Daz 3D is part of 
Connect
DAZ Productions, Inc.
224 S 200 W, Suite #250
Salt Lake City, UT 84101
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2024 Daz Productions Inc. All Rights Reserved.Daz 3D is part of
Connect
DAZ Productions, Inc.
224 S 200 W, Suite #250
Salt Lake City, UT 84101
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2024 Daz Productions Inc. All Rights Reserved.
Comments
So, I'm trying this again.
I have my SSS Color set as above for the SMD of .01, and the SMD set to .01.
I picked a Transmitted Color from the texture map as suggested, and set the TMD to 1
The skin is littered with chromatic aberration. With no dome set On, its black... with dome set On its bright.
What am I missing? Is it the Translucency? Doesn't seem to be. Tried lots of options there. Refraction weight is 0.0.
.1 SMD works...
Which seems counterintuitive to me. Isn't an SMD of .1 greater than an SMD of .01? Why would .1 be more "solid" than .01?
I think this is a known issue. You might try using Mono instead of Chromatic. I wish I could be more help. I know a lot of people have had similar issues with the Chromatic setting not playing nice unless you have the settings at specific levels and colors.
turning Draw Dome on will fix itt oo, at the expense of having a background to the render. It is, I think, fixed in the 4.11 Public beta
The idea is to find physically plausible settings that work, so... changing to Mono defeats the purpose of the experiment. As noted, turning Dome "On" does not work, it just changes the abberation from dark to light. It does seem to be corrected in the beta, which is good, but the beta is wildly unstable.
In the current beta there's an "Invert Transmission Normals" option added to Uber Shader. By default it's on, but I'm finding my renders look much better with it turned off for skin surfaces. It looks like it affects how SSS interacts with bump and normal mapping.
I believe this is actually a correction to previous behavior, if I remember reading the release notes properly. "On" is actually "Correct Behavior" and "Off" is physically inaccurate. IF, as I said, I remember what I read properly.
That's how I read it too at first. But it could also mean that they fixed the shader so that it stops inverting the transmission normal, and then added the property to reproduce the original buggy behavior, in case there are some users out there relying on it. In which case you would want it off for the physically correct behavior.
Here's a couple more renders I did... I'm using a tiling normal map to add the skin pores here. In the first one, with "Invert Transmission Normal" turned on, the highlights are broken up more and the surface of the skin looks hard like it's made out of clay or plastic. Turned off, there is more scattering inside the pores and the surface looks smoother. The lip highlights are what really convince me. They always look correct when this property is off, no matter what lighting I use. But when it's on they usually look pretty strange (though they're not too bad here), and I have to fiddle with the Top Coat settings a lot to get them to look semi-decent, and if I change the lighting I usually have to adjust it again.
The invert Normals On is correct.
@arnold c, this formula Scattering Coefficient = log(Scattering Color) / (-1 * Scattering Measurement Distance) requires a Base for the Log function. Not being much of a maths person, I have no idea what that is.
Little help?
I return, for a question regarding sss directon. for physically accurate, is it supposed to be positive or negative? I remember mec4D posting some early settings saying -0.5 was correct, but then sometimes I see people use 0.5?
I have no idea which is physically more accurate, but a good rule of thumb is just to do some comparison renders and see which one LOOKS more realistic and natural. :) The end results, imo, really matter more than trying to be perfectly physically accurate (though they do go hand in hand frequently). :)
OK I NEED Y'ALLS EYEBALLS. which looks more "right"
These are octane renders btw.
Well to my eye both of the -0.5 look more correct because I'm guessing the the dude it supposed to be pale and the -0.5 is paler. Why, I don't know, maybe the parameter being set to -0.5 makes the skin behave as if it was more translucent and so paler. However, none of them I can say are more correct than others because I don't know what the dude used as a real model's source photographs look like, if there was a 'real dude' used at all.
oh I forgot to mention the texture is Owen 8 lol.
The -0.5, on the first one, because for that amount of light you would expect some glow through the skin. I do a lot of photography and that seems to be more what I would expect out of camera. For that amount of darkness to occur on his shadow side, there would have to be some light blocker in the scene (like a black panel) on the 0.5. Now it is possible to get both effects out of camera with a wider/narrower aperture but in general I would expect the first one (not knowing anything about your scene set-up)
I can't guess. They all look reasonable to me. V3Digitimes used various positive and negative (if I remember correctly) with her physically based settings in the Iray Advanced Skin Shader (?) packages for Genes 2 and 3. That leads me to believe that under certain circumstances it could be either. And I guess that would depend on the other settings. In general, I prefer the results of the positive scatter direction in your images.
Sorel,
You would think that this question has an easy answer. Having fussed about withi this very same parameter, in Octane no less, I've come to a few conclusions recently.
Rather than to think of which one is more correct, maybe better to think about what each setting does to the final skin.
-0.5
This setting means that much more of the incoming light is allowed to reflect back to the camera, rather than being absorbed or scattered within the skin surface making the skin seem "lighter." If we assume that a seeting -1.0 means that all light is bounced back toward the camera and none is absorbed or scattered, this leads us to a material that behaves in a rather diffuse sort of manner. The result is a more "solid" appearance, perhaps even "harder" as well. Most often we gauge the correct lightness of a skin tone by the way it appearsa next to other surfaces, such as a wood table or a white tshirt. I find that the negative settings means that I can use less light intensity one that fits nicely with the assumed values of the other surfaces in the scene
0.5
The positive setting means that more light is absorbed and scattered than is sent back toward the camera. For this reason the look will seem darker, but the results will seem more "softened" and less solid and more obvious in its translucency. some have used the term "sallow" to describe this. A positive setting will make the skin sonewhat harder to light, in that the brightnessd response of the skin is subdued by the absorption.
The first problem is getting people to agree on the proper amount of translucency.
Personally I find -0.5 to be a little too hard and similar to diffuse for my tastes, however though I like the look of +0.5 I find it sometimes disagrees with the responses of the other surfaces in the scene under standard easy lighting. My current setting of choice is -0.25, because while I like the brighter response of the negative values, I feel that the correct amount of absorption is probablgy negative.
You know, you could always just go with a value of 0 for scattering indirection, and I'm certain you will be happy with the results.
Like Evildead observes, they all look correct, generally, and probably wou;dnt matter much which way you decide to go.
I'm going to upload some recent renders soon to demonstrate more of what I am talking about. Fun fun as always!!
Well now that you have said the skin is Owen 8 they's no doubt that the -0.5 is the value that works as intended with DAZ Studio & their DO iRay Skin material setup for Genesis 8. They all look OK but I have enough experience with DS & Genesis iRay skin materials I've purchased from DAZ 3D to know that the skin isn't lloking reflective and translucent as it should be with pale skins. In real life, the paler the skin, the more translucentcy is obvious because things such as viens and blood vessels will be but the 'DAZ' way of doing it makes reds & oranges overwhelm pale skin so one if forced to dial the tranlucency/transmitted down.
Nonesuch,
The vert first thing I'll be doing with Iray is figuring out a better way to do the SSS than what we have now. Not that what we have inst great, but I'm convinced we can go further.
Sorel,
The problem is that its very difficult to truly discern just how light or dark a skin tone might be. So much depends on the lighting, sin ce human skin does absorb some light where other surfaces int he room might not.
What I am saying is there are varying degrees of whiteness, and blackness or whaterness to skin tones. So really just how light or dark your renders should appear is hard to say, both can be correct.
As promised here are some recent examples of what I am doing. Scattering direction is -0.25 in all cases. There are medium and dark skin tones. Please forgive the lifeless expressions and posing. I didnt see the use in spending hours tweaking a pose when the study at this time is merely the skin surfacing.
All rendered in Octane 4. Genesis 1
First Image and second image use the same texture, however the roughness is slightly greater in the first Image which creates a nice sheen on the skin. However the second image closeup I decided to lower the roughness.
You will notice that Second Image and Thrid image are the exact same morph, but with a medium skin tone. For crazy fun sake I applied that exact same medium skin texture to a female.
I'm very happy with my current hair shader, and with Carrara's implementation of hair.
So how am I doing in terms of realism? Feedback is greatly appreciated.
Thanks for your time.
I'd say you are doing pretty well. I'll see how I feel about that -.25 with my skins. I think your models need some specular work. The skin seems nice, but... dry. The second fellow, his lips in particular, I think are more "real" than the rest. Base skin is good. But something is missing (its hard to ignore those dead stares).
Evilded777,
Thanks for looking at the images! The specular you mention is actually something I was hoping to discuss. I have been looking at people under typical indoor lighting situations and have come to realize that most often they do not display too many sharp or crispness in the specular reflections, skin seems more matte generally, provided the light sources are broad shaped, like ceiling lights that are several feet long and are several inches wide. The specular is very different when the model is placed close to a small light source, like a bulb roughly the size of the palm of a human hand. As you will observe with the third image, he is placed within the room in a place where he is not in the direct path of any of the ceiling light sources. All of his light seems to be from indrect bounces within the room. The young lady wearing the same texture is indeed directly beneath one of those ceiling lights, but as is the roughness setting is a bit higher and that is why she gets that sort of sheen effect around her nose. I've seen humans with rougher skin that this actually happens in real life. I've also seen people with skin so smooth especially in certain areas where the specular can seem almost glasslike and more pronounced. So maybe I need to use a map to specify the roughness in different areas of the face instead of using a global parameter setting. I've also noticed that human foreheads and tips of the nose tend to be much shinier than other parts of the face, and that the face ov erall is sually about double the shininess of the rest of the body.
One of the things I am doing is I'm packing sufficient details into the normals/bump type maps that on their own they do a lot of breaking up of the specular for me, the roughness ends up serving as a master control that twekas it all upwars and downwards as needed.
As I stated before none of these maps are derived from applying actual photos in any way, all tissue samples are derived from painting with brushes, every pore, every wrinkles, every vein or other shading detail. I have done this because I'm no longer convinved that photo based skins have the level of detail that I can produce using brushes and painting in 3d. The lip for example is 100% faked, and actually not my personal favorite aspect but I'm glad it was pleasing to you, means I should rethink my aversion.
Anyhow, later I will upload a few examples of these same skin setting but under a strongly specular light such as the direct sun (relatively small but bright, to give strong hard shadows and sharp intensie specular highlights.
More on the way! As always your feedback is greatly appreciated, Evilded777!
THe fact that they are hand painted is even more impressive. You may be correct in your surmise that photographic resources are not the best solution. I also agree that a map denoting the areas that would be expected to have higher shine is the way to go. I so rarely see a map worth its salt in this area. A good map would accentuate the "T" zone of the face, but for the rest of the human figure would be largely uniform (nails being the exception). I'm very interested in seeing future developments. And anything you might share about materials or techniques for creating these textures.
Those look pretty great! I have to say though, I feel perhaps the bump is a little too strong in certain places, but I as well agree that bumps that are just b&w diffuse maps are no longer cuttting it. I have been dabbling in zbrush a lot more lately, I hope to produce a character eventually. I know modeling all the pores and other skin details can be time consuming but the overal result looks so much better. I also made some additional adjustments to my shader using the -0.5 and I think I am content with it. I will post some resultts later.
Those look awesomely soft skinned in the way they look. Care to share the shader & render settings?
As far as realism I forgot to comment but those are better in my eye than harsh heavily bump & normal mapped style because that's how I see people in real life too. If I was doing a portfolio of beatiful model types for a new clothing line you'd do them like that and not with the harsh bumpier realism. Those are all excellent!
Tried using the eye settings and I ended up with a white ring around the Iris. I inputted all the info exactly as you have it laid out in your settings image post. I got a little improvement using your skin settings which I'm happy about but the ring around the Iris is not so good. Can you help with that J Cade? Side note, my eyes are using the correct UV maps for Genesis 8.
Sorry, I was away for some time.
For the Log function Iray uses the "natural logarithm" (ln), and in reverse the "exponential function", to the base e (approximately equal to 2.71828182845904).
The formulas to calculate the appropriate Transmitted Color/Scattering Color from a known coefficient are
log(Transmitted Color) = Absorption Coefficient * (-1 * Transmitted Measurement Distance)
and
log(Scattering Color) = Scattering Coefficient * (-1 * Scattering Measurement Distance).
Since most coefficients are given in a cm scale, you need to convert them to the base Measurement Distance of DAZ Studios now 1 meter world scene-space scale (simply multiply them with 100), and you also have to keep in mind that DAZ's Iray Uber shader hidden and internally convert it back to a 1 cm world scene-space scale. The exponent to the base e of the log(Transmitted Color) and log(Scattering Color) will get you a decimal that is the corresponding color value (in linear space) which has to be put into DAZ Studios color picker. So you need to calculate three color values for the Red, Green and Blue color channels of the color picker.
On the picture attached the decimal numbers in Red are the corresponding color values for the measured coefficents (average of 21 individuals) given in the box at the bottom of the table for a medium caucasic skin type, for the used 1.00 and 0.01 Measurement Distances. The Scattering Measurement Distance has to be that low because scattering is much stronger than absorption and a higher MD value would exceed the color picker's abilities. The last thing one has to keep in mind that color values of 1.00 and 0.00 will get you very odd (and wrong) rendering results and should be avoided.
This just goes straight over my head. Its like... I know its in English, but suddenly I speak English at a 2nd grade level.
I will read this a few dozen more times, and maybe something will gel.
edit: my post may have been wrong