Fiddling with Iray skin settings...

1798082848591

Comments

  • JimbowJimbow Posts: 557
    RAMWolff said:
    A shader COULD be universal.  Just hold down Ctrl, tell it to ignore maps and you have a better (hopefully) look to your overall set up. 

    Save

    Bear in mind that you can save your shader without maps by clicking off the image maps and UVs during the shader's save process. I do it often for setups I like that I want to be independent of maps and UVs.

  • RAMWolffRAMWolff Posts: 10,140

    OH.  I learned something.  Thanks for the tip Jimbow!  :-)

  • evilded777evilded777 Posts: 2,437
    edited August 2016
    RAMWolff said:
    Arnold C said:
    Some surfaces do have weird glossy colors due to stuff too complex to model. But the problem I run into is that even setting roughness high, it has a tendency to lighten and wash out color. Or is that a good time to start using top coat?

    I really had to think about those surfuces you speak of, but the only surfaces that came to mind were... metals. wink

    That could not only depend on your roughness settings alone. Translucency color and - strength as well as the SSS Reflectance Tint has an impact on the overall result. Not to forget the volume distribution function parameters. And/or the use of a not-so-optimal specular map...

    For skin I always use Top Coat, for there are two different materials, with different optical properties, directly visible: the outermost stratum corneum tissue layer, and the sebum layer.

     

    RAMWolff said:

    I think one of things they could improve on the flow of the surface options is to move the Transmitted Color up under the area that makes more sense since the info that Arnold has imparted to us clearly shows that both are relevent to.  Perhaps I'm wrong but don't understand why it's way way down on the list....

    The Transmitted Color is a parameter belonging to the Volume Distribution Functions of MDL. So IMO it is located at the best place on the user interface as part of the "Volume" group along it "cousins". And you only need those when Thin Walled is turned to "Off".

     

    RAMWolff said:

    It would also be great if someone would make a shader preset that had all the correct values for skin, for nails, for the various eye components and inner mouth and teeth.  I'd buy a set like that and not have it geared for any specific figure, it's off putting when it's for say Genesis 3.  Not all of us use these figures. 

    Me personally, I really don't like much the way shaders are used in DS: having to select all the appropriate surfaces, apply the preset, select another set of surfaces, apply another preset... and so on.

    But one could develop a set of shader presets and make a one-click material preset from those for each of the DAZ figures. Ideally, such a "General Skin Preset" should be added to the "Starter Essentials" package for each figure (in case of M4/V4 to the "Studio CF" package. One really don't needs it if he/she just only uses Poser). For G2F/G2M we already have one (even if it would need an intense overhaul). angel 

    But we would still have a bit of a problem with the specular maps from the older figure lines, which unfortunately come in many different dark and lighter greyscales. The ones for Victoria 6 start to look good at a Glossy Weight of 3.25, the ones for the Genesis 2 Base Female needs even 20.0 to show a convincing effect. Thanks to the bad habit of the "good ole" 3Delight times when everyone tended to cook his/her own soup instead of developing and agreeing to a general method of how things had to be done and stick to that. frown 

    (That are those who now put fancy, ridiculous colors into their Glossy Color parameter... cheeky).

    Hmmm, well OK about the Transmitted ....

    Hmmm, I think you missed that little part I mentioned... I don't use the DAZ figures (rarely).  I develop content for the Hivewire figures so would need to be a universal shader.  Notice I didn't say Material preset.  A shader COULD be universal.  Just hold down Ctrl, tell it to ignore maps and you have a better (hopefully) look to your overall set up.  For a content creator a few extra clicks to get things in place is better than hours and hours of pulling out my hair waiting on test renders. 

    Save

    No, a shader (preset) could not be universal in the way you are asking. A shader preset does not retain the information for different materials.  This is why the various content creators around here have provided there visions in different delivery methods: this one does a material preset, that one does multiple shader presets and another gives you a script that applies a material preset and then other scripts to make global changes to that material preset.

    Its the nature of the beast. It really has to be a material preset if you want it to contain the different information for the different surfaces, and the only way you would get that to be non-figure specific is if you coded a rather large script that could determine what presets to apply where.

     

    And you could always make your material preset for the figures you use... and save it without maps so you can add the ones you want after (or before).

    I have a several base presets that I use based on the different generations, some require a script to copy maps, some don't.  Its relatively easy to do for one's self.

    Post edited by evilded777 on
  • RAMWolffRAMWolff Posts: 10,140

    Sorry, I have to disagree.  A shader can be applied to any selected surface regardless of the UV's, I'm looking for correct surface settings not the mapped areas which can be "ignored" or set up to not load maps according to info that Jimbow related just above. 

  • evilded777evilded777 Posts: 2,437
    RAMWolff said:

    Sorry, I have to disagree.  A shader can be applied to any selected surface regardless of the UV's, I'm looking for correct surface settings not the mapped areas which can be "ignored" or set up to not load maps according to info that Jimbow related just above. 

    I'm not following what you are looking to accomplish then. It sounded to me like you wanted a one click option to set up multiple surfaces.

  • JimbowJimbow Posts: 557
    RAMWolff said:

    Sorry, I have to disagree.  A shader can be applied to any selected surface regardless of the UV's, I'm looking for correct surface settings not the mapped areas which can be "ignored" or set up to not load maps according to info that Jimbow related just above. 

    I think the best thing about a shader preset is it doesn't need to match surface naming like a material preset does. If you're making something specific to a model with definitive surface names, then you can still save a material preset and switch off the image maps and UV sets during the save dialogue thingy.

  • RAMWolffRAMWolff Posts: 10,140
    RAMWolff said:

    Sorry, I have to disagree.  A shader can be applied to any selected surface regardless of the UV's, I'm looking for correct surface settings not the mapped areas which can be "ignored" or set up to not load maps according to info that Jimbow related just above. 

    I'm not following what you are looking to accomplish then. It sounded to me like you wanted a one click option to set up multiple surfaces.

    HAHAHA.. no, my original post about this sort of option would be multiple options, one for skin, one for nails, one for each of the eye surfaces, one for inner mouth and one for teeth! 

  • RAMWolffRAMWolff Posts: 10,140
    Jimbow said:
    RAMWolff said:

    Sorry, I have to disagree.  A shader can be applied to any selected surface regardless of the UV's, I'm looking for correct surface settings not the mapped areas which can be "ignored" or set up to not load maps according to info that Jimbow related just above. 

    I think the best thing about a shader preset is it doesn't need to match surface naming like a material preset does. If you're making something specific to a model with definitive surface names, then you can still save a material preset and switch off the image maps and UV sets during the save dialogue thingy.

    But if it's clicked on won't it actually remove the maps without the end user holding down the Ctrl key for options?

  • evilded777evilded777 Posts: 2,437
    RAMWolff said:
    Jimbow said:
    RAMWolff said:

    Sorry, I have to disagree.  A shader can be applied to any selected surface regardless of the UV's, I'm looking for correct surface settings not the mapped areas which can be "ignored" or set up to not load maps according to info that Jimbow related just above. 

    I think the best thing about a shader preset is it doesn't need to match surface naming like a material preset does. If you're making something specific to a model with definitive surface names, then you can still save a material preset and switch off the image maps and UV sets during the save dialogue thingy.

    But if it's clicked on won't it actually remove the maps without the end user holding down the Ctrl key for options?

    I assume you can save a shader preset in the same way as a material preset and simply tell it not to include the map information and then it will not overwrite them at all.

  • RAMWolffRAMWolff Posts: 10,140

    I'll have to give that a try and see what happens! 

  • JimbowJimbow Posts: 557
    RAMWolff said:
    But if it's clicked on won't it actually remove the maps without the end user holding down the Ctrl key for options?

    Good question. I don't think so, as it's an empty channel. Neither one way or the other.

  • RAMWolffRAMWolff Posts: 10,140

    OK, that works just fine.  I guess I'll have to sit down and make my own shaders for my future projects.  I can do that!  lol cheeky

  • Arnold CArnold C Posts: 740
    edited August 2016

    Even at the highest reflectivity option, I am still getting very flat, almost no gloss.

    I think I need to look at something else, like the glossy weight or the maps.

    What's a good estimation for glossy weight? I don't think the "specular maps" have improved all that much over the older ones, they may be a little lighter, but often they are not.

    That depends on the specular map used. The darker they are, the greater the weight value would have to be. Some "medium-grey" specmaps will work with an amount of 3.25+, for those that are almost black you need an even higher value. I usually try increments of 5 (5.00, 10.00, 15.00, etc) to see when the glossines will start to be visible.

     

    RAMWolff said:

    Hmmm, I think you missed that little part I mentioned... I don't use the DAZ figures (rarely).  I develop content for the Hivewire figures so would need to be a universal shader.  Notice I didn't say Material preset.  A shader COULD be universal.  Just hold down Ctrl, tell it to ignore maps and you have a better (hopefully) look to your overall set up.  For a content creator a few extra clicks to get things in place is better than hours and hours of pulling out my hair waiting on test renders. 

    Well, a material preset for the Hivewire3D characters would also be doable, but I guess here would be the wrong place to ask. JavierMicheal  has already a product you're looking for on his shelf:  JM HumanShader for Iray. Can't say much about its quality, though, but some of the renders do look good so far.

     

    Post edited by Arnold C on
  • evilded777evilded777 Posts: 2,437
    Arnold C said:

    Even at the highest reflectivity option, I am still getting very flat, almost no gloss.

    I think I need to look at something else, like the glossy weight or the maps.

    What's a good estimation for glossy weight? I don't think the "specular maps" have improved all that much over the older ones, they may be a little lighter, but often they are not.

    That depends on the specular map used. The darker they are, the greater the weight value would have to be. Some "medium-grey" specmaps will work with an amount of 3.25+, for those that are almost black you need an even higher value. I usually try increments of 5 (5.00, 10.00, 15.00, etc) to see when the glossines will start to be visible.

     

    RAMWolff said:

    Hmmm, I think you missed that little part I mentioned... I don't use the DAZ figures (rarely).  I develop content for the Hivewire figures so would need to be a universal shader.  Notice I didn't say Material preset.  A shader COULD be universal.  Just hold down Ctrl, tell it to ignore maps and you have a better (hopefully) look to your overall set up.  For a content creator a few extra clicks to get things in place is better than hours and hours of pulling out my hair waiting on test renders. 

    Well, a material preset for the Hivewire3D characters would also be doable, but I guess here would be the wrong place to ask. JavierMicheal  has already a product you're looking for on his shelf:  JM HumanShader for Iray. Can't say much about its quality, though, but some of the renders do look good so far.

     

    I won't speak to "quality" but I can speak to its "plausibility" and its not plausible. I don't know if the re-shade version is more plausible, but I tend to doubt it.

  • Arnold CArnold C Posts: 740
    edited August 2016

    I won't speak to "quality" but I can speak to its "plausibility" and its not plausible. I don't know if the re-shade version is more plausible, but I tend to doubt it.

    Ah, ookay, thanks for the hint. I guess, fancy colors on Glossy Color, and many things mostly guesswork?

    Post edited by Arnold C on
  • evilded777evilded777 Posts: 2,437
    Arnold C said:

    I won't speak to "quality" but I can speak to its "plausibility" and its not plausible. I don't know if the re-shade version is more plausible, but I tend to doubt it.

    Ah, ookay, thanks for the hint. I guess, fancy colors on Glossy Color, and many things mostly guesswork?

    Bingo

  • RAMWolffRAMWolff Posts: 10,140

    Yes, I have that but some of the results are good like the eye settings but the skin settings not so much! 

  • Arnold CArnold C Posts: 740

     

    RAMWolff said:

    Yes, I have that but some of the results are good like the eye settings but the skin settings not so much! 

    In that case you could use its files as a starting point, work in the neccessary changes on things that aren't done correctly, and finally save it as a new shader preset. Saves you a bit of time in contradiction to building your presets from scratch.

    Bingo

    Hm, why doesn't that surprise me much? At least he didn't advertise his product as "being based on real-world measurements", so it can be seen as an "artistical freedom" work. That's a plus.

  • evilded777evilded777 Posts: 2,437
    Arnold C said:

     

    RAMWolff said:

    Yes, I have that but some of the results are good like the eye settings but the skin settings not so much! 

    In that case you could use its files as a starting point, work in the neccessary changes on things that aren't done correctly, and finally save it as a new shader preset. Saves you a bit of time in contradiction to building your presets from scratch.

    Bingo

    Hm, why doesn't that surprise me much? At least he didn't advertise his product as "being based on real-world measurements", so it can be seen as an "artistical freedom" work. That's a plus.

    Along the same lines, does anyone know, exactly, what this means: "based on real world simulation"? And before I get in trouble again, allow me to clarify that there is no negative intent in asking a question. I only want to know what it means.

  • Arnold CArnold C Posts: 740
    Arnold C said:

     

    RAMWolff said:

    Yes, I have that but some of the results are good like the eye settings but the skin settings not so much! 

    In that case you could use its files as a starting point, work in the neccessary changes on things that aren't done correctly, and finally save it as a new shader preset. Saves you a bit of time in contradiction to building your presets from scratch.

    Bingo

    Hm, why doesn't that surprise me much? At least he didn't advertise his product as "being based on real-world measurements", so it can be seen as an "artistical freedom" work. That's a plus.

    Along the same lines, does anyone know, exactly, what this means: "based on real world simulation"? And before I get in trouble again, allow me to clarify that there is no negative intent in asking a question. I only want to know what it means.

    Ah, yeah, that one. smiley What that has to mean knows only the wind (or the PA) alone... from a customer's perspective I'd assume that it would be "based on real-world data and measurements". From an experienced amateur's perspective, who does know a little bit about Physically Based Shading, I can tell that it doesn't include any.

    And so I see it to be just a full-mouthed "hollow phrase" at best, and dangerously close to "False Advertising" at worst. (No negative intent from my side, too! But in my country, exaggeration in business-advertizing is allowed, but telling untruths is against the laws). And from what I know it's also forbidden in the U.S, too. It did cost "Red Bull" 13 million dollars that their drink doesn't "give you wings". wink 

    Adjusting Gamma to make that shader work? ***Err, no comment.*** At least he dropped his cloth shader product from his shelf, so it seems he's not completely resistant to advice.

    That's not to say that the solutions I recommend would be the only one correct, many ways lead to Rome, but at least I spent my time learning the basics of MDL first, and "beg, borrow or steal" measured data wherever I could find them. And if you don't exactly know how or why something works in a certain way, you can always pay a visit to the NVIDIA forums, the folks there are very nice and patient and answer any question, regardless of how stupid it might sound. Registration is free.

  • Rashad CarterRashad Carter Posts: 1,798

    Arnold C,

    I must say a grand thank you to you. Because even if we wanted to, many of us simply don't have the time and patience to dig as deeply as you have to unravel this knot known as Iray. I've been really having a difficult time predicting the outcomes of moving the sliders and activating the various channels. The only way to know what works is to conduct millions of test renders, attempting to isolate the impact of a single parameter alteration... fun in some ways but kinda not as well. Some of this back and forth is fun for me, I'm process oriented anyhow and care less about results. Learning from you that some values are internally squared, others are interanlly square rooted...for Disney purposes and the like...really helps to lift the veil.

    In my opinion, Iray can tend to "price out" the casual hobbyist, using Iray isnt nearly as fun as some of the other engines that do similar things. But what Iray has going for it is that it is highly capable and it is supported by some very smart people.

    Skin Plausibility; All of the human skin shaders so far that I have seen are based in some degree on artistic assumptions. I personally have never seen a chart or graph that displays the exact values for caucasian skin, or african skin. I've never seen where the refraction index of human flesh has been measured. Exactly how much light is supposed to shine though a person's ear when backlit? When a person holds their hand up to a very bright light source how much light should we see gathering along the edges of the fingers? Should we be able to see the silhouttes of the finger bones? Without real world values to pin down these parameters regarding SSS, we still have to judge by eye so even if all of the channels are used correctly, the values placed in some of those channels will need to be judged by eye. What I'm saying is that the information needed to truly remove the need for any artistic license whatsoever in skin shading presets is information that is currently unavailable. So to that end, all shaders are created somewhat equally.

  • evilded777evilded777 Posts: 2,437

    Arnold C,

    I must say a grand thank you to you. Because even if we wanted to, many of us simply don't have the time and patience to dig as deeply as you have to unravel this knot known as Iray. I've been really having a difficult time predicting the outcomes of moving the sliders and activating the various channels. The only way to know what works is to conduct millions of test renders, attempting to isolate the impact of a single parameter alteration... fun in some ways but kinda not as well. Some of this back and forth is fun for me, I'm process oriented anyhow and care less about results. Learning from you that some values are internally squared, others are interanlly square rooted...for Disney purposes and the like...really helps to lift the veil.

    In my opinion, Iray can tend to "price out" the casual hobbyist, using Iray isnt nearly as fun as some of the other engines that do similar things. But what Iray has going for it is that it is highly capable and it is supported by some very smart people.

    Skin Plausibility; All of the human skin shaders so far that I have seen are based in some degree on artistic assumptions. I personally have never seen a chart or graph that displays the exact values for caucasian skin, or african skin. I've never seen where the refraction index of human flesh has been measured. Exactly how much light is supposed to shine though a person's ear when backlit? When a person holds their hand up to a very bright light source how much light should we see gathering along the edges of the fingers? Should we be able to see the silhouttes of the finger bones? Without real world values to pin down these parameters regarding SSS, we still have to judge by eye so even if all of the channels are used correctly, the values placed in some of those channels will need to be judged by eye. What I'm saying is that the information needed to truly remove the need for any artistic license whatsoever in skin shading presets is information that is currently unavailable. So to that end, all shaders are created somewhat equally.

    I disagree. There actually is quite a bit of information on human skin available; translating that to MDL, that's the challenge. There are scientific principles at work and defined methods available to us. Most of the people peddling material presets pay no attention to either.

    Certainly, there is always going to be some room for artistic expression, but that's wiggle room for tweaking setting to your liking, not going off and defining your own rules because you think it looks right.

  • Rashad CarterRashad Carter Posts: 1,798

    It's in the translation into Iray values that I'm talking about. What I mean to say is that there is no information about these values available in a manner that can be represented readily in Iray. The units of measure, vocabulary and all are completely different in the technical documents than in the render thread related sites. Perhaps if we had the option to initiate a fully Scientific Nomenclature Layout option that would be awesome, including nm values and other relevant values. But alas, unless I;ve really missed something, this information doesnt exist. If someone could devise such a thing, that would be incredible and very useful.

  • evilded777evilded777 Posts: 2,437

    It's in the translation into Iray values that I'm talking about. What I mean to say is that there is no information about these values available in a manner that can be represented readily in Iray. The units of measure, vocabulary and all are completely different in the technical documents than in the render thread related sites. Perhaps if we had the option to initiate a fully Scientific Nomenclature Layout option that would be awesome, including nm values and other relevant values. But alas, unless I;ve really missed something, this information doesnt exist. If someone could devise such a thing, that would be incredible and very useful.

    Point taken, sir. And a well made one.

  • Rashad CarterRashad Carter Posts: 1,798
     

    Certainly, there is always going to be some room for artistic expression, but that's wiggle room for tweaking setting to your liking, not going off and defining your own rules because you think it looks right.

    Training the eye is all. If people were better at knowing what looks right in the first place, they should still manage to arrive at arbitrary values that are at least similar to the correct values. I'm  convinced too many of us are not paying close enough attention to the real thing. For example a random sculptor may not have exact measurements for the human form he's carving, but if he pays close enough attention to his resource material he should still end up in the ball park.

    That's not to say that imperical measuremnts aren't incredbily useful, because they remove the guesswork. But with something like human skin, derived from maps produced in a million different ways by different artists; I convinced there will always be a certain degree of guess work.

    If I am to be honest, the biggest problem facing anyone trying to build skin is that too many of us still haven't mastered light. Light still comes first. Just because we jumped to physically based rendering options doesnt change the fact that light is still the essential key. How and why light behaves as it does. Too many of us rely solely on HDRI images because we have no notion of how to properly illuminate a 3d environemnt otherwise. Not that HDRI are not awesome, but they shouldn't be the only situations we've tested skins under if we want to distribute those skins to others as flexible enough to be useful in almost any situation.

    And flexibility is the true test. If a shader cannot withstand lots of different lighting arrangements, then it probably still isn't ready for release or very close to being physically accurate.

     

  • Arnold CArnold C Posts: 740
    edited August 2016

    Arnold C,

    I must say a grand thank you to you. Because even if we wanted to, many of us simply don't have the time and patience to dig as deeply as you have to unravel this knot known as Iray. I've been really having a difficult time predicting the outcomes of moving the sliders and activating the various channels. The only way to know what works is to conduct millions of test renders, attempting to isolate the impact of a single parameter alteration... fun in some ways but kinda not as well. Some of this back and forth is fun for me, I'm process oriented anyhow and care less about results. Learning from you that some values are internally squared, others are interanlly square rooted...for Disney purposes and the like...really helps to lift the veil.

    In my opinion, Iray can tend to "price out" the casual hobbyist, using Iray isnt nearly as fun as some of the other engines that do similar things. But what Iray has going for it is that it is highly capable and it is supported by some very smart people.

    Skin Plausibility; All of the human skin shaders so far that I have seen are based in some degree on artistic assumptions. I personally have never seen a chart or graph that displays the exact values for caucasian skin, or african skin. I've never seen where the refraction index of human flesh has been measured. Exactly how much light is supposed to shine though a person's ear when backlit? When a person holds their hand up to a very bright light source how much light should we see gathering along the edges of the fingers? Should we be able to see the silhouttes of the finger bones? Without real world values to pin down these parameters regarding SSS, we still have to judge by eye so even if all of the channels are used correctly, the values placed in some of those channels will need to be judged by eye. What I'm saying is that the information needed to truly remove the need for any artistic license whatsoever in skin shading presets is information that is currently unavailable. So to that end, all shaders are created somewhat equally.

    You're welcome, Rashad. In case of the Disney stuff you'd have to thank AndyGrimm, too, for the discovery has been a result of an internal conversation and deep reconaissannce for his "The 5 simple formulas you must know for mastering IRAY" collection, along studying the bricks on the Shader Mixer Tab and reading through the "Physically-Based Shading at Disney" document (which is cited in the "irayubermaterial.mdl" in .\Program Files\DAZ 3D\DAZStudio4\shaders\iray\daz_3d). smiley

    Actually, they're only get squared, but if you want to use a measured value on your shader, you'd have  either to calculate the square root for it first, to get the value used you wanted, or simply disable Roughness Squared and Glosiness Squared on the Surfaces Tabwink

    Measurements of optical properties of human skin can be found, among others, in:

    Optical properties of human skin
    Tom Lister, Philip A. Wright and Paul H. Chappell
    Journal of Biomedical Optics 17, (September 2012)


    OPTICAL PROPERTIES OF HUMAN SKIN, SUBCUTANEOUS AND MUSCOUS TISSUES IN THE WAVELENGTH RANGE FROM 400 TO 2000 NM

    A. N. BASHKATOV, E. A. GENINA, V. I. KOCHUBEY and V. V. TUCHIN

    Journal of Physics D: Applied Physics
    Volume 38, Issue 15 (2005) 2543-2555
    INSTITUTE OF PHYSICS PUBLISHING
    DOI: 10.1088/0022-3727/38/15/004


    OPTICAL PROPERTIES OF SKIN, SUBCUTANEOUS AND MUSCLE TISSUES: A REVIEW

    ALEXEY N. BASHKATOV, ELINA A. GENINA, and VALERY V. TUCHIN, J.

    Journal of Innovative Optical Health Sciences
    Vol. 4, No. 1 (2011) 9-38
    World Scientific Publishing Company
    DOI: 10.1142/S1793545811001319


    Optical properties of the human skin
    Zorica Gajinov, Milan MatiĆ, Sonja PrĆiĆ, Verica Đuran.
    Institute of Child and Youth Health Care of Vojvodina, Novi Sad, Serbia
    Clinic of Dermatovenereology Diseases, Clinical Center of Vojvodina, Novi Sad, Serbia
    DOI: 10.2478/v10249-011-0029-5


    The Appearance of Human Skin
    Takanori Igarashi, Ko Nishino, and Shree K. Nayar
    Technical Report: CUCS-024-05
    Department of Computer Science
    Columbia University


    The BASHKATOV, GENINA, and TUCHIN publications cover the whole spectrum of visible light, and so they're IMO optimal to derive parameter values for MDL-based shaders.


    In vivo determination of skin near-infrared optical properties using diffuse optical spectroscopy
    Sheng-Hao Tseng, Alexander Grant, and Anthony J. Durkin
    University of California, Irvine Beckman Laser Institute Laser Microbeam and Medical Program
    Journal of Biomedical Optics 2008
    DOI: 10.1117/1.2829772.

    includes a comparision of the optical properties of African, Asian and Caucasian descent human skin types. Unfortunately it covers only the 650 nm to 1000 nm wavelenth range, for our purpose we need the range from 420 nm to 750 nm.

     

    It's in the translation into Iray values that I'm talking about. What I mean to say is that there is no information about these values available in a manner that can be represented readily in Iray. The units of measure, vocabulary and all are completely different in the technical documents than in the render thread related sites. Perhaps if we had the option to initiate a fully Scientific Nomenclature Layout option that would be awesome, including nm values and other relevant values. But alas, unless I;ve really missed something, this information doesnt exist. If someone could devise such a thing, that would be incredible and very useful.

    A big +1 on this. It is there, but not that openly published. Unfortunately, the DAZ Uber Iray Documentation isn't very helpful about to get the information of how someone has to put his/her measured values into, and where, and even the official older NVIDIA MDL Handbook and Specs were short on some vital information. So you need to do a "Holmes & Watson" and find out for yourself. Or have someone to find out for you. smiley

     

    Training the eye is all. If people were better at knowing what looks right in the first place, they should still manage to arrive at arbitrary values that are at least similar to the correct values. I'm  convinced too many of us are not paying close enough attention to the real thing. For example a random sculptor may not have exact measurements for the human form he's carving, but if he pays close enough attention to his resource material he should still end up in the ball park.

    That's not to say that imperical measuremnts aren't incredbily useful, because they remove the guesswork. But with something like human skin, derived from maps produced in a million different ways by different artists; I convinced there will always be a certain degree of guess work.

    If I am to be honest, the biggest problem facing anyone trying to build skin is that too many of us still haven't mastered light. Light still comes first. Just because we jumped to physically based rendering options doesnt change the fact that light is still the essential key. How and why light behaves as it does. Too many of us rely solely on HDRI images because we have no notion of how to properly illuminate a 3d environemnt otherwise. Not that HDRI are not awesome, but they shouldn't be the only situations we've tested skins under if we want to distribute those skins to others as flexible enough to be useful in almost any situation.

    And flexibility is the true test. If a shader cannot withstand lots of different lighting arrangements, then it probably still isn't ready for release or very close to being physically accurate.

    Another +1. There's no absolute measurements for skin, for there are many, many different skin types, from dry to oily around, and the Melanin concentration isn't the same for each individuum. But there still are ranges to get your data from... or simply use averages.

    For shader development I use a (neutral) spotlight AndyGrimm recommends: 

     Luminous Flux (Lumen): 15000.0
     Temperature (K): 5000.00
     Light Geometry: "Disc"
     Width: 10.00 (cm)

    positioned in 100 units (cm) from the figure's head in height of the eyes and a photo reference (Adriana Lima, no makeup).

     

    Post edited by Arnold C on
  • j cadej cade Posts: 2,310
    edited August 2016
     

    Certainly, there is always going to be some room for artistic expression, but that's wiggle room for tweaking setting to your liking, not going off and defining your own rules because you think it looks right.

    Training the eye is all. If people were better at knowing what looks right in the first place, they should still manage to arrive at arbitrary values that are at least similar to the correct values. I'm  convinced too many of us are not paying close enough attention to the real thing. For example a random sculptor may not have exact measurements for the human form he's carving, but if he pays close enough attention to his resource material he should still end up in the ball park.

    That's not to say that imperical measuremnts aren't incredbily useful, because they remove the guesswork. But with something like human skin, derived from maps produced in a million different ways by different artists; I convinced there will always be a certain degree of guess work.

     

    That's why I like skin, there's always at least one perfectly good example near by :) different fabrics, for instance, require actual moving to get reference. I'm less math and physics based than Arnold and Andy. I'm more simple experiments I can conduct myself and some basic logical thinking. But then I love doing test renders and experiments

     

    If I am to be honest, the biggest problem facing anyone trying to build skin is that too many of us still haven't mastered light. Light still comes first. Just because we jumped to physically based rendering options doesnt change the fact that light is still the essential key. How and why light behaves as it does. Too many of us rely solely on HDRI images because we have no notion of how to properly illuminate a 3d environemnt otherwise. Not that HDRI are not awesome, but they shouldn't be the only situations we've tested skins under if we want to distribute those skins to others as flexible enough to be useful in almost any situation.

    And flexibility is the true test. If a shader cannot withstand lots of different lighting arrangements, then it probably still isn't ready for release or very close to being physically accurate.

    My first test is always to kill all other lights and strongly backlight a figure, with so many skin material settings  the whole body will glow red... argh. Humans don't do that. Of course, when you're using the general ambient light that tends to come from most HDRIs its not immediately evident whats going on even if you have a good enough eye to tell somethings off, but the moment someone sticks those settings in less ambient lighting it will look really off.

     

     

    @evildead I keep my fingernails with mostly the same settings as the skin but I set the refraction weight to .2 the refraction index to 1 and lower the sss amount to .45 or so

    to be extra fancy I add a bump map to the top coat to simulate nail ridges (a super simple map, just black and white vertical lines)

    I have the base glossy match the skin exactly (a lot of the daz skins have the reflectivity set to something to other than .5 I am not a fan of it)  I control the overall strength and roughness of the nail gloss in the topcoat

     

    The most important thing IMO is the refraction weight though. The gloss and the ridges really aren't noticeable unless the hand is super close up, but the semi-transparency really helps it match up to the skin. Its pretty real world accurate too, I mean if your nails aren't completely down to the stubs take a close look at that area between where they're attached to the skin and where they turn white. Its pretty transparent there, no? (thats why the dirt shows up so easily sad )

    nailss.png
    600 x 780 - 902K
    Post edited by j cade on
  • RAMWolffRAMWolff Posts: 10,140

    This is awesome information jcade!  Thanks so much!  :-)

  • Arnold CArnold C Posts: 740
    edited August 2016
    j cade said:

    @evildead I keep my fingernails with mostly the same settings as the skin but I set the refraction weight to .2 the refraction index to 1 and lower the sss amount to .45 or so

    to be extra fancy I add a bump map to the top coat to simulate nail ridges (a super simple map, just black and white vertical lines)

    I have the base glossy match the skin exactly (a lot of the daz skins have the reflectivity set to something to other than .5 I am not a fan of it)  I control the overall strength and roughness of the nail gloss in the topcoat

    A Refraction Index of 1 would mean, that the nails would be made of vacuum, which's result would be, that no refraction at all would occur. Normally, nails have a Refraction Index of between 1.54 to 1.56. A solution I prefer is to make the Base define the material of the nail, and the Top Coat define a possible used varnish. That way you can switch from unvarnished to varnished nails by just disable/enable the Top Coat, and you can use this preset on girls and guys. wink

    Using Refraction Weight and the Nail Ridges bump map is a good idea though. smiley

    Post edited by Arnold C on
  • j cadej cade Posts: 2,310
    Arnold C said:
    j cade said:

    @evildead I keep my fingernails with mostly the same settings as the skin but I set the refraction weight to .2 the refraction index to 1 and lower the sss amount to .45 or so

    to be extra fancy I add a bump map to the top coat to simulate nail ridges (a super simple map, just black and white vertical lines)

    I have the base glossy match the skin exactly (a lot of the daz skins have the reflectivity set to something to other than .5 I am not a fan of it)  I control the overall strength and roughness of the nail gloss in the topcoat

    A Refraction Index of 1 would mean, that the nails would be made of vacuum, which's result would be, that no refraction at all would occur. Normally, nails have a Refraction Index of between 1.54 to 1.56. A solution I prefer is to make the Base define the material of the nail, and the Top Coat define a possible used varnish. That way you can switch from unvarnished to varnished nails by just disable/enable the Top Coat, and you can use this preset on girls and guys. wink

    Using Refraction Weight and the Nail Ridges bump map is a good idea though. smiley

     

    To my mind refraction index of 0 is "cutout opacity without killing volume" But, in any case, the nails themselves aren't completely physically accurate either. They're thicker than mine certainly and don't sit perfectly flush against the skin, not that I'd expect them to either, modelling that would be an absolute nightmare, (actually the underside of the nails would need to sit very slightly under the skin for the refraction look as expected, but that would run into problems too where the nail no longer sits against the skin.) all in all it would be way more difficult for the very minimal effect of a bit of refraction over somthing less than a half a milimmetre thick

Sign In or Register to comment.