Diffuse Strength + SSS Strength = 100%?

wancowwancow Posts: 2,708
edited May 2013 in Daz Studio Discussion

Okay... so I was just told that Diffuse and SubSurface Strength on any shader in 3DL should always add up to 100% for it to work as it is supposed to...

Is this true? I'm looking at a render of Bree, the only thing I changed was dialing down the diffuse to 50% (because the SSS was at 50% in the default...) and I'm a little... unclear as to why I would be told this...

Oh, I only did it to the skin, Lips through Toenails..

Bree-test2.jpg
477 x 603 - 95K
Post edited by wancow on
«1

Comments

  • Richard HaseltineRichard Haseltine Posts: 102,480
    edited December 1969

    Well, there's the idea that if you are trying to be physically correct, or even plausible, the outgoing light (specularity+diffuse+SSS) should not exceed the incoming light, which would be the case if their strengths added up to 100% (or even a bit less, since some light is lost - absorbed and remitted as heat).

  • SzarkSzark Posts: 10,634
    edited December 1969

    It is a new one on me but I use SSS strength maps so I always have it on 100% and Diffuse on 80%.

  • wancowwancow Posts: 2,708
    edited December 1969

    The math is beginning to make my head spin... unless it's not based on incoming light and only based on what's reflected... So If I set everything at 33.3, Diff, Spec and SSS I would be "correct?"

    Let me just add this: I have never seen light coming through the ear the way I see on some renders. I don't know how to even get that...

  • Richard HaseltineRichard Haseltine Posts: 102,480
    edited December 1969

    There's often more SSS than is strictly realistic, which can produce excessive light coming from unexpected places (as can not having sahdows on on lights with Translucence on).

  • ZilvergrafixZilvergrafix Posts: 1,385
    edited December 1969

    people is obsessed with SSS since the eons of times...in real world is a minimal parameter when many others values do more for a realistic skin.
    in MaxwellRendes I saw about 3-4 skin maps like dermal, subdermal, oil map...OIL MAP???

  • wancowwancow Posts: 2,708
    edited December 1969

    ZilverG, I know! I know... what you say is true. It's just that those guys using Firefly have lorded it over us so long... I wanna SHOW THEM! :P

  • SickleYieldSickleYield Posts: 7,644
    edited December 1969

    wancow said:
    ZilverG, I know! I know... what you say is true. It's just that those guys using Firefly have lorded it over us so long... I wanna SHOW THEM! :P

    It's harder to get that glowy-eared translucent plastic kewpie doll look in 3Delight, but if that's what you're going for, crank the velvet up to 100% and set it to a medium tone on top of everything else.


    I don't think SSS should approach or exceed diffuse. When I try to render with skins with settings like this they look unnaturally luminous and don't seem to react to scene lights, and if I'm not creating a supernatural character, that's not usually what I want.

  • evilded777evilded777 Posts: 2,465
    edited May 2013

    Due tothe physical law of Conservation of Energy, (diffuse+specular+sss ) = 100% total. Bowing to richard, SSS ought to take the absorption into effect.

    This ain't the real world.

    Would it be nice if the shader, somewhere, took this into account and did the balance for us? Yes.

    Are the results accurate? Well, no not really.

    SSS is actually a much larger part of the visual appearance of many things, human skin being a particular example of where it is very important. Diffuse reflectance is actually pretty damn low on the percentage of what we actually see. But what we have here in Studio and other renderers is not a physically accurate representation: its a best guess, an approximation, a replica.

    This is also ART. Some of it really OUGHT to be up to the eye of the artist, attempts at photo realism somewhat to the side.

    Until Studio has a shader like ezSkin (yes, I know, its not a shader per se; but you know what it does have: conservation of energy built in, or at least an approximation there of), Studio is going to be more on the artistic interpretation side and less on the realism side.

    And I'm not dissing Studio, you'll find me in the Studio forums over the Poser forums 9 times out of 10. These are simple facts.

    Post edited by evilded777 on
  • evilded777evilded777 Posts: 2,465
    edited December 1969

    wancow said:
    The math is beginning to make my head spin... unless it's not based on incoming light and only based on what's reflected... So If I set everything at 33.3, Diff, Spec and SSS I would be "correct?"

    Let me just add this: I have never seen light coming through the ear the way I see on some renders. I don't know how to even get that...

    No, that would not be correct.

    To be "correct" your diffuse would need to be way down (this number 0f 6% sticks in my head from something i read, but I may be wrong), with specular somewhere above that and the majority of it coming from SSS.

    If you use the US2 shader, you might be able to get some backscatter, I have not tested it; but there is a backscatter boost built in. I find it unlikely that you are going to see much backscatter with other shaders without a serious sss map (and I don't it even then).

  • evilded777evilded777 Posts: 2,465
    edited December 1969

    people is obsessed with SSS since the eons of times...in real world is a minimal parameter when many others values do more for a realistic skin.
    in MaxwellRendes I saw about 3-4 skin maps like dermal, subdermal, oil map...OIL MAP???

    That's not quite accurate. SSS is the PRIMARY factor in what you see when you look at human skin. Diffuse reflectance and specularity are minimal in comparison.

    in your brief mention of these maps... well yes, to be truly accurate you are going to need all of those and probably more, and all of those that you mention would be involved in sss calculations: oil would apply more to specularity, but if its reflected by the oil is not going to penetrate and scatter, the dermis is going to absorb some and reflect some and some will penetrate (we are talking about light here) and the same with subdermal (though this is where a lot of the scattering happens).


    I have begun to truly realize what an awesome process this is, and it does not surprise me in the least that we are not able to simulate it casually.

  • evilded777evilded777 Posts: 2,465
    edited December 1969

    oh, and you NEED global illumination of some kind or a ton of lights (which is not as accurate or desirable IMO).

    And when you start adding in the various control maps, you're "Accurate" percentages go out the window.

    Fudge it. Use light tricks. Use all the tools in your arsenal, because its a COMPLEX process that changes in an eyeblink.... so just do your best, and satisfy your own eyes.

  • Joe CotterJoe Cotter Posts: 3,259
    edited May 2013

    .. SSS is the PRIMARY factor in what you see when you look at human skin. Diffuse reflectance and specularity are minimal in comparison.

    The top layers of skin are almost totally transparent, think of when one peals skin away. If one were to try to mimic reality they would put the 'Diffuse Map' in the SSS channel and have a semi opaque setting in the diffuse it would seem. I've seen some skin textures where they do this. I've played with doing this on various materials but haven't done enough yet to put up any conclusions, other then at times it gives the exact same results putting the diffuse map in diffuse or sss. So, I'm guessing the differences come with lighting and various settings I haven't worked out all of the details on yet.

    Post edited by Joe Cotter on
  • wancowwancow Posts: 2,708
    edited December 1969

    I always render with a map in the SSS Chanel, most often the diffuse map because I'm too lazy to create custom SSS maps... however, most of what's been said here I really didn't know or had never thought of.

    Thank you Evilded, you really shed quite a bit of "Light" on the situation for me. You jacked up the intensity! You turned off the damn limits! :P

  • SickleYieldSickleYield Posts: 7,644
    edited December 1969

    people is obsessed with SSS since the eons of times...in real world is a minimal parameter when many others values do more for a realistic skin.
    in MaxwellRendes I saw about 3-4 skin maps like dermal, subdermal, oil map...OIL MAP???

    That's not quite accurate. SSS is the PRIMARY factor in what you see when you look at human skin. Diffuse reflectance and specularity are minimal in comparison.

    in your brief mention of these maps... well yes, to be truly accurate you are going to need all of those and probably more, and all of those that you mention would be involved in sss calculations: oil would apply more to specularity, but if its reflected by the oil is not going to penetrate and scatter, the dermis is going to absorb some and reflect some and some will penetrate (we are talking about light here) and the same with subdermal (though this is where a lot of the scattering happens).


    I have begun to truly realize what an awesome process this is, and it does not surprise me in the least that we are not able to simulate it casually.

    Hmm. I wonder if geometry shells fit "tight" enough to simulate an epidermis? Probably not, but it might be worth trying. Most of the time I just don't think about the biology because it's so irrelevant to how models are built and rendered (we are very squishy, complicated creatures compared to these!).

  • wancowwancow Posts: 2,708
    edited December 1969

    Hmm. I wonder if geometry shells fit "tight" enough to simulate an epidermis? Probably not, but it might be worth trying. Most of the time I just don't think about the biology because it's so irrelevant to how models are built and rendered (we are very squishy, complicated creatures compared to these!).

    You know that thought occurred to me too... experiments might be revealing...

  • SickleYieldSickleYield Posts: 7,644
    edited December 1969

    I'm going to say no. First pic is default Bree, second is with a geometry shell at 16% opacity and her own mat with some settings changed. Note nose aura.

    BreeNoseAura.jpg
    1000 x 1000 - 411K
    BreeDefault.jpg
    1000 x 1000 - 431K
  • wancowwancow Posts: 2,708
    edited December 1969

    that's unfortunate... :( It was a good thought, however! At least we know!

  • ZilvergrafixZilvergrafix Posts: 1,385
    edited December 1969

    I ever wonder why SSS from high end [they claim]render engines produces that green glow around their SSS skin shaders

    sss.jpg
    1920 x 1200 - 579K
  • JaderailJaderail Posts: 0
    edited May 2013

    Nevermind was stupid on second thought.

    Post edited by Jaderail on
  • evilded777evilded777 Posts: 2,465
    edited December 1969

    I ever wonder why SSS from high end [they claim]render engines produces that green glow around their SSS skin shaders

    That's not SSS producing that, it CAN"T be. The SSS settings are clearly shown in the top left, and there's no green. That is most likely a diffuse rim effect, like you can get with pwSurface.

  • ZilvergrafixZilvergrafix Posts: 1,385
    edited December 1969

    I ever wonder why SSS from high end [they claim]render engines produces that green glow around their SSS skin shaders

    That's not SSS producing that, it CAN"T be. The SSS settings are clearly shown in the top left, and there's no green. That is most likely a diffuse rim effect, like you can get with pwSurface.
    evilded, you speak with so much expertise, can you post your images validating all your knowledge? :-/

  • ReDaveReDave Posts: 815
    edited December 1969

    Zilver, the bottom part of the jaw looks green too, but the underside of the arms don't, so I wonder if you have a different setting on the head and ears?
    It's strange, Keyshot uses the counseling of Jenssen, the guy who wrote all those articles on SSS, and, as far as I know, is the only program doing the computations in the proper colour space (although I'm going to say the only proper colour space is just plain physics models, rather than any of these colour spaces people have invented just to annoy everyone else). In any event, that colour space that Keyshot uses alters colour significantly, in particular greens become very strong, so try a redder colour (or less green red) on the first channel.

    @wancow: Diffuse+SSS can be anything in general. You need two conditions to justify starting to apply real physics (and thus justify the Diffuse+SSS+Specular+ambient+everything else=1) to rendering: 1) you must have bounces for the light; 2) you must be in a space where you can add colours and obtain something close to real. <---Condition 2 never happens in any renderer, otherwise they would be rife with Boltzman constants and stuff.</p>

  • ZilvergrafixZilvergrafix Posts: 1,385
    edited December 1969

    ReDave said:
    Zilver, the bottom part of the jaw looks green too, but the underside of the arms don't, so I wonder if you have a different setting on the head and ears?

    Keyshot considers this Cookie figure like a bunch of glued geometry, for Keyshot, ears are different from head, and body is a third geometry, AFAIK it renders individually at different times even with welded geometries on the export settings it just render separately.
  • Scott LivingstonScott Livingston Posts: 4,344
    edited December 1969

    I can see why this Diffuse+Spec+SSS=100% math might not work in a biased renderer. Has anyone tried it in LuxRender? I may just try to play around with it in Luxus later...

  • Joe CotterJoe Cotter Posts: 3,259
    edited December 1969

    ... that glowy-eared translucent look in 3Delight,...

    Unless I misheard, that's what Omnifreaker said Fresnel was for in his video on US2.

  • Joe CotterJoe Cotter Posts: 3,259
    edited December 1969

    I can see why this Diffuse+Spec+SSS=100% math might not work in a biased renderer. Has anyone tried it in LuxRender? I may just try to play around with it in Luxus later...

    Regardless of biased vs unbiased, from what I understand, any global or omnidirectional light source will by definition not provide much if any specularity by nature. If one wants to use said light source and doesn't want to loose the specularity, they would probably have to make up for it some way. Of course the real test is as you say, to set up some test cases and see what the results are.

  • Joe CotterJoe Cotter Posts: 3,259
    edited December 1969

    ...produces that green glow around their SSS skin shaders

    I could be wrong, but that very definitely looks like SSS to me

  • evilded777evilded777 Posts: 2,465
    edited December 1969

    I ever wonder why SSS from high end [they claim]render engines produces that green glow around their SSS skin shaders

    That's not SSS producing that, it CAN"T be. The SSS settings are clearly shown in the top left, and there's no green. That is most likely a diffuse rim effect, like you can get with pwSurface.


    evilded, you speak with so much expertise, can you post your images validating all your knowledge? :-/

    Well, over in the sister thread, 3Delight Thread in the Commons, you can see several.

    It is my opinion that if you look at that image and note where the primary or hottest light is coming from (top right) that greenish tinge can not correspond to SSS because of where it appears and how SELECTIVELY it appears. It appears most prominently around the outside edges of the ears, and only there on the ears -- one of which in on the completely opposite side of where the light is shining. I think the ear rims might even be part of the texture. The other areas, at the right armpit, under the "breasts", under the mouth, etc, all look to me be a diffuse rim effect.

  • ZilvergrafixZilvergrafix Posts: 1,385
    edited December 1969


    It is my opinion that if you look at that image and note where the primary or hottest light is coming from (top right) that greenish tinge can not correspond to SSS because of where it appears and how SELECTIVELY it appears. It appears most prominently around the outside edges of the ears, and only there on the ears -- one of which in on the completely opposite side of where the light is shining. I think the ear rims might even be part of the texture. The other areas, at the right armpit, under the "breasts", under the mouth, etc, all look to me be a diffuse rim effect.

    saw your images.

    uh...The green effect was vanished when the progressive render finished, Keyshot does realtime rendering and I don't take an screenshot of the final work.

    I could tell that "green" fx is only a sub process of the shader itself.

    as far as I know, the only champion of SSS in DazStudio is the user known as LATICIS, his render about SSS on deviantart proves that.

  • SzarkSzark Posts: 10,634
    edited May 2013

    Interesting thread

    as far as I know, the only champion of SSS in DazStudio is the user known as LATICIS, his render about SSS on deviantart proves that.

    It only seems that way as He has made his findings public and he can do a lot of testing and with help from others ;) I thought I knew a lot about this subject due to my testing over the years and only recently have I started to get a handle of things. But some aspects I found I know very little when I started to get deeper.

    I always thought the SSS shading rate was better higher than being low , opposite of the render setting shading rate.

    Never heard of the 100% balance between Diffuse and SSS until recently and I am still unsure if that is correct for Uber Surface etc. As some have already mentioned that the skin texture maps we use have SSS burnt in given the nature of how they are made so in essence everything is a cheat. I honestly don't think there is one hard and fast rule when it comes to this, yes there are guides lines that we could use to set up Uber Surface form the get go and then tweek as required.

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