendering with a lot of yellow areas just until yesterday, and I didn't update after that. Not sure why it's not doing that now. Wonder if it's because there's a background image set to be a light. But from what I was trying yesterday, it only stopped being yellow after I unplugged the textures that this character has plugged into the PhysicalSurface Scatters - Material looks like this:

-I have NO idea what's going on this the hair there. Looks so much darker in P12, and then there are stripe-things at what seems to be each polygon down the hair cards. I can't even try to figure out what's going on, as my hairstyles use bagginsbill's hair material, which is a node noodle that I can't even begin to understand and SOME OF IT looks like this:

I could make materials specific for P12 Superfly with what I know of Cycles now that we have the PrincipledBSDF and PrincipledHairBSDF nodes but 1. would be a bother to have separate materials for P11 and P12 for all my hair products and 2. CyclesRoot doesn't show you a decent preview, so I'm kinda lowkey hoping that whatever's going on there is a bug and not a change in how Superfly works.

I tried without the Progressive Refinement with 20 samples and it looks crisp enough, and really fast - thank you!

Alright, so. To test, I applied my character's regular lip color, to avoid that change in glossiness, and deleted the Background material I had. Lo and behold:

The eyebrow bug is back! And the yellowed skin (look at the lips - they're very rosey pink in P11). Eyes looking sickeningly red! Ouch.

Eye shader:

I think I read somewhere that P12 uses lower numbers in scatter? If that's true, that must be what's happening to the eye there. But the yellowing in the rest of the skin is not just low numbers - that keeps happening in very low numbers unless I unplug the SeparateRGB+Scatter map from ScatterDist there, I've checked this a lot.

Also, that one streak of lighter hair on top of her hair? I suspect that's a hair card that has its normals pointing down, so it's rendering the backside differently. Shouldn't happen - this would break longer hair even if we make hair with all cards aligned. Buuut might be something something in the node noodle from that hair material, again, I don't know. Will play around with another attempt of hair shader just to test.

I loaded a Background picture again just so that the eyebrow bug and red eyes wouldn't distract me while I attempt to make a hair shader using PrincipledBSDF - but the bugs are still there. Huh. Was it only away when I loaded it directly from the file saved in P11? Why so finicky, P12?

Yikes. All this makes me want to stick with Firefly.

I wonder how Snarly's doing with the new version of EZskin.

#1 The eyebrow issue with LaFemme is a known bug and is being looked at with high priority by the devs. #2 glossiness of nodes is something that changed with blender to get a wider range of roughness/glossiness out of those nodes. See my post on the glossy node. #3 the image size bug is noted and being looked at by devs.

I'll add that I messed about with LaFemme and found the same issues you had so you aren't crazy. With the default textures and shaders she rendered just fine. After altering her SSS shaders to the ones I use I got the crazy eyebrow glow. no way to get rid of it. Yes, the default shaders make her look a bit jaundice.

Shifting the radius of the SSS down eliminates the eyebrow glow. I shifted it down by two decimal places. I'll try with a different shader setup.

Any ideas about what might be going on with the hair though? As I'm focusing my sales on hair, I gotta know what to worry about ?

Looking at the skin shader that you are using for P12, I would not use a red subdermal map for subsurface scatter. The SeparateRGB node is separating your image into 3 basic channels and is being fed into corresponding connectors, you are feeding a predominant red image into each channel that, I believe, is giving the wrong color information. I would use a regular base skin texture image without makeup or tattoos as your donor for RGB information. As for the scatter distance mm value for each RGB channel, I recommend using R=5, G=4.5, B=0.9; these values represent the amount of skin penetration for each channel that is close to real physical conditions. This may also help with the self glow. If you are using any SSS radius, the number I found that work is 1,0.2,0.1.

If the hair shader you are using is from Bagginsbill EZSkin3, don't use it for P12. There are error connections to the far right that impede how hair appears in renders. I shifted entirely to ghostman hair shaders, ATM for better renders.

Don't use SSS for eyes, it is entirely unnecessary.

Just investigating the skin thing at the moment. glowing eyebrows gone with change of SSS radius. Using the Cycles SSS shader setup by Bopper on Share CG which I use for all my figures. Looks nicer and less yellow. You can see the default shaders on the arms and the Cycles shaders on the head and torso.

Two things here, just form a pure user side :

1. For skin shaders, until EZSkin is able to function with a P12 dpecific shader, most of as don't have the means to painstakingly alter every single shader we used before P12 to match what was changed. A lot of users will simply not use P12 until this is fixed. I adore the speed of rendering in Superfly, but I can't spare the time it needs to manually alter all shaders used. Deride us if you want, but end users really just want to use what they already paid for.

2. As to your hair, Ohki: I mentioned in a few months back that there are some hairs that one currently can't use in P12 because, even if it's sold recently, they render less than optimally in P12. I can't afford dropping money on products that don't work in the latest version of the program, even though I frankly adore your hairs. I hope this can also be fixed and at least a simple shader how-to can be posted in a place it's easy to find/access.

I'm not bashing anyone. I understand that progress had to be made and I think P12 is totally a step in the right direction. I understand this is an early release, I filed bugs myself. I hope the most glaring ones will be fixed optimally soon and that EZSkin Python 3 can be released equally soon for those of us who don't do programming.

Vendors can update their products with P12-specific materials as poses/materials in the library, that should make it pretty easy for users to make them work in P12 (and still work in P11 too) without even needing to play with the nodes themselves.

Of course that can't happen until the vendors know what needs to be changed. And personally, if I still had a product in the store (which I don't), I wouldn't update anything until P12 stabilized. You don't want to go to the effort of updating your product and then have it not work in the next release a week or two later.

Regarding the hair shader posted above, a bunch of issues have arisen with glossy, anisotropic, etc, so I would look at whatever is plugged into alternate specular. It's not visible in the window though.

CyclesRoot doesn't show you a decent preview

It's often all white for me. Is that just me? It's pretty hard to pose a figure correctly when it's all white, especially getting the eyes right. I don't know whether this is inevitable but it does make me want to do more with PhysicalSurface and not Cycles.

I recognize that hair shader as a BB construct and it works beautifully in P11 but in P12 it renders incorrectly. I traced back the problem to a couple of node connections but since BB constructs are usually far from simple, it is virtually impossible to unravel the connectors and figure out which one of the nodes are trouble spots. I say look for a simpler hair shader that will work in both P11 and P12.

But I do agree that until P12 reaches a point where all the problems are corrected, it's better to find some quick alternatives to keep the software afloat.

It's often all white for me. Is that just me? It's pretty hard to pose a figure correctly when it's all white, especially getting the eyes right. I don't know whether this is inevitable but it does make me want to do more with PhysicalSurface and not Cycles.

No, it's not just you. I find that quite difficult to deal with. I wonder if anything can be done about it.

-On the skin: that skin shader came from an entire thread where BB helped me come up with something for the Physical Surface that worked with a subsurf map - in the thread, Poser bugs were found and workarounds were figured out, and things were explained like how the Scatter values depend on the diffuse map colors too. I'll have to test if this is no longer true for P12. But if it still happens in P12, the values hborre gave there won't help, as they would be specific for each skin map.

-Rhia I'm fully expecting to have to issue out updates for my hair styles (and for Mayu, which is the character you see here) with extra mats for P12 Superfly. BUT:

-I'll only do that once P12 is completely stable at least when it comes to materials and rendering, for the reasons others have mentioned here

-I'm avoiding to use Cycles root and nodes specifically because they preview all white, which is a bummer because I spent this entire year in blender and got quite comfortable with Cycles nodes meanwhile. But from what I understand, the Eevee realtime renderer has different terms of use than Cycles so we probably can't ever hope for nice previews of cycles nodes anyway because they can't add Eevee to Poser. Which makes me wanna stick to PoserSurface and PhysicalSurface.

-Looks like I'll have to open that thread about the PhysSurf skin again and go through BB's steps to check how the RGB SubsurfDist is working now.

-Where even is BB these days, btw? ?

-... What ghostman hair shaders? ? Where would one acquire them, and can they be used and distributed in commerciall products? I'm still very unknowledgeable in mathy things and can't think in node noodle language.

BB is around. I saw him posting in EClark's "Memba Deese?" nostalgia thread not long ago. (Remembering old toys.)

randym77 posted at 5:55AM Sat, 26 December 2020 - #4408590

BB is around. I saw him posting in EClark's "Memba Deese?" nostalgia thread not long ago. (Remembering old toys.)

-ooh - good to know! Though for now I don't think I need to pester him, unless he's willing to fix his hair shader for Poser LMAO

ok I'm out of bed. Experience time! Will work with the lessons seen in this thread: https://www.renderosity.com/rr/mod/forumpro/?thread_id=2935202

Except this time things should be a lot faster - thanks P12, Optix and my new GPU. Experiments to follow...

So far so good. Scatter values seem to get stronger for P12 but not by much.

Eyebrows bugging even at full transparency with nothing there xD I hope they'll fix this soon.

(I'm ignoring the lips, to make things easier for me) Ok, this looks like good enough values for scatter without a skin map.

Then I add the skin map:

Yup, seems like the diffuse map does still affect the scatter values/color/whatever there.

These values seem good enough to me.

Now I plug in my SSSMap...

I like what it does to the eyes and nose. I hate what it does to the ear. WTF, is the eyebrow bug gone???

Hello I'm Ohki and I'm confuse. xD

I realize that my main problem with the scatter in this was yellowed lips. So I'm including the lips material now.

Without my SSSMap:

... the lips are yellowed anyway. It wasn't yellowing before, when I unplugged the SSSMap. Poser, why do you do this to me. You seem completely inconsistent.

With the map:

Lip looks the same to me now? Huh.

I'm gonna try these without the diffuse map, to try to see what the SSSMap is doing by itself.

Without:

With:

(Why are the lips more yellow and the ears less yellow and the eyes correct when they're all purplish in the map...)

With, same numerical values:

With, but now the SSSMap is an oversaturated version of the skin map:

... I think it doesn't like purple. Which makes me wonder if this is a new&revamped version of the "won't render blue scatter" bug found in the thread I linked above. Let's see what happens if I try to scatter blue.

Works. Purple?

... Poser, sir, what are you doing there?

Judging on the other scatter colors there, I don't think it should be rendering the opposite color of my scatter where it's front-lit, huh? Should I report a bug?

I saved that with the purple scatter and rendered it in P11:

Renders correctly. The green where front-lit with purple scatter is a problem introduced in P12.

So, perhaps it's a problem with specific render options?

-I had rendered this with Optix, so now I tried rendering with GPU non-Optix. Same render settings otherwise.

Same thing. Perhaps it's a GPU render thing?

Apart from takin ten times as long, didn't seem to change much. Adaptive Sampling perhaps?

Same diff. Perhaps the problem is with adaptive sampling AND gpu rendering?

Nah. I take it it's a P12 bug.

So, I can't really do anything about that, so I'm moving on. Will check the other stuff first so I know what to report as bug later.

Moving on to hair.

Same lighting, I changed my figure to Anuli because she uses a Cycles Root and seems to not have those eyebrow and scatter bugs. And then I loaded my Juliana Hair because it has more clear, straight hair cards that will probably be easier for me to see what's happening - all my hair styles use the same materials anyway.

In P11, in all of (or part of) its noodle glory:

(anyone tired of my image spam yet? I just want to make sure I'm documenting and testing these correctly.)

And in P12, with so much difference going on I don't know where to start.

Someone mentioned something about speculars. So I'm unplugging Alternate Specular there to see what happens.

I can't see any difference. Seems like P12 can't read that at all.

... And I'm utterly failing to make a hair material with the Physical Surface. Specular is too limited. Sigh Cycles root it is.

The one thing you haven't tried with the subsurface scatter is to plug the Diffuse map into both color and SeparateRGB and render with the values you felt comfortable with. That is the way I currently set up SSS for some of my models. Soon I will be looking at darker skin textures to determine what works best for Afro characters. I will try to post the screencap of ghostman's hair shader, it appeared in a separate post but I am unable to find it ATM. However, I did create a custom material MT5 in my Library so it's easier for me to load the nodes into a blank model. The current BB hair shaders for P12 don't work well anymore. I'll screencap that also not that it will mean much to you.

hborre posted at 9:09AM Sat, 26 December 2020 - #4408609

The one thing you haven't tried with the subsurface scatter is to plug the Diffuse map into both color and SeparateRGB and render with the values you felt comfortable with. That is the way I currently set up SSS for some of my models.

Would be a workaround, but they need to fix the bug with the purple scatter.

@ Ohki it was pointed out to me late last night (right before bedtime) that the SSS issue with glowing eyebrows only happens in Optix renders. I haven't been having any issues with my renders of V4 and V3 so I did not fully understand the issues. It IS a bug and it IS being looked at by the devs. Sorry to have led you down that rabbit hole.

ghostship2 posted at 9:12AM Sat, 26 December 2020 - #4408612

@ Ohki it was pointed out to me late last night (right before bedtime) that the SSS issue with glowing eyebrows only happens in Optix renders. I haven't been having any issues with my renders of V4 and V3 so I did not fully understand the issues. It IS a bug and it IS being looked at by the devs. Sorry to have led you down that rabbit hole.

Pretty sure I've had it without Optix too. See last page - you can notice the eyebrows even in some CPU renders of mine, in those attempts of scatter options I did, even though they're set to be fully transparent there.

But if it's a bug being looked at, I'll ignore it in my tests, I'm patient and can wait for the fix!

I don't have an Optix card and render exclusively CPU. I had seen the eyebrow problem with V4 and LF before and solved it but, unfortunately, I don't remember the process of how I solved it. I believe it's associated with the proximity of SSS material causing self glow and value settings. I found that if certain values are set too high for SSS, depending on the type of PhysicalSurface used for the model, the greater the chances that any skin-to-skin closeness will cause some glowing.

BTW, BB's absence from the P12 forum is due to not owning a copy of P12 and he can only render CPU. That limits his capacity to troubleshoot.

you can use my hair shader if you like. Just give me some credit for it. It seems to work well.

As promised, I have posted the Ghostman Hair shader that works in P12. This is the workaround for the wonky Anisotropic connection in the PrincipledHairBsdf node. Please note, I created two compound nodes to clean up the Material Room workspace for easy access. The Anisotropic/Tangent node was created just to protect the Tangent node settings.

I'm getting a little confused, I'm giving credit to the wrong person. My apologies to you ghostship2, this is your hair shader I am posting, LOL.

I can confirm the eyebrow bug is alive and well with very old GTX cards as well, guys. I posted somewhere weeks ago about it. For us not inside the node cults: should another bug report be filed or is this already being worked on please?

I think that a report has already been submitted but I'm not quite sure that it is a bug. I have test three different sets of cycle nodes, the PhysicalSurface, the PrincipledBsdf, and Ghostship's Uber 3, and fine-tuned the SSS settings for each. For every instance, I found that each node required very low SSS values, anything that is considered too high causes glowing. And some of those values are based on posted Blender tutorials that support what I am seeing.

Thank you for the answer, hborre. Out of curiosity: what is that value in your experiments? And: here is the question then. Will EZSkin for Python 3 fix this by including an 'adjust SSS value for fixing eyebrow glow'? Or will this need to be an entirely manually adjusted function?

So many questions!

@ghostship2 thank you muchly! I'll test this in my hair props - can I include it in my comercial products as long as I credit you in the readme for them?

I don't know which route Snarly will take when EZSkin is finally released. There are 3 flavors of Cycles nodes to choose from. Straight cycle nodes would be more complex and wouldn't help individuals that don't have a Material Room background. PhysicalSurface is better and can be used in both P11 and P12 and easy to work with. PrincipledBsdf is exclusive to P12 and is also to work with. I can give examples of what I have put together that works for me.

hborre, that would be much appreciated and thank you. I now mostly work in Superfly given how the render times improved dramatically with P12, it it would be a great help.

@ghostship2 :

May I ask what the Trans Mask is in your setup please? I'm familiar with the rest of the 'sghetti otherwise but that threw me.

@Rhia474 it's the transmap that comes with whatever hair you apply the shader to. Same with the texture maps and bump maps. It still uses those from your hair. I personally use color and bump maps from OutOfTouch but this one is using the color and bump maps from the Juno hair itself. Biscuits and OutOfTouch make fantastic maps with no shadows or highlights. This business with the glow thing is interesting one of the vendors here is on the beta team and told me that it WAS and issue with Optix but you and others now have stated that it is popping up with older cards without Optix. Sounds like a change of Cycles node standards to me just like the glossy node issue.

@Ohki Yes go ahead and use my shaders in your commercial stuff, just give me credit for them! What the hell am I going to do with them otherwise? lol

@ghostship2, thank you, off to futz. Yes, those hair maps are great, and 3Dream also put out some Superfly mats for their hair that they added as freebies with no highlights as well.

@Ohki: thank you for working on this, I love your work and would gobble up all your hairs at once if I had the funds right now.

I going to start with the PrincipledBsdf node in P12. I came across a Blender tutorial based on its PrincipledBsdf and each of the channels. I will attempt to explain the additional nodes as I go along. Note, I crossed out 2 nodes that I typically connect to the Specular channel on the PrincipledBsdf node, it is an IOR to Specular converter based on a Blender tutorial and specific to the dielectric material.

Whenever I use a character diffuse map, I try to find a comparable base map that contains no makeup or tattoos. In this particular case, I am using a character map from V4 and her base map that was originally included with her release ages ago. I use the base map as a subdermal texture map that is connected to a CombineRGB, red channel, and eventually connected to the PrincipleBsdf Subsurface Color. The reason I use the CombineRGB is to separate the RGB but yet have some control over what color is expressed by the Subsurface Color. I played with various combinations by changing numbers in the green and blue channels. Use your imagination. If I make a Star Trek Vulcan character, I can recreate green subsurface scatter because those aliens have copper as a hemoglobin base as opposed to humans that have iron as a base component. For a bump map, again I use an unblemished base map. I lifted the bump map nodes from EZSkin3 and bundled them into a compound node. If you like to see that, let me know.