Fiddling with Iray skin settings...

1565759616291

Comments

  • fasttam said:

    Iray for Maya

    "Also available is the first edition of vMaterials, a free set of digital material offerings verified for accuracy, control and consistency. With vMaterials, designers no longer have to spend time sourcing and creating their own materials."

     

    Yep, I got those downloaded and there are some pretty nice ones. No skin yet, but it was because of this set of materials that I greatly suspect NVIDIA may be working on one.

  • SzarkSzark Posts: 10,634
    edited December 2015
    AndyGrimm said:

    @Szark .. yes.. the sunsky does simulate lightcolor based on daytime...

    this is the reason why i always said - the "brutal" sun sky test at 12 midday.. (where a photographer normaly would NOT take outside portraits).. is the best test for a skin....  the resulting render is NOT a charming one - same as in reality.. but it is colorneutral to HDRis....  if red goes to much saturated there.... the skinsettings are far off.

    and that is what I have found so far. Yes I am still playing about with this on and off when the feel takes me but since I don't have the content I don't feel inspired to make a super realistic image, or learn to be honest. Instead what I am doing is to see what is needed to get better skins maps looking better in Iray. Which functuons do I change dpending on the colour saturation of diffuse maps. So far I have gotten consistant resukts with keeping the SSS colour a medium staturated Orange/Red with every M4 and M5 skins set I have tested. What I have also found is to change the Diffuse and SSS Tint colours to offset the redness iof the maps. These colours differ between sets but most of the time I am just dropping the red channel keeping the colour saturated as possible. 

    This is Genesis M5 Dave (Eye maps from Actual Eyes for Genesis). Adjusted and converted bump maps in to Normal and used both Bump and Normal maps, Bump in the Top Coat Bump Channel. No Gloss Wieght or Roughness maps tested.

     

    PBR Metalicity?Roughness Shader for everything

     

    Diffuse Colour for this image pure white with diffuse maps. All diffuse maps set to Gamma 1.80 via the Image Editor inside Daz Studio.

    I used 0.60 Transluceny Weght

    Scatter and Transmit

    Diffuse Maps in the Transluceny Colour

    Drop the red down in the SSS Reflectance Tint (For this image 0.87, 1.00, 1.00 (RGB 240, 255, 255))

     

    Gloss Weight 0.50

    Roughness 0.45

     

    Now I am not sure if this is needed

    Base Thin Flim about 1500000.00 depending on which part of the skin it is for.

    Base Thin Flim IOR 1.44

     

    Top Coat Weight 0.40

    Top Coat Color 1.00.0.96, 0.83 (RGB 255,250, 234)

    Scatter and Transmit

    Top Coat Roughness 0.70

    Top Coat IOR 1.46

    Top Coat Thin Film 2000.0

    Top Coat Thin Film 1.46

     

    SSS settings from the top.

    Transmitted Measurement Distance 2.00

    Colour 0.71, 0.032, 0.0037 (RGB 218, 53, 20)

    Scattering Measurement Distance 0.25

    SSS Amount 1.67

    SSS Direction -0.50

    HDRI copy.jpg
    1414 x 2000 - 864K
    Sun copy.jpg
    1414 x 2000 - 1005K
    Post edited by Szark on
  • Szark, the Hue is different between them, just looking at the non-skin stuff. 70 vs 26 on the silver chain on the left of the image (right shoulder).

    TestPoint1_001.png
    1287 x 787 - 686K
    TestPoint1_002.png
    1314 x 805 - 618K
  • SzarkSzark Posts: 10,634

    Yep  like what was said earlier the Iray Sun is a lot warmer than the included HDRI which is something people need to take in account when setting up skin, as already mentioned. I was just showing what that contrast between them looks like when you only change one thing, the light source. Also you have to take in to consideration that these were rendered as PNGs and saved as compressed JPG's for the forum.

  • ZarconDeeGrissomZarconDeeGrissom Posts: 5,412
    edited December 2015

    Still, (as I'm sampling more points between them), JPG from PNG would not cause that much change in hue. saturation and 'Value' (brightness) possibly, not drastic color, lol.  174 vs 216 one chain link over. The blue shirt also has a drastic change in apparent color.  Just call it a 'Hollywood lighting trick' to change the color of the set, lol.

    Forgot to say, good looking skin settings there.

    Post edited by ZarconDeeGrissom on
  • SzarkSzark Posts: 10,634

    Yeah a big difference in Hue and that justs adds to how important it is to have a base to test in and that we are assuming the Sun/Sky is as real as they say Iray is. :)

     

    Thanks yeah I think it came out well.

  • SzarkSzark Posts: 10,634

    I am an idiot sometimes, the Sun/Sky I used is the default. I am doing another with the sun higher in the sky from behind to show the SSS for the ears.

  • kyoto kidkyoto kid Posts: 41,260
    edited December 2015

    pearbear, those are incredible renders. The kind I would expect from a supercomputer, or a film Production computer cluster.  Unfortunately, most of us do not have such luxuries. And as such, we have to make do with what looks best with minimal compute resources.

    I'll say it again. I would love to have the hair on the girl in kyoto kid's sig, even tho I would never be able to render it. There is one of me on a single computer, not an entire division of Pixar, lol.  Now looking back, I've seen some incredible results using the incredibly limited Daz Default shader, and it dose not even have Sub-dermis whatever. That was baked into the Diffuse and Specular maps. Velvet, what is that, not in the Daz Default shader, lol.

    . ...that hair took Pixar nearly three years to create., including the programming of proprietary software and development of hardware powerful enough to accomplish the task.  I came close, but no where near as well detailed as the actual character.

    For my version of the character, I used two instances of 3Dream Mairy's Bolina Hair, the colours of each just being slightly different from each other to try and create a sense of more depth.

    merida new test 2.png
    900 x 900 - 1M
    Post edited by kyoto kid on
  • @Szark

    "Drop the red down in the SSS Reflectance Tint (For this image 0.87, 1.00, 1.00 (RGB 240, 255, 255))"

    are you still using 4.8 ?

  • SzarkSzark Posts: 10,634

    nope latest 4.9 beta 

  • @Szark

    then you should be able to ignore tint (default white) and change transmitted color (bring in more blue and green) instead...

    But i like your render 1....   thin film dosent make sense to me on skin.. and your glossiness settings could maybe improve without... other then that.. nice result.

  • SzarkSzark Posts: 10,634
    edited December 2015

    and this is what it looks like with the sun higher and behind

     

    only let it render for an hour

    sunhighbehind.jpg
    1414 x 2000 - 1M
    Post edited by Szark on
  • SzarkSzark Posts: 10,634
    AndyGrimm said:

    @Szark

    then you should be able to ignore tint (default white) and change transmitted color (bring in more blue and green) instead...

    But i like your render 1....   thin film dosent make sense to me on skin.. and your glossiness settings could maybe improve without... other then that.. nice result.

    I was thinking what Arnold and I were discussing a month ior so back about sebum having ;, a thickness, b: a colour, c; an IOR and lastly fresnel properties I have found that light bends better around the curves of the skin with this thin flim top coat than without. Base thin film; I haven't found if it does make a difference or not. I need to reset that a render a few shots.

  • SzarkSzark Posts: 10,634

     I have just added this to my test render post

    "All diffuse maps set to Gamma 1.80 via the Image Editor inside Daz Studio."

  • AndyGrimmAndyGrimm Posts: 910
    edited December 2015

    When you look on the forhead in the backlight render - there are this not so nice "light bleeding" areas from SSS.....    i guess this coming from using high red  and the other colors very low in TMC and to much red saturation and luminance (gamma 1.8 correcture) in the base texture....  but you are very close to a nice skin setting which works in all HDRis...

    Post edited by AndyGrimm on
  • jaq's suggestion about using plus scatter instead minus - solved many similar problems in my case...

  • SzarkSzark Posts: 10,634

    Yeah I need to do a series of renders with testing one thing at a time. like + or - SSS Direction and see if it makes any difference.

  • mmkdazmmkdaz Posts: 335

    @Szark, You have amazing results! Thank you.

    @fasttam

    Is the process for installling the vmaterials to Daz easy? If so, what is the process?

  • SotoSoto Posts: 1,440
    edited December 2015
    fasttam said:

    Iray for Maya

    "Also available is the first edition of vMaterials, a free set of digital material offerings verified for accuracy, control and consistency. With vMaterials, designers no longer have to spend time sourcing and creating their own materials."

    Can those be translated to DAZ Studio somehow or how do they work?

    Post edited by Soto on
  • MBuschMBusch Posts: 547
    edited December 2015
    AndyGrimm said:

    i come from the advertising and photoindustry....   (i dont say anymore, "thats how it is" here laugh).....  we used over decades for neutral  colors  (photo neutral) 5k.... for Photo and Print..
    5 k is horizontal daylight  5.5k vertical (sky)....   photolamps do usally 4.8 - 5,5k  because of this.

    That is in USA and mainly in print industry. SWOP USA print standard is a little bit more cyan than European Euroscale/FOGRA print standard which uses D65 and is a little bit more magenta. Film and video industry uses D65. Anyway, I think the real reason is the large adoption of sRGB as standard RGB color space which uses D65 as reference. From wikipedia: "The sRGB color space is well specified and is designed to match typical home and office viewing conditions, rather than the darker environment typically used for commercial color matching."

    There are and interesting note in wikipedia: "The sRGB specification assumes a dimly lit encoding (creation) environment with an ambient correlated color temperature (CCT) of 5000 K. It is interesting to note that this differs from the CCT of the illuminant (D65). Using D50 for both would have made the white point of most photographic paper appear excessively blue.[5] The other parameters, such as the luminance level, are representative of a typical CRT monitor. For optimal results, the ICC recommends using the encoding viewing environment (i.e., dim, diffuse lighting) rather than the less-stringent typical viewing environment."

    https://en.wikipedia.org/wiki/SRGB

    Oh! I think you will found this just more interesting.

    https://en.wikipedia.org/wiki/Adobe_RGB_color_space

    Post edited by MBusch on
  • AndyGrimmAndyGrimm Posts: 910
    edited December 2015

    @MBusch

    yes the story about Adobe RGB, the incorrect profile, confused the whole graphic and print sector .. that was still with analog film as source and so we used "film" scanner profiles and everybody which could afford,  invested in  CMYK  raster thermosublimations printer for pre-print proofs  and developed their own custom profiles....that was the only way to get accurate results ... memories smiley

    DAZ  spot 6.5k is not a mistake but it is good to know that for adjusting a natural reacting skin it is best set to 5k because of  "natural whitepoints" in HDRi's... 3 - 7k is about the range in normal daylight photos so 5k is the middle (another simple reason)

     

    Post edited by AndyGrimm on
  • kyoto kidkyoto kid Posts: 41,260
    edited December 2015

    ...for what it's worth, more messing around with the Josie 7 skin settings. Had to edit the face map to make the eyebrows match.. Hair is Koz's Messy Hair with a Iray shader from Mec4D's Unshaven.

     

    Tracey J7.png
    636 x 900 - 630K
    Post edited by kyoto kid on
  • SzarkSzark Posts: 10,634
    magnumdaz said:

    @Szark, You have amazing results! Thank you.

    Thank you, doing some more testing and there is a slight difference in plus V minus SSS direction, after dinner I post the results.

  • Hellboy said:

    Can those be translated to DAZ Studio somehow or how do they work?

    They are .mdl materials. Maybe someone knows how to convert them to .duf.

  • agent unawaresagent unawares Posts: 3,513
    edited December 2015
    fasttam said:
    Hellboy said:

    Can those be translated to DAZ Studio somehow or how do they work?

    They are .mdl materials. Maybe someone knows how to convert them to .duf.

    Drag the MDL file into your Shader Mixer pane to open it. There's a moment while it compiles so be patient.

    EDIT: Tried a couple of the gemstone shaders and DS is just treating them like they're a weird black rubber/glass hybrid material.

    EDIT (EDIT EDIT): Never mind that, it was a scale issue. Only looks right on very small objects.

    shader mixer.jpg
    679 x 414 - 100K
    Post edited by agent unawares on
  • RAMWolffRAMWolff Posts: 10,256

    So studied SickleYield's info on lighting.  Not sure but I think I'll have to go back over the skin settings now.   Looks a bit saturated and a bit more tan that I really wanted but this is using ONE mesh light.  The settings are at the bottom of the image.  The top settings are from the Render Settings> Tone Mapping

     

     

    Lighting Tests using Sunny16 - 13000000.00 lms - 03.jpg
    938 x 917 - 355K
  • KhoryKhory Posts: 3,854

    I gather that your creating a mesh light rather than using a photometric light. I don't really get the logic in that and since it is so much easier to work with the photometrics and I honestly think they are faster than emissive lighting in most cases but all of that could be me. At any rate what color did you set the light up as and did you start with the Iray uber emissive preset because something is seriously off with it currently.

  • RAMWolffRAMWolff Posts: 10,256

    Hi,

    The Emmision color is pure white (1.00 - 1.00 - 1.00)

  • RAMWolffRAMWolff Posts: 10,256

    OK.... studied a bit more.  Found that the Emission Temperature plays into how much red plays into the skin tones.  So new render found a sweet spot with the three settings. 

    As you can see though, under the settings box I inserted into the image there is that strange lightness but this time it came out as a weird box shape so wondering if the lighting is playing tricks along with the bad code in 4.8 around this issue.  I really hope this is cleared up with 4.9.  I've tried changing quite a few, if not all of the settings for the skin and lights and can't seem to figure out how to make that weird bit go away with out sacrificing the quality of the skin settings or the lighting! 

     

     

    Lighting MUCH better.jpg
    938 x 917 - 186K
  • FishtalesFishtales Posts: 6,162
    edited December 2015

    The setting for the light temperature at 1940 is red. Temperatures below 5500 are in the red end of the spectrum and above are in the blue end.

    https://en.wikipedia.org/wiki/Color_temperature

    Post edited by Fishtales on
Sign In or Register to comment.