Fiddling with Iray skin settings...

1808183858691

Comments

  • Arnold CArnold C Posts: 740
    j cade said:

    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

    On the Iray Uber, the Refraction Index/IOR doesn't have an impact on the volume distribution functions, it's for specular reflection: "the ior parameter controls how much the light the layer parameter’s BSDF reflects" (MDL Handbook [as of 18.07.2016], Page 67). Even in real-world there's no IOR lower than 1 (vacuum), meaning that a lightray will travel trough a medium at the speed of light without any velocity loss.

    But since you're using mainly your Top Coat to determine specular reflection, that's okay. In that case you could set Share Glossy Inputs to "Off" and also set Base's Glossy Weight to 0.00 and Glossy Color (Glossy Specular) to 0.00 black to prevent to come into effect.

  • j cadej cade Posts: 2,310
    Arnold C said:
    j cade said:

    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

    On the Iray Uber, the Refraction Index/IOR doesn't have an impact on the volume distribution functions, it's for specular reflection: "the ior parameter controls how much the light the layer parameter’s BSDF reflects" (MDL Handbook [as of 18.07.2016], Page 67). Even in real-world there's no IOR lower than 1 (vacuum), meaning that a lightray will travel trough a medium at the speed of light without any velocity loss.

    But since you're using mainly your Top Coat to determine specular reflection, that's okay. In that case you could set Share Glossy Inputs to "Off" and also set Base's Glossy Weight to 0.00 and Glossy Color (Glossy Specular) to 0.00 black to prevent to come into effect.

    Exactly. Thats the point, because, cutout opacity does. I am using "refraction" purely as a tool to make the nails transparent without losing the volumeetrics it already has (my skin settings).

  • Oso3DOso3D Posts: 14,874
    edited August 2016

    Working on skins, I have vaguely assumed that most stuff should be 'scatter & transmit.'

    I _think_ I remember reading that glossy weight & top coat weight should add up to 1. Testing, it seems if I set glossy weight to 1 regardless, the color will shift a little and it doesn't look quite right, and that 'add up to 1' works mostly. (I'm mainly trying to set a 'dry look' and then add wetness on top with top coat)

    Does translucency weight play into that?

    I've noticed that the translucence of the ears (for example) drops when top coat goes up, even with scatter & transmit. ... um.

     

    I'm also debating whether I should bother with top coat at all... using just glossy might be fine. Hrm

    Post edited by Oso3D on
  • SorelSorel Posts: 1,389

    So I'm sure it has been mentioned somewhere in this thread, but is the specular/reflection color of human skin supposed to be just white? As is full white?  

  • Oso3DOso3D Posts: 14,874

    I've found that for all things Iray, I can get a good enough result by just using Glossy Weight 1, Glossy Color pure weight, and then juggling reflectivity and roughness.

     

  • nonesuch00nonesuch00 Posts: 17,890
    Sorel said:

    So I'm sure it has been mentioned somewhere in this thread, but is the specular/reflection color of human skin supposed to be just white? As is full white?  

    http://forum.unity3d.com/threads/official-5-0-pbr-calibration-charts.289416/

    They have both Metallic and Specular setup charts.

    Also, basically says dielectrics like skin will be a shade of grey and human skin is at 0.028.

    https://seblagarde.wordpress.com/2011/08/17/feeding-a-physical-based-lighting-mode/

     

  • SorelSorel Posts: 1,389

    hmmm, when I use that grey color I get no shine >_>  

  • SorelSorel Posts: 1,389

    I should also mention I render with octane, so I guess that is where things would get a bit more complicated, since the only parameter I have for specular is...well, specular/reflection

  • nonesuch00nonesuch00 Posts: 17,890
    edited September 2016
    Sorel said:

    hmmm, when I use that grey color I get no shine >_>  

    Oh, for that I learned the easy what is to delete the Top Coat Color greyscale map, I think it's called a Spectral Map, leave the Topcoat weight at 1.0, and the Top Coat Roughness at about 0.66 will create a bit of an oily sheen.

    Alternatively you can do something I haven't tried and that's keep that Top Coat greyscale / Spectral greyscale map and ramp up the Top Coat Weight way, way up to 10.0 or even higher because in real life normally the face is much oilier than other parts of the body (the reason why they made that greyscale map to begin with).

    Sorry, I don't have Octane but the problem I think is the Spectral Map needs much more weight assigned to it for the faster & easier to try solution to your problem. 

    +++++

    Seems I can only make the Top Coat Weight go up to 2.0...but without deleting the Top Coat Map that is not enough to make her face look that oily really. So the alternative is to delete the Top Coat Map as already mentioned or edit the Top Coat Map and make the lightest greyscale on that map be a value of 250 250 250 and each other grey scale value likewise shifted to 255 255 255 be the same amount...

     

    Post edited by nonesuch00 on
  • SorelSorel Posts: 1,389

    Yeah I figure there are not a lot of octane users here, but I try to take from threads like this and see how much of what I can translate over. I've been stalking this thread for months and there has been a lot of info put in here.  I ended up just inverting the grey color and now it actually looks pretty good.  I've tried researching other people's octane shaders but a lot of them dont look realistic enough, or the ones that do dont give any instruction on what they did haha.  

  • nonesuch00nonesuch00 Posts: 17,890
    Sorel said:

    Yeah I figure there are not a lot of octane users here, but I try to take from threads like this and see how much of what I can translate over. I've been stalking this thread for months and there has been a lot of info put in here.  I ended up just inverting the grey color and now it actually looks pretty good.  I've tried researching other people's octane shaders but a lot of them dont look realistic enough, or the ones that do dont give any instruction on what they did haha.  

    Yes, deleting the map or shifting all the grey scale values on the map to 255 255 255 so that the highest is at say, 250 or 245 is the best but you don't want to go too close to 255 because of the water wet look starts to take over.

  • RafmerRafmer Posts: 564
    Sorel said:

    Yeah I figure there are not a lot of octane users here, but I try to take from threads like this and see how much of what I can translate over. I've been stalking this thread for months and there has been a lot of info put in here.  I ended up just inverting the grey color and now it actually looks pretty good.  I've tried researching other people's octane shaders but a lot of them dont look realistic enough, or the ones that do dont give any instruction on what they did haha.  

    Yes, deleting the map or shifting all the grey scale values on the map to 255 255 255 so that the highest is at say, 250 or 245 is the best but you don't want to go too close to 255 because of the water wet look starts to take over.

    You can edit the parameters of any slider and remove their limits.
  • nonesuch00nonesuch00 Posts: 17,890
    Rafmer said:
    Sorel said:

    Yeah I figure there are not a lot of octane users here, but I try to take from threads like this and see how much of what I can translate over. I've been stalking this thread for months and there has been a lot of info put in here.  I ended up just inverting the grey color and now it actually looks pretty good.  I've tried researching other people's octane shaders but a lot of them dont look realistic enough, or the ones that do dont give any instruction on what they did haha.  

    Yes, deleting the map or shifting all the grey scale values on the map to 255 255 255 so that the highest is at say, 250 or 245 is the best but you don't want to go too close to 255 because of the water wet look starts to take over.

     

    You can edit the parameters of any slider and remove their limits.

    Thanks

  • Arnold CArnold C Posts: 740
    edited September 2016
    Sorel said:

    So I'm sure it has been mentioned somewhere in this thread, but is the specular/reflection color of human skin supposed to be just white? As is full white?  

    Specular reflection from skin is always white (1.0, 1.0, 1.0 in linear space or 255, 255, 255 sRGB). See NVIDIA's GPU GEMS 3, Chapter 14:

    "Specular Reflectance from Skin Is White

    The tissue cells and oil in the outermost layer of skin are dielectric materials that reflect light without coloring it (whereas metals such as gold color the light that is reflected due to highly varying indices of refraction over visible wavelengths of light.) Thus, a physically based skin shader should use a white specular color. In other words, the specular reflection of a white light from skin will be white, and the specular reflection of a colored light will be the same color as that light—regardless of the color of the underlying skin..."


    Though I'm not that familiar with the Octane render itself, too (and looking at the prizes Otoy charges that will most likely never,ever happen...), but taking a glimpse at the Materials Tab from the Octane for DAZ Studio plugin, ("Specular Material" type), their "Reflection" parameter would be identical to DAZ's Iray Uber's "Glossy Color".

    It seems they don't use different mixes like DAZ, and they don't have a "Glossy Reflectivity" or "Glossy Specular" parameter. So I'd assume that "Specular Albedo/Reflectance" is just only determined by their "Index of refraction" parameters.

    My recommendation: leave the "Reflection" parameter at 255, 255, 255 and put in a valid Refractive Index. Alternatively, for human skin you could make use of the "Material Mix", if possible:

    - a "Glossy Material" with an Albedo/Diffuse texture, "Specular" 1, "Roughness" 0.3 and an "Index of refraction)" of 1.52, to mimic the stratum corneum outermost visible skin layer,

    - and a "Specular Material" with "Roughness" 0.3 (or just slightly lower, but not much) and an "Index of refraction" of 1.5 to mimic the sebum (thin oily film) on top of it.


    Also, basically says dielectrics like skin will be a shade of grey and human skin is at 0.028.

    https://seblagarde.wordpress.com/2011/08/17/feeding-a-physical-based-lighting-mode/

    Sébastien Lagarde's article (or NVIDIA's GPU GEMS 3) is a bit ambigous on this matter, since he uses the term "Specular Color" to explain "Specular Albedo/Specular Reflectance", where NVIDIA explaines the color of the specular reflection.

    In case for DAZ's Iray Uber, the "Specular Color" ("Glossy Color" on the Iray Uber) for a dielectric has always to be plain white, the "Specular Albedo/Specular Reflectance" is determined differently, depending on the "Base Mixing" mode used, and in a range between

    "Specular Reflectance" [F(0°) (or R0 in case of  Schlick's approximation)] is determined by either the "Glossy Reflectivity" parameter (on the PBR Metallicity/Roughness Mix), or the "Glossy Specular" parameter (on the PBR Specular/Glossiness Mix). Both are based on a material's Refractive Index, and unfortunately have to be pre-calculated manually.

    Using the formula mentioned on Legarde's Blog, and a Refractive Index of 1.52 (in vivo measured value for "Skin (Stratum Corneum)": [Tearney G. J., Brezinski M. E., Southern J. F., Bouma B. E., Hee M. R. and Fujimoto J. G.: “Determination of the refractive-index of highly scattering human tissue by optical coherence tomography.” Optics Letters, Volume 20, Issue 21, 1995, Pages 2258–2260]), you'll get a value of 0.0465210 0.042580 (*). That's more than 1.5 times larger than the "sample value" Legarde mentiones (0.028, corresponds to an IOR of 1.402). I'd recommend to always use measured values, and calculate the appropriate F(0°) or "Glossy Reflectivity" for them, above a general "sample value". Legarde's 1.4 looks like to be an average for the IOR's of the stratum corneum, the epidermis and dermis skin tissues... but the latter are more appropriate to determine SSS parameters than Specular Reflection, the stratum corneum and sebum are the two only things which can be viewed directly, and so have a direct impact on that.

    I made myself an OpenOffice calcsheet up a while ago to help calculate R0, IOR, Glossy Reflectivity and Glossy Roughness parameter values for use on the Iray Uber. Those who are interested could try it out. It's been made up with the german language version of OpenOffice (although the text is in english), I hope it will work on other language versions, too.


    (*) EDIT: Oops, sorry, wrong line of my crib sheet. The 0.0465210 is of course for an IOR of 1.55. blush

    Post edited by Arnold C on
  • nonesuch00nonesuch00 Posts: 17,890
    edited September 2016
    Arnold C said:
    Sorel said:

    So I'm sure it has been mentioned somewhere in this thread, but is the specular/reflection color of human skin supposed to be just white? As is full white?  

    Specular reflection from skin is always white (1.0, 1.0, 1.0 in linear space or 255, 255, 255 sRGB). See NVIDIA's GPU GEMS 3, Chapter 14:

    "Specular Reflectance from Skin Is White

    The tissue cells and oil in the outermost layer of skin are dielectric materials that reflect light without coloring it (whereas metals such as gold color the light that is reflected due to highly varying indices of refraction over visible wavelengths of light.) Thus, a physically based skin shader should use a white specular color. In other words, the specular reflection of a white light from skin will be white, and the specular reflection of a colored light will be the same color as that light—regardless of the color of the underlying skin..."


    Though I'm not that familiar with the Octane render itself, too (and looking at the prizes Otoy charges that will most likely never,ever happen...), but taking a glimpse at the Materials Tab from the Octane for DAZ Studio plugin, ("Specular Material" type), their "Reflection" parameter would be identical to DAZ's Iray Uber's "Glossy Color".

    It seems they don't use different mixes like DAZ, and they don't have a "Glossy Reflectivity" or "Glossy Specular" parameter. So I'd assume that "Specular Albedo/Reflectance" is just only determined by their "Index of refraction" parameters.

    My recommendation: leave the "Reflection" parameter at 255, 255, 255 and put in a valid Refractive Index. Alternatively, for human skin you could make use of the "Material Mix", if possible:

    - a "Glossy Material" with an Albedo/Diffuse texture, "Specular" 1, "Roughness" 0.3 and an "Index of refraction)" of 1.52, to mimic the stratum corneum outermost visible skin layer,

    - and a "Specular Material" with "Roughness" 0.3 (or just slightly lower, but not much) and an "Index of refraction" of 1.5 to mimic the sebum (thin oily film) on top of it.


    Also, basically says dielectrics like skin will be a shade of grey and human skin is at 0.028.

    https://seblagarde.wordpress.com/2011/08/17/feeding-a-physical-based-lighting-mode/

    Sébastien Lagarde's article (or NVIDIA's GPU GEMS 3) is a bit ambigous on this matter, since he uses the term "Specular Color" to explain "Specular Albedo/Specular Reflectance", where NVIDIA explaines the color of the specular reflection.

    In case for DAZ's Iray Uber, the "Specular Color" ("Glossy Color" on the Iray Uber) for a dielectric has always to be plain white, the "Specular Albedo/Specular Reflectance" is determined differently, depending on the "Base Mixing" mode used, and in a range between

    "Specular Reflectance" [F(0°) (or R0 in case of  Schlick's approximation)] is determined by either the "Glossy Reflectivity" parameter (on the PBR Metallicity/Roughness Mix), or the "Glossy Specular" parameter (on the PBR Specular/Glossiness Mix). Both are based on a material's Refractive Index, and unfortunately have to be pre-calculated manually.

    Using the formula mentioned on Legarde's Blog, and a Refractive Index of 1.52 (in vivo measured value for "Skin (Stratum Corneum)": [Tearney G. J., Brezinski M. E., Southern J. F., Bouma B. E., Hee M. R. and Fujimoto J. G.: “Determination of the refractive-index of highly scattering human tissue by optical coherence tomography.” Optics Letters, Volume 20, Issue 21, 1995, Pages 2258–2260]), you'll get a value of 0.0465210. That's more than 1.5 times larger than the "sample value" Legarde mentiones (0.028, corresponds to an IOR of 1.402). I'd recommend to always use measured values, and calculate the appropriate F(0°) or "Glossy Reflectivity" for them, above a general "sample value". Legarde's 1.4 looks like to be an average for the IOR's of the stratum corneum, the epidermis and dermis skin tissues... but the latter are more appropriate to determine SSS parameters than Specular Reflection, the stratum corneum and sebum are the two only things which can be viewed directly, and so have a direct impact on that.

    I made myself an OpenOffice calcsheet up a while ago to help calculate R0, IOR, Glossy Reflectivity and Glossy Roughness parameter values for use on the Iray Uber. Those who are intersted could try it. It's been made up with the german language version of OpenOffice (although the text is in english), I hope it will work on other language versions, too.

    Thanks very much. I was reading that webpage of LeGarde's that was the closest thing to a direct answer I could find. May I never do another websearch for human skin PBR values again.

    Post edited by Chohole on
  • Arnold CArnold C Posts: 740
    edited September 2016

    For those who already downloaded my OpenOffice calcsheet: it got an overhaul and now includes a formula to calculate the appropriate "Glossiness" parameter value for the PBR Specular/Glossiness Mix [taking into account that DAZ added a "Glossiness Squared" property to the NVIDIA Iray Uber material on DAZ Studio 4.9.2.54 (and above)].

    Post edited by Arnold C on
  • nonesuch00nonesuch00 Posts: 17,890

    Thanks

  • Thanks

    You're twice welcome. smiley 

    Do you use a different language version than german, and does it work for you in there? I'm too lazy to deinstall OO and reinstall an english version. blush

  • nonesuch00nonesuch00 Posts: 17,890
    Arnold C said:

    Thanks

    You're twice welcome. smiley 

    Do you use a different language version than german, and does it work for you in there? I'm too lazy to deinstall OO and reinstall an english version. blush

    Well, lucky I still remember enough Gernman from when I worked in Switzerland. Thanks.

  • Well, lucky I still remember enough Gernman from when I worked in Switzerland. Thanks.

    Ah, nice. But... "swiss" german isn't really a "german" german. For all those "chr's" they build in words ending on "k"... cheeky Just kidding! smiley

    That's not what I meant, the descriptive text in my document is held in english. But I wasn't sure wether the formulas are all kept intact on a different language version, as well as if the difference in punctuation may could cause trouble. We germans use a "," (komma) where in the U.S. a "." (point) is used. I still remember having some trouble using U.S. formatted documents back in the days when I used M$'s "Word" and "Excel". smiley

     

  • nonesuch00nonesuch00 Posts: 17,890

    Oh, sorry. well I only opened and looked but not used the formulas yet. I want to combine with Will's TP2 shaders after I finish remodeing the living room.

    For what I see though it looks like it is written numerically using American notation even though I have a non-English language chosen for the keyboard I'm using. I agree, Swiss and European numeric notation is confusing to Americans that arent used to it. I have it opened in MS Excel 2015. I only have that because Micrsoft gave it to me free.

    I have software keyboards where in Windows 10 I can switch easy between different countires but the OS subsystem itself remains American notation underneath. To actually effect Windows 10 compliant language specifications I have to change the Windows 10 Locale to be Germany or Switzerland or so on. Even then, the Windows 10 OS won't change the format of your document even if i change the laguage & region for my PC to be Germany. It will though then start causing new documents done in Microsoft Windows 10 OS locale compliant SW to be created using the numerical notation for the local chosen.

  • SorelSorel Posts: 1,389
    Arnold C said:
    Sorel said:

    So I'm sure it has been mentioned somewhere in this thread, but is the specular/reflection color of human skin supposed to be just white? As is full white?  

    Specular reflection from skin is always white (1.0, 1.0, 1.0 in linear space or 255, 255, 255 sRGB). See NVIDIA's GPU GEMS 3, Chapter 14:

    "Specular Reflectance from Skin Is White

    The tissue cells and oil in the outermost layer of skin are dielectric materials that reflect light without coloring it (whereas metals such as gold color the light that is reflected due to highly varying indices of refraction over visible wavelengths of light.) Thus, a physically based skin shader should use a white specular color. In other words, the specular reflection of a white light from skin will be white, and the specular reflection of a colored light will be the same color as that light—regardless of the color of the underlying skin..."


    Though I'm not that familiar with the Octane render itself, too (and looking at the prizes Otoy charges that will most likely never,ever happen...), but taking a glimpse at the Materials Tab from the Octane for DAZ Studio plugin, ("Specular Material" type), their "Reflection" parameter would be identical to DAZ's Iray Uber's "Glossy Color".

    It seems they don't use different mixes like DAZ, and they don't have a "Glossy Reflectivity" or "Glossy Specular" parameter. So I'd assume that "Specular Albedo/Reflectance" is just only determined by their "Index of refraction" parameters.

    My recommendation: leave the "Reflection" parameter at 255, 255, 255 and put in a valid Refractive Index. Alternatively, for human skin you could make use of the "Material Mix", if possible:

    - a "Glossy Material" with an Albedo/Diffuse texture, "Specular" 1, "Roughness" 0.3 and an "Index of refraction)" of 1.52, to mimic the stratum corneum outermost visible skin layer,

    - and a "Specular Material" with "Roughness" 0.3 (or just slightly lower, but not much) and an "Index of refraction" of 1.5 to mimic the sebum (thin oily film) on top of it.


    Also, basically says dielectrics like skin will be a shade of grey and human skin is at 0.028.

    https://seblagarde.wordpress.com/2011/08/17/feeding-a-physical-based-lighting-mode/

    Sébastien Lagarde's article (or NVIDIA's GPU GEMS 3) is a bit ambigous on this matter, since he uses the term "Specular Color" to explain "Specular Albedo/Specular Reflectance", where NVIDIA explaines the color of the specular reflection.

    In case for DAZ's Iray Uber, the "Specular Color" ("Glossy Color" on the Iray Uber) for a dielectric has always to be plain white, the "Specular Albedo/Specular Reflectance" is determined differently, depending on the "Base Mixing" mode used, and in a range between

    "Specular Reflectance" [F(0°) (or R0 in case of  Schlick's approximation)] is determined by either the "Glossy Reflectivity" parameter (on the PBR Metallicity/Roughness Mix), or the "Glossy Specular" parameter (on the PBR Specular/Glossiness Mix). Both are based on a material's Refractive Index, and unfortunately have to be pre-calculated manually.

    Using the formula mentioned on Legarde's Blog, and a Refractive Index of 1.52 (in vivo measured value for "Skin (Stratum Corneum)": [Tearney G. J., Brezinski M. E., Southern J. F., Bouma B. E., Hee M. R. and Fujimoto J. G.: “Determination of the refractive-index of highly scattering human tissue by optical coherence tomography.” Optics Letters, Volume 20, Issue 21, 1995, Pages 2258–2260]), you'll get a value of 0.0465210 0.042580 (*). That's more than 1.5 times larger than the "sample value" Legarde mentiones (0.028, corresponds to an IOR of 1.402). I'd recommend to always use measured values, and calculate the appropriate F(0°) or "Glossy Reflectivity" for them, above a general "sample value". Legarde's 1.4 looks like to be an average for the IOR's of the stratum corneum, the epidermis and dermis skin tissues... but the latter are more appropriate to determine SSS parameters than Specular Reflection, the stratum corneum and sebum are the two only things which can be viewed directly, and so have a direct impact on that.

    I made myself an OpenOffice calcsheet up a while ago to help calculate R0, IOR, Glossy Reflectivity and Glossy Roughness parameter values for use on the Iray Uber. Those who are interested could try it out. It's been made up with the german language version of OpenOffice (although the text is in english), I hope it will work on other language versions, too.


    (*) EDIT: Oops, sorry, wrong line of my crib sheet. The 0.0465210 is of course for an IOR of 1.55. blush

    Oh god I was right the first time. I need to stop questioning myself so much. I do currently use a mix material of specular/specular/diffuse. Don't like the glossy node for flesh. But thank you for this info, i will play with these settings
  • Rashad CarterRashad Carter Posts: 1,798
    edited September 2016

    Sorel,

    A few things

    1. As a fellow Octane user I find there is very little about the way Iray operates that translates very well to Octane. All the effort and collaboration that is going on with Iray users would possibly make an Octane user think that we too should be struggling to find the best solution. But alas, Octane is just a hundred times more straightforward than Iray, so you can get there much faster than with Iray. Yes, trust yourself. Not to say that you might not have room to improve, but to realize you are already on the right path. Or to put it another way; you've probably gone about as far as you can using these low resolution models and photocaptured textures. Applying your current skin settings as incomplete as you may think them now to one of those billion polygon real life geometry scans based on proper PBR texture inputs and you'd then see what is still lacking with any shader considerations on your own end. I'll bet you're already closer than you think. As you know I am a big fan of your skins. The primary thing I want to encourage you to do from here is to use much more bump instead of relying entire on specular effects and specular maps in general. Newborn babies have silky pearly smooth skin. But adults, just don't, no matter how pampered and priviledged. The simple act of smiling every day and eating food and looking around forces skin to have some "give" by the time we reach the age of 18. Pores give skin the ability to flex and move, as well as release perspiration. Faces shouldn't be porcelain is my point. Bump and normal mapping should be the inputs that "break up" the otherwise uniform specular. As in if the bump settings are correct, you can get away with quite high specular settings. I think it a mistake to tweak specular reflections without also carefully considering the bump and normal mapping that accompany it.

    2. I've said this before to several people but I think it bears repeating. I think we tend to make our own lives harder by trying to find the "perfect" solution for a given skin. That's why I keep encouraging users to use lots of different skins, or to skip skins altogether and use uniform skin-like colors while they study the effects of parameters like SSS. Because if the favorite skin you like to use already has a good deal of baked in SSS, it can mislead us if we assume other skins are operating the same way. Real skins have so many details such as wrinkles and pores and freckles that they distract us from seeing certain things clearly such as SSS, reflections, and bump. Honestly, with the right settings, a flat color can look as realistic as a photo based diffuse map provided the bump and specular mapping inputs are of the appropriate quality.

    3. In your specular/diffuse mix material, you should be using reflection for the shine on the skin, not the specular node on the diffuse part of the mix. And your Mix amount should strongly favor the specular part of the material, the diffuse should account for less than 20% of the total result. My own skins use 90% specular and 10% diffuse. While some people have found success making skins with only a Specular node, to me there is not quite enough control. The added Diffuse node is needed as a tweak to finish it off.

    4. Also Sorel, the exact SIZE of a light source and its proximity to a model will greatly affect the look of the specular/ reflections. Larger. wider light sources create very even looking soft diffuse lighting which "feels' realistic, but at the cost that the specular highlights can appear extremely flat. That same light source at half the size but with double the intensity would produce the same amount of total light but would do so in a more concentrated way resulting in a more pronounced specular reflections. Also, you are allowed to use settings above 1.0. If you need to raise the reflection to 2.0 or 85.0 no matter, just do it. Again, if you're breaking it up with the bumping, it will work well enough.

    5. I agree, the glossy node is not for skin, because skin has depth and the glossy node cannot account for that. But glossy is exactly the right answer for plant leaves with are paper thin.  Mix material Diffuse/Glossy will give much better results than specular/ diffuse for plants.

    6. Iray doesn't allow users to mix materials in the way Octane does. And if used properly Iray doesn't need them. But alas, Iray is so odd to me, just can't get with it.

    7. There are exceptions to the rule of human skin reflectance being fully white, such as make-up coverage. Any make-up with metalic elements is going to have a color tinted specular. There are also powders which have metallic properties, and several lipsticks as well.

    Post edited by Rashad Carter on
  • SorelSorel Posts: 1,389

    Sorel,

    A few things

    1. As a fellow Octane user I find there is very little about the way Iray operates that translates very well to Octane. All the effort and collaboration that is going on with Iray users would possibly make an Octane user think that we too should be struggling to find the best solution. But alas, Octane is just a hundred times more straightforward than Iray, so you can get there much faster than with Iray. Yes, trust yourself. Not to say that you might not have room to improve, but to realize you are already on the right path. Or to put it another way; you've probably gone about as far as you can using these low resolution models and photocaptured textures. Applying your current skin settings as incomplete as you may think them now to one of those billion polygon real life geometry scans based on proper PBR texture inputs and you'd then see what is still lacking with any shader considerations on your own end. I'll bet you're already closer than you think. As you know I am a big fan of your skins. The primary thing I want to encourage you to do from here is to use much more bump instead of relying entire on specular effects and specular maps in general. Newborn babies have silky pearly smooth skin. But adults, just don't, no matter how pampered and priviledged. The simple act of smiling every day and eating food and looking around forces skin to have some "give" by the time we reach the age of 18. Pores give skin the ability to flex and move, as well as release perspiration. Faces shouldn't be porcelain is my point. Bump and normal mapping should be the inputs that "break up" the otherwise uniform specular. As in if the bump settings are correct, you can get away with quite high specular settings. I think it a mistake to tweak specular reflections without also carefully considering the bump and normal mapping that accompany it.

    2. I've said this before to several people but I think it bears repeating. I think we tend to make our own lives harder by trying to find the "perfect" solution for a given skin. That's why I keep encouraging users to use lots of different skins, or to skip skins altogether and use uniform skin-like colors while they study the effects of parameters like SSS. Because if the favorite skin you like to use already has a good deal of baked in SSS, it can mislead us if we assume other skins are operating the same way. Real skins have so many details such as wrinkles and pores and freckles that they distract us from seeing certain things clearly such as SSS, reflections, and bump. Honestly, with the right settings, a flat color can look as realistic as a photo based diffuse map provided the bump and specular mapping inputs are of the appropriate quality.

    3. In your specular/diffuse mix material, you should be using reflection for the shine on the skin, not the specular node on the diffuse part of the mix. And your Mix amount should strongly favor the specular part of the material, the diffuse should account for less than 20% of the total result. My own skins use 90% specular and 10% diffuse. While some people have found success making skins with only a Specular node, to me there is not quite enough control. The added Diffuse node is needed as a tweak to finish it off.

    4. Also Sorel, the exact SIZE of a light source and its proximity to a model will greatly affect the look of the specular/ reflections. Larger. wider light sources create very even looking soft diffuse lighting which "feels' realistic, but at the cost that the specular highlights can appear extremely flat. That same light source at half the size but with double the intensity would produce the same amount of total light but would do so in a more concentrated way resulting in a more pronounced specular reflections. Also, you are allowed to use settings above 1.0. If you need to raise the reflection to 2.0 or 85.0 no matter, just do it. Again, if you're breaking it up with the bumping, it will work well enough.

    5. I agree, the glossy node is not for skin, because skin has depth and the glossy node cannot account for that. But glossy is exactly the right answer for plant leaves with are paper thin.  Mix material Diffuse/Glossy will give much better results than specular/ diffuse for plants.

    6. Iray doesn't allow users to mix materials in the way Octane does. And if used properly Iray doesn't need them. But alas, Iray is so odd to me, just can't get with it.

    7. There are exceptions to the rule of human skin reflectance being fully white, such as make-up coverage. Any make-up with metalic elements is going to have a color tinted specular. There are also powders which have metallic properties, and several lipsticks as well.

    Yeah are defintiely right about daz models/textures.  I had a feeling the specular was right before, I think it is the lack of correct bump/normals/displacement which is where I find it might be hard to get to look good.  I'm going to download one of those high res face models and see what results I get.

  • Sorel said:

    Oh god I was right the first time. I need to stop questioning myself so much. I do currently use a mix material of specular/specular/diffuse. Don't like the glossy node for flesh. But thank you for this info, i will play with these settings

    Well, if that's doable, then I'd say keep your current layout. I only read that the "Mix Material" setup would allow up to two different materials to be used and assumed, that this would be the upper limit.

  • Arnold C said:
    Sorel said:

    Oh god I was right the first time. I need to stop questioning myself so much. I do currently use a mix material of specular/specular/diffuse. Don't like the glossy node for flesh. But thank you for this info, i will play with these settings

    Well, if that's doable, then I'd say keep your current layout. I only read that the "Mix Material" setup would allow up to two different materials to be used and assumed, that this would be the upper limit.

    Mix Material only allow two inputs at a time with an "Amount" slider that controls the wieght of the influence. So if you have a Specular/Diffuse Material, then the Mix Amount will default to 0.5. In this case "0" would shift the influence entirely toward the Diffuse node while a setting of 1.0 would lean the influence entirely toward the Specular end of the material. However, the Mix Amount itself can be controlled with a slider or with image maps, you can even do a mix multiple image maps. So in some sone, one can say that a Mix is a type of node that can appear anywhere.

    As far as I can tell, with Octane's node based system, there are no limits to the number of nodes one can introduce and the way they can be combined. It's amazing to me how stable Octane can be while still allowing users to do all these crazy things.

    While the pricing of Octane might seem prohibitive when compared with free Iray available in DS, Iray simply doesn't replace the ease of use that comes with Octane. What I'm saying is that Octane offers clear advantages to Iray, though both platforms are obviously capable of alot. Octane is truly truly FUN to use. I've yet to hear Iray desribed as being fun by anyone. Octane is worth every penny, even if you're comfy with Iray.

  • Arnold CArnold C Posts: 740
    edited September 2016

    Mix Material only allow two inputs at a time with an "Amount" slider that controls the wieght of the influence. So if you have a Specular/Diffuse Material, then the Mix Amount will default to 0.5. In this case "0" would shift the influence entirely toward the Diffuse node while a setting of 1.0 would lean the influence entirely toward the Specular end of the material. However, the Mix Amount itself can be controlled with a slider or with image maps, you can even do a mix multiple image maps. So in some sone, one can say that a Mix is a type of node that can appear anywhere.

    As far as I can tell, with Octane's node based system, there are no limits to the number of nodes one can introduce and the way they can be combined. It's amazing to me how stable Octane can be while still allowing users to do all these crazy things.

    While the pricing of Octane might seem prohibitive when compared with free Iray available in DS, Iray simply doesn't replace the ease of use that comes with Octane. What I'm saying is that Octane offers clear advantages to Iray, though both platforms are obviously capable of alot. Octane is truly truly FUN to use. I've yet to hear Iray desribed as being fun by anyone. Octane is worth every penny, even if you're comfy with Iray.

    Thanks, Rashad. I just discovered that Otoy even offers a demo version of the DAZ Studio plugin. So I gave it a test ride.

    To (6.) from your posting above, I partially disagree. Iray does, too, you can mix materials and stack layers to your heart's content. Through MDL scripting. You can even mix your own custom shader within DS's Shader Mixer. But you're right, Octane's user interface offers a more user-friendly solution. DS's drawback is the lack of a comprehensive and recent User Manual for the Shader Mixer... and other tools. But that's the "same old, same old"...

    To (2.): the main big problem is the need to find a solution for the use of those old RDL texture assets which aren't made with PBR in mind. And as those are even different from creator to creator, you have to find more than one workaround. Unfortunately, the bad habit of "cooking the own soup" of the "good ole" 3Delight times, instead of first developing a single fitting method to solve things and follow that to an extent, is still existent.

    Compairing Otoy's Octane and DAZ's Iradium, they both work on the same PBR basics, and in the range from 0.00 to 1.00 in linear space. Octane's a bit more straight forward, and it doesn't seem to use fancy "hidden", and only sparse documented, methods like those based on Disney's "principled" BRDF as Iradium does (but that's DAZ, not NVIDIA). Like "Squaring Roughness/Glossiness", or using a normalized range to determine the incident specular amount ("Glossy Reflectivity") instead of an explicit index-of-refraction. Octane only uses the IOR to determine specular reflection. Plain, simple, comprehensible.

    Octane also seems to be more hardware friendly. At first I thought there's something broken. When monitoring "Memory Used" during rendering with GPU-Z, it didn't show any increasement, although the "GPU Load" showed the GPU's doing some work here. Was a bit irritating at the beginning. Another "pro" are the "System" and the "Textures" tab. Having the ability to monitor the VRAM usage without any fancy third party stuff is very comfortable. If the readings there are correct, Octane's less a "resource hog" than Iray is one.

    "Cons". Yes there are, IMHO, some: everthing is squished into that little "OctaneRender" tab, and adjusting materials can't be comprehended on Studio's viewports. Maybe I'm spoiled, but I like to have a glimps on how things would look when putting something together. Applying and adjusting materials is IMO a bit less comfortable.

    Post edited by Arnold C on
  • Arnold C said:

    Mix Material only allow two inputs at a time with an "Amount" slider that controls the wieght of the influence. So if you have a Specular/Diffuse Material, then the Mix Amount will default to 0.5. In this case "0" would shift the influence entirely toward the Diffuse node while a setting of 1.0 would lean the influence entirely toward the Specular end of the material. However, the Mix Amount itself can be controlled with a slider or with image maps, you can even do a mix multiple image maps. So in some sone, one can say that a Mix is a type of node that can appear anywhere.

    As far as I can tell, with Octane's node based system, there are no limits to the number of nodes one can introduce and the way they can be combined. It's amazing to me how stable Octane can be while still allowing users to do all these crazy things.

    While the pricing of Octane might seem prohibitive when compared with free Iray available in DS, Iray simply doesn't replace the ease of use that comes with Octane. What I'm saying is that Octane offers clear advantages to Iray, though both platforms are obviously capable of alot. Octane is truly truly FUN to use. I've yet to hear Iray desribed as being fun by anyone. Octane is worth every penny, even if you're comfy with Iray.

    Thanks, Rashad. I just discovered that Otoy even offers a demo version of the DAZ Studio plugin. So I gave it a test ride.

    To (6.) from your posting above, I partially disagree. Iray does, too, you can mix materials and stack layers to your heart's content. Through MDL scripting. You can even mix your own custom shader within DS's Shader Mixer. But you're right, Octane's user interface offers a more user-friendly solution. DS's drawback is the lack of a comprehensive and recent User Manual for the Shader Mixer... and other tools. But that's the "same old, same old"...

    To (2.): the main big problem is the need to find a solution for the use of those old RDL texture assets which aren't made with PBR in mind. And as those are even different from creator to creator, you have to find more than one workaround. Unfortunately, the bad habit of "cooking the own soup" of the "good ole" 3Delight times, instead of first developing a single fitting method to solve things and follow that to an extent, is still existent.

    Compairing Otoy's Octane and DAZ's Iradium, they both work on the same PBR basics, and in the range from 0.00 to 1.00 in linear space. Octane's a bit more straight forward, and it doesn't seem to use fancy "hidden", and only sparse documented, methods like those based on Disney's "principled" BRDF as Iradium does (but that's DAZ, not NVIDIA). Like "Squaring Roughness/Glossiness", or using a normalized range to determine the incident specular amount ("Glossy Reflectivity") instead of an explicit index-of-refraction. Octane only uses the IOR to determine specular reflection. Plain, simple, comprehensible.

    Octane also seems to be more hardware friendly. At first I thought there's something broken. When monitoring "Memory Used" during rendering with GPU-Z, it didn't show any increasement, although the "GPU Load" showed the GPU's doing some work here. Was a bit irritating at the beginning. Another "pro" are the "System" and the "Textures" tab. Having the ability to monitor the VRAM usage without any fancy third party stuff is very comfortable. If the readings there are correct, Octane's less a "resource hog" than Iray is one.

    "Cons". Yes there are, IMHO, some: everthing is squished into that little "OctaneRender" tab, and adjusting materials can't be comprehended on Studio's viewports. Maybe I'm spoiled, but I like to have a glimps on how things would look when putting something together. Applying and adjusting materials is IMO a bit less comfortable.

    Arnold,

    I'm glad you had a chance to play with Octane. Otoy really is doing its best to produce a first rate product. However Iray still has Octane beat for most users around here because it is free, and with a little help from friends like you to unravel the knot; extremely capable. Having both CPU and GPU render options is itself worth several hundred dollars, and the full level of integration with DS makes it even more valuable. Octane implementations in each application are different.

    Would it be feasible to request of Daz to perhaps change the way the problematic inputs such as Glossy and Reflection are presented in the next version of DS? Simplifying it? There could easily be a conversion algorithm to maintain consistency between new and old scenes.

    I would really like to know why it was decided to implement these options as they have been presented in the DS version of Iray.

  • Arnold CArnold C Posts: 740
    edited September 2016

    Arnold,

    I'm glad you had a chance to play with Octane. Otoy really is doing its best to produce a first rate product. However Iray still has Octane beat for most users around here because it is free, and with a little help from friends like you to unravel the knot; extremely capable. Having both CPU and GPU render options is itself worth several hundred dollars, and the full level of integration with DS makes it even more valuable. Octane implementations in each application are different.

    Would it be feasible to request of Daz to perhaps change the way the problematic inputs such as Glossy and Reflection are presented in the next version of DS? Simplifying it? There could easily be a conversion algorithm to maintain consistency between new and old scenes.

    I would really like to know why it was decided to implement these options as they have been presented in the DS version of Iray.

    Yes, that was a nice experience. Another thing I forgot to mention is a better displacement solution. The display of the estimated render duration, memory usage and GPU temperature on the Viewport is also a nice thing to have. I'd guess, with a more reasonable licencing plan, f.e. an ability to purchase only the thing you want when you'd never make use of the Standalone, would Octane most likely grant a whole lot of more friends.

    I guess the main reason was to implement a more intuitive, and so more artist-friendly solution, given that the artist knows what he/she is doing. In the "irayubermatrial.mdl" it's mentioned that it would be "based on the IRay material for 3DS MAX", so they wouldn't be the only ones using some of Disney's methods.

    They aren't that bad, but require to know that something like that is used in the first place to make the most of it... and avoid confusion. smiley 

    Let's see how Iray evolves, they just have started, and their most recent build has some more nice and new features to offer.

    Post edited by Arnold C on
  • CPU-Z and GPU-Z need permission to run, presumably because at least some of the information they report requires privleged access. I wouldn't want to have to do that with DS to get the temperatures in the UI.

Sign In or Register to comment.