Loading...
> >

Fiddling with Iray skin settings...

1…6970727475…91

Comments

  • Arnold CArnold C Posts: 740
    RAMWolff said:

    So basically what's being said here is to ignore the SSS parameters settings (SSS Amount and SSS Direction... any others??) unless you want to mask what's going on under the Translucency weight paremter (i.e. taking the glare off of eyebrows, facial hair, ect..)???

    Uh, a definite no to that, if you have set "Thin Walled" to Off, you should use all of the "Volume" group parameters (Transmitted Measurement Distance, Transmitted Color, Scattering Measurement Distance, SSS Amount and SSS Direction) to make your material look as real as possible.

    If you're using a Translucency Color instead a Diffuse map in the "Translucency Color" texture slot (you need translucency for skin to make the "Volume" group parameters work), you should also better use a texture map there to mask out eyebrows, hairstubs, etc to avoid those being tinted by the Translucency Color as well. Depending on the Translucency Color used especially eyebrows will look from somewhat odd to completely stupid in the final render.

    Don't use any texture map, be it grayscale, color or anything else, on the "Translucency Weight" parameters' texture slot, because a texture map there will block your Translucency Color having an effect. [Looks like Iray expects a float there, and with a texture in it it doesn't know what to do next.]

  • RAMWolffRAMWolff Posts: 10,256

    Thanks for the tips.  Hmmmm, doesn't seem to matter what I do I just can't seem to wrap my head fully around all this!  lol  There was a tutorial on understanding all this better but I had to return it as it was written in Spanish.  My Bruna character for Dawn (Hivewire) is a tanned Italian woman.  I've studied other settings for other characters but there is just something off about my settings.  regardless of what I try it seems.  I know there is a fine balance for some settings but being from the 3Delight camp TRYING to understand this as quickly as possible just isn't panning out.  I can release her with very basic settings but I'd feel badly as I know she could look better! 


    indecision

  • KhoryKhory Posts: 3,854

    Iris size is adjustable so that shouldn't be a factor in the discussion... unless someone were to do a replacement eye and not include that morph. In fact any replacement eye would need to have all of the morphs that the eye already has including things like cat eye pupils. Even if the eyeball is hyper realistic it still needs to be usable with fantasy type figures since they are rendered in Iray as well.

  • AndyGrimmAndyGrimm Posts: 910
    Khory said:

    Iris size is adjustable so that shouldn't be a factor in the discussion... unless someone were to do a replacement eye and not include that morph. In fact any replacement eye would need to have all of the morphs that the eye already has including things like cat eye pupils. Even if the eyeball is hyper realistic it still needs to be usable with fantasy type figures since they are rendered in Iray as well.


    A correct scaled cornea does not fit in the eye when the iris is to big...the cornea would cut into the eyelids/lashes..  this is one of the limitations of a ultra realistic eye. A way around this problem could be a minus bulge cornea morph smiley But i am agreed with what you saying... the eye should work also for fantasy types.. 

  • AndyGrimmAndyGrimm Posts: 910
    edited March 2016
    PA_ThePhilosopher said:

    Andy,
    Perhaps the eyes can be improved without having to alter any geometry.
    The goal, it seems, is to produce a softer transition between the sclera and iris, which I think can be done using an oppacity map. 
    Here is something I put together last night using the base G3F model (see attachment). The only thing I altered was the face and eye maps. Let me know what you think;

    i think you are up for a good start with your settings.. very nice renders

    the cornea is not clear down to the edge of the iris... that's why i used a very small refraction ring gradient mask there....  
    a iris displacement map gives more surfaces with the right angle to the caustic glow....  the glow is opposite to the light entrance (reflection) this is to less to see now. do you have caustic on in your render settings?

    Post edited by AndyGrimm on
  • PA_ThePhilosopherPA_ThePhilosopher Posts: 1,039
    edited March 2016

     

    nonesuch00 said:

    Well, I can't speak to the technical things but the model looks very good except the irises are too deep a blue, unless they are wearing blue contact lenses. The irises themselves look one or maybe two millimeters too big in total diameter.

    I hope my attempt at realism is as good as yours when I get that far.

    Nonesuch00,

    Took your advice. I made a few more adjustments to the base G3F eyes for my experiment; reduced the iris size and sclera mats, toned down the blue, etc.
    not sure yet though....cornuculas are a bit tricky.... I'm still new to Daz myself...so much to learn.

    -P

     

    test6.jpg
    680 x 748 - 331K
    test6b.jpg
    680 x 748 - 393K
    Post edited by PA_ThePhilosopher on
  • PA_ThePhilosopherPA_ThePhilosopher Posts: 1,039
    edited March 2016

     

    AndyGrimm said:

    i think you are up for a good start with your settings.. nice

    the cornea is not clear down to the edge of the iris... that's why i used a very small refraction ring gradient mask there....  
    a iris displacement map gives more surfaces with the right angle to the caustic glow....  the glow is opposite to the light entrance (reflection) this is to less to see now. do you have caustic on in your render settings?

    Andy,
    Caustics is turned off for this model. I tend not to use them in order to keep overhead down to a minimum.
    I need to add more moisture; I think that will help improve realism significantly, as it does in this maya/zbrush model.
    Also notice the darkening of the sclera as it nears the cornucula.

    -P

    9.jpg
    1200 x 1200 - 313K
    Post edited by PA_ThePhilosopher on
  • nonesuch00nonesuch00 Posts: 18,333
    PA_ThePhilosopher said:

     

    nonesuch00 said:

    Well, I can't speak to the technical things but the model looks very good except the irises are too deep a blue, unless they are wearing blue contact lenses. The irises themselves look one or maybe two millimeters too big in total diameter.

    I hope my attempt at realism is as good as yours when I get that far.

    Nonesuch00,

    Took your advice. I made a few more adjustments to the base G3F eyes for my experiment; reduced the iris size and sclera mats, toned down the blue, etc.
    not sure yet though....cornuculas are a bit tricky.... I'm still new to Daz myself...so much to learn.

    -P

     

    Those are are much more realistic looking although I've seen people wwith blue eyes and blue contacts so the first coloring is possible but usually results in people asking if there eyes are really that blue. The size looks OK now too although they look to be the size size as earlier, maybe the deep blue was making them seem bigger to me.

  • AndyGrimmAndyGrimm Posts: 910
    edited March 2016
    PA_ThePhilosopher said:

     

    AndyGrimm said:

    i think you are up for a good start with your settings.. nice

    the cornea is not clear down to the edge of the iris... that's why i used a very small refraction ring gradient mask there....  
    a iris displacement map gives more surfaces with the right angle to the caustic glow....  the glow is opposite to the light entrance (reflection) this is to less to see now. do you have caustic on in your render settings?

    Andy,
    Caustics is turned off for this model. I tend not to use them in order to keep overhead down to a minimum.
    I need to add more moisture; I think that will help improve realism significantly, as it does in this maya/zbrush model.
    Also notice the darkening of the sclera as it nears the cornucula.

    -P

    yes - Super! great finetuning....   you added what i am working on myself (in the background - not public on a G3 yet smiley) Peach fuzz ? ....   
    I worked on fiber mesh hairs over the last days... and one thing is for sure.. Iray handels a high polygon count phantastic..  so here is your next product idea ...   smiley..

    Caustic should improve the light play on the iris a lot....  give it a try.. (looks as you can render fast enough laugh... rendered in Maya ? would be great to compear the same model/render with Daz Iray.

    Edit: I just noted that i might had missunderstood your example render as your own work ...  but still .. peach fuzz adds a lot to realism.

    Post edited by AndyGrimm on
  • PA_ThePhilosopherPA_ThePhilosopher Posts: 1,039
    edited March 2016

    Andy,
    Yes the second one is not my work. It is one of the reference images I have been using to design peach fuzz for Daz (do you think there would be a demand for that as a product?). I also have been using the eyes as refence. But Daz is capable is equally as stunning work. Here is another reference image I have been using (see attached). This one was done soley in Daz....

    daz3d_iray_skin_shader_settings_2_by_second_circle-d8trfv9.jpg
    1024 x 1029 - 196K
    Post edited by PA_ThePhilosopher on
  • KhoryKhory Posts: 3,854

    Good point Andy.

    "Yes the second one is not my work. It is one of the reference images I have been using to design peach fuzz for Daz (do you think there would be a demand for that as a product?). "

    I would think so.

     

  • AndyGrimmAndyGrimm Posts: 910
    edited March 2016

    yes - peach fuzz.. for close ups should sell i think....  different props: density. faceareas, maybe a lenght morph... if you are good with fibermesh there are many products missing for g3 right now.. Armpits, chesthairs, legs, arms.. and so on... members here asked for eyebrows just some days ago...   there are many possibilities.

    I just rendered a hairstyle with 120k strains and 12 segments on my poor 1gb card over the weekend.. (no textures are needed for good looking hairs in iray).... this convinced me even for full fiber hairstyles... (fibers make 3d delight rendering really slow.. but not so in iray)... 

    Post edited by AndyGrimm on
  • PA_ThePhilosopherPA_ThePhilosopher Posts: 1,039
    edited March 2016

    I've been pleased with the speed of iRay as well for full fiber hair. It seems to handle it just fine.

    Thankfully, because peach fuzz requires many fibers to be convincing. It good to know there would be interest for it. I'll continue working on it then. I've been working on eyebrows as well, though there are a bit more hurdles to jump over if it is to be released as a product (such as the removal of the base skin eyebrow prior to application. But it may be doable, possible scripting). 

    Here is one of my first [very rough] test renders...still have a lot more work yet....

     

    test7.jpg
    600 x 600 - 40K
    Post edited by PA_ThePhilosopher on
  • Arnold CArnold C Posts: 740
    edited March 2016
    PA_ThePhilosopher said:

    Here is another reference image I have been using (see attached). This one was done soley in Daz....

    Those Iray material setttings are a ("good") example of an artistic approach... but have less in common with physically based shading (PBS).

    - "Translucency Weight" is too low for a light caucasian skin. Should be a value between 0.40 and 0.50

    - "Glossy Color".  A common serious mistake: when too much glossiness occurs, many artists and even PA's tune their glossy colors down. Specular reflection (not to be confused with specular color, that's a different thing) from dielectric materials (non-metals) is always White, because only metals are able to change the coloring of specular reflection. Problems will begin when a colored light hits the material. When using the PBR Metallicity/Roughness mix, don't put any "Glossy Specular" values from the PBR Specular/Glossiness mix into "Glossy Color". In the PBR Metallicity/Roughness mix, the "Glossy Reflectivity" parameter will take care of the specular color. How to calculate the appropriate reflectivity value from a given IOR can be found on page 33 of this thread in AndyGrimm's formula collection.

    - "Top Coat Color": see "Glossy Color"

    - "Transmitted Color" in combination with "Transmitted Measurement Distance": again for a light caucasian skin way too high. The corresponding absorption coefficents for those settings will be at around 2.22/3.93/4.24, more in the range of african skin. For caucasian skin tones, try a RGB color of 204-144-110 in DAZ Studio's color-picker in combination with a "Transmitted Measurement Distance" of 1.00. These values are based on the average of real-world measurements.

    The formula how transmission and scattering are calculated in Iray can be found here.

    Post edited by Arnold C on
  • Arnold CArnold C Posts: 740
    edited March 2016
    RAMWolff said:

    Thanks for the tips.  Hmmmm, doesn't seem to matter what I do I just can't seem to wrap my head fully around all this!  lol  There was a tutorial on understanding all this better but I had to return it as it was written in Spanish.  My Bruna character for Dawn (Hivewire) is a tanned Italian woman.  I've studied other settings for other characters but there is just something off about my settings.  regardless of what I try it seems.  I know there is a fine balance for some settings but being from the 3Delight camp TRYING to understand this as quickly as possible just isn't panning out.  I can release her with very basic settings but I'd feel badly as I know she could look better! 

    You're welcome. smiley

    Yep, I almost bought this tut, too. In the last moment I was aware of it's drawback: Io no habla espanol. surprise 

    I own Dawn too (SR1 and SR2), but I'm not very fond of her. Her shape's too toonish for my taste. (Her male counterpart I find even worse, looks IMO way too feminine. There's no room for softies in my content library laugh).

    I've been able to have a look on the settings of her Iray materials from the SE, and like the Queen I'm not very amused. From what I've seen it's more done by the eye than based on anything PBS. If I remember correctly the creator even used the bump map also as specular map.

    PBS is IMO way less confusing than 3Delight shading. Anything's based on real-world measurements, so one key to success is to have some valid measurements at hand and the knowledge where to put them into.

    To learn the basics of PBS I'd recommend the articles Basic Theory of Physically-Based Rendering, Tutorial: Physically Based Rendering, And You Can Too!, and if you plan to work on textures Tutorial: PBR Texture Conversion. Sébastien Lagarde's Feeding a physically based shading model article is also highly recommendable.

    Post edited by Arnold C on
  • RAMWolffRAMWolff Posts: 10,256

    WOW... thanks so much Arnold.  Those links will, hopefully, prove useful!  Off to check them out! 

  • RAMWolffRAMWolff Posts: 10,256

    I see most of these are from the Toolbag pages.... so you use Toolbag Arnold?  I've been considering it for a while now... just curious about your experience with it in case you do indeed use it in your workflow...

  • Arnold CArnold C Posts: 740
    RAMWolff said:

    I see most of these are from the Toolbag pages.... so you use Toolbag Arnold?  I've been considering it for a while now... just curious about your experience with it in case you do indeed use it in your workflow...

    You're welcome. smiley

    Nah, haven't used it yet. If it would have a vast sample material library I would, if only for looting it for measured values. But there's nothing what NVIDIA's vMaterials doesn't offer. Since all materials there are in MDL format, it's much easier to get them into DS via Shader Mixer.

    The best thing about PBR is, unlike all the non-PBR renderer like 3Delight or Poser's FireFly, which all work on different principles, the basic principles of PBR never change, the only differences lie within the methods the renderer uses to calculate their results. A roughness value of 0.3 for a material will be the same regardless of the renderer in use. Since you said you're from the "3Delight camp" I guess you can tell a thousand stories about converting Poser materials to 3Delight. Me, I'm happy I don't have to do that anymore.

    Another good source of information is Allegorithmic’s Comprehensive PBR guide.

    Happy learning. smiley

  • AndyGrimmAndyGrimm Posts: 910
    edited March 2016

    Hairs:

    I googled a lot but cant find a document which describes hairs for  a PBR workflow.

    Specular for hairs should be in the 0.02 - 0.05 range - definitly hairs are not metal... but most PBR hair shaders use colored specular.. the better shaders even a primary and a secondary specular color.....

    Are hairs the big exception of physically correct shading? 

    Edit: The same applies also for velvet and satin... no metal.. but they are done with using specular color and backscattering color.. (often without a diffuse color at all!).... definitly not physically correct... are there any datas or measuerements of this kind of materials in the net? 


    Fiber hairs in IRAY...

    I struggle to control the direction of anisotropy on Fibermesh hairs...    it seems to have just a effect on a single hair and not over the whole hairstyle...  i guess it has to do with using or not using a UV ? .... any advise here? Or is this a limit in IRAY using only the real mesh (aka each single hair ) ?

    In other words: Is anisotropy direction driven by the mesh or by the UV ? How should a UV look for controlling this effect for fibers? Is somehwere a example to find ? (any PBR engine).

     

    Post edited by AndyGrimm on
  • SorelSorel Posts: 1,407
    Yeah I find it is hard to find info for things that aren't metal, skin, it glass. I did this image with "silk". Basically set it up as a metal and adjusted roughness and bump to desired levels. http://www.daz3d.com/gallery/#images/85666
  • AndyGrimmAndyGrimm Posts: 910
    edited March 2016

    @Sorel

    nice render... just gave you a thumbs up for the silk shader smiley

    yes....  there is a information gap for advanced/average PBR users.... the documents ether explain the basics using always the same materials (rost on metal - hey not everybody wants to create weapons for gamessmiley) or then go directly into formulas for shader coders..... 



     

    Post edited by AndyGrimm on
  • KhoryKhory Posts: 3,854

    That fabric looks fab Sorel!

  • AndyGrimmAndyGrimm Posts: 910
    edited March 2016
    evilded777 said:
    Arnold C. said:

    The "SSS" designation of the maps currently used on the Translucency Color texture slots is a bit misleading, since they don't have a direct impact on Subsurface Scattering (although you need a material to be translucent to have Subsurface Scattering happening). All they do is additionally darken the color set in the Translucency Color parameter and/or mask out parts where translucence shouldn't have an impact. A better fitting one could be "TM" (Translucency Mask) or "TCM" (Translucency Color Map)... if it would be one.

    Is it so much to ask to use nameing conventions that are correct/relevant? You wouldn't think so...but someone comes up with something and then everyone else runs with it whether its correct or not. Can be quite frustrating, not to mention down right confusing.

    I do nowhere see a problem with naming conventions in the postings above...

    there are just two point of views... those which split the effects by formulas exactly.. and those which see transparency as the sum of  SSS and TMC...  

    there is no SSS without translucency OR transparency (refraction)....  translucency weight drives the amount of SSS and TMC.  You want more SSS effect ? you have to go up with translucency weight...  as such the only thing which is confusing is that translucency weight and translucency color are in another section of the ubershader...  

    Post edited by AndyGrimm on
  • throttlekittythrottlekitty Posts: 173
    AndyGrimm said:

    Hairs:

    I googled a lot but cant find a document which describes hairs for  a PBR workflow.

    Specular for hairs should be in the 0.02 - 0.05 range - definitly hairs are not metal... but most PBR hair shaders use colored specular.. the better shaders even a primary and a secondary specular color.....

    Are hairs the big exception of physically correct shading? 

    Edit: The same applies also for velvet and satin... no metal.. but they are done with using specular color and backscattering color.. (often without a diffuse color at all!).... definitly not physically correct... are there any datas or measuerements of this kind of materials in the net? 


    Fiber hairs in IRAY...

    I struggle to control the direction of anisotropy on Fibermesh hairs...    it seems to have just a effect on a single hair and not over the whole hairstyle...  i guess it has to do with using or not using a UV ? .... any advise here? Or is this a limit in IRAY using only the real mesh (aka each single hair ) ?

    In other words: Is anisotropy direction driven by the mesh or by the UV ? How should a UV look for controlling this effect for fibers? Is somehwere a example to find ? (any PBR engine).

     

    Some types of hair product use metals/minerals for tinting as part of their chemistry, particularly the "fun" colors, so you also end up with tinted specular highlights in real life. For renders, anything goes, it's just a matter of what look you want. Though hair is a bit more complicated to make "realistic", since each strand is slightly transparent and scatters light all over, and over, and over...

    Typically anisotropy starts with UVs. Say for a slightly bulged plane or hair strip, the uvs are typically aligned upward, pretty much what you see in the viewport for reference. But you could rotate the uv shell 90 degrees and get a different aniso highlight. Many times, the shader will have a setting for vertical or horizontal alignment of the UVs regardless. Some shaders also use a Flow Map to control the effect on a pixel level- like if you had a wavy hair texture but the mesh is just a simple flat plane. (or a stroked/buffed aluminum sheet) here's a few game-oriented links to get you started if you want: http://wiki.polycount.com/wiki/Anisotropic_map and http://wiki.polycount.com/wiki/Flow_map

    Personally I wouldn't use anisotropic for fibermesh hair. The effect is meant to mimic the kind of highlights seen when you have a bunch of micro tubes. So if you have enough hair geometry, it's not entirely necessary. If you're struggling with it, try duplicating a long, thin cylinder into a row, and maybe bend that set around to see a little better what's happening on a smaller scale.

  • AndyGrimmAndyGrimm Posts: 910
    edited March 2016

    @throttlekitty

    thx..  i just play around with a simple fibermesh hairstyle.. the glossy color reflections i get are definitly wrong for hairs.... actually i expected to see the anisotropic effect beeing just there (enough geometry)...  but it wont happen... do you have a "working" example or a render done in iray with fibers where the glossy effect looks right ? 

    thats why i started to look into UV driven anisotropy direction... 

    Edit: could it be that the fiber profile must be smoother? 

    Post edited by AndyGrimm on
  • AndyGrimmAndyGrimm Posts: 910
    edited March 2016

    using organic values for specular (dark grey values) does also not look correct... the tinting of specular is way to strong on fibers...  custom fresnel which i tought is the problem does not change that...  there is way to much specular tint for F0... in iray (for fiberhairs)...

    just brainstorming here.... and maybe someone has a fibermesh solution or advice for a working or cheating setup in iray allready...

    Post edited by AndyGrimm on
  • throttlekittythrottlekitty Posts: 173
    AndyGrimm said:

    @throttlekitty

    thx..  i just play around with a simple fibermesh hairstyle.. the glossy color reflections i get are definitly wrong for hairs.... actually i expected to see the anisotropic effect beeing just there (enough geometry)...  but it wont happen... do you have a "working" example or a render done in iray with fibers where the glossy effect looks right ? 

    thats why i started to look into UV driven anisotropy direction... 

    Edit: could it be that the fiber profile must be smoother? 

    I don't have any examples personally, I haven't had the time to really dig in with iRay, sadly. Mind posting a pic of your tests?

  • AndyGrimmAndyGrimm Posts: 910
    edited March 2016

    well... my test renders do not really help yet - they only open questions smiley....   it is my plan to dig in fibermesh and Iray hair shaders now... so for sure i will come up with test results which make some sense (not the ones i do right now)...

    but your answer just lead me on the possible track that the fiberprofiles must be way smoother to reflect the light to the next fiber ( i use only 3 sides right now)....  more to come..

    meanwhile i am thankfull for every link/advice/experience and ideas for hair fiber shaders with iray or other pbr engines (not low poly game hairs) smiley​

    Post edited by AndyGrimm on
  • Oso3DOso3D Posts: 15,047

    For hair, I like using bump maps (and normal maps, if I can get them) tiled a LOT to capture the look of strands of hair (for transmapped hair). It's pretty ideal for anything but a REALLY close-up shot, and at that point you kind of need fiber based hair of some kind anyway.

    At one point I tried the same thing with displacement, but Iray displacement is such a hog for mediocre results, most of the time.

     

  • Arnold CArnold C Posts: 740
    edited March 2016
    AndyGrimm said:

    Hairs:

    I googled a lot but cant find a document which describes hairs for  a PBR workflow.

    Specular for hairs should be in the 0.02 - 0.05 range - definitly hairs are not metal... but most PBR hair shaders use colored specular.. the better shaders even a primary and a secondary specular color.....

    Are hairs the big exception of physically correct shading? 

    Yep, it's definatey not... I'd speculate that their creators are the same who put any grey, blue, pink, whatever into the "Glossy Color" of their skin materials.

    Hair has an IOR of around 1.55, so it would have a Specular Color / Reflectivity value of around 0.046521 / 0.581507.

    If you're looking for a quick solution, Mec4D's Unshaven 2 product's Iray shaders can be applied to almost any transmapped hairstyle, keeps bump, opacity and displacement maps in place, only removes the diffuse texture maps and adds color values instead which saves a bit of VideoRam.

    If you're looking for inspiration you could have a look at NVIDIA's Fermi Hair Demo. Would be nice to have something similar in DS. Maybe with DAZ Studio 10? smiley

    There are some documents on the NVIDIA Siggraph 2008 presentation.  Maybe you'll find something helpful threre.

    EDITH:

    As I just saw, DimensionTheory's Subsurface Workshop product offer hair shaders, too.

    Aww, I just got my wishlist down to 100 items...frown

    Post edited by Arnold C on
Sign In or Register to comment.