Daz 3D is part of 
Connect
DAZ Productions, Inc.
7533 S Center View Ct #4664
West Jordan, UT 84084
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2025 Daz Productions Inc. All Rights Reserved.
Comments
COLOR is made with LIGHT..
From the RED you create another color based on the LIGHT.. which is exactly absorbed with RGB in our single sctter = final color AND luminance
And you can calculate this in two directions!.. like algebra
that it scatters is NOT relevant.. wne all 3 wavelentghs have the same absorbition amount... it does still sctter.. and project ALL pixels the same back on our surface... just with a very high luminance ...
but that's easy... set the transmitted color to a lower grey.. and the luminance sinks. Because ALL colors just get darker..... and the difference between R G B.. well is in the albedo map... darker = more absorbation.
And now - where do i not understand absorbation and scatter... absorbation is simple colormixing. i did this a decade or more...
scatter... well trace a light beam... as longer as the way is so darker it goes.... but when the way for all light beams the same is... i can also send a dark color from beginning.. and have exactly the same ratio at the end of the trace /ratio to the other colors..
it is actually easy... test it .. grey texture, full albedo.. no color absorbation.. or ALL the same ... and it should look good.
Absorption and scattering are completely linked. I'm sure of what I say, that is a part of what I was teaching to my students.
In order to scatter, light must propagate inside a volume (because what we speak about here is volume scattering, and not the surface scattering of rough surfaces). This volume absorbs light before it goes out. That is what links them.
the amount of light exiting the volume for each wavelength (or color channel) depends on the linear coefficient of absorption for this wavelength and the propagation path of the light inside the material. They are defined by TMD+TMC (+TC) and also by geometry vs source vs base shader. By wavelength you must understand color channel in our specific case.
Absorption is not a simple "colormixing", it has a decreasing exponential dependency % propagation length PER WAVELENGTH. Absorption coefficient it different for each wavelength. Never the transmitted color will be the one you set as TMC, except for the propagation length being exactly the TMD. The rest must be computed using Beer Lambert law or an equivalent law Nvidia would have decided to use. TC is a global correction of that transmitted color (not TMC, but computed transmitted color) for translucency computation.
The nuance in all that is that in DS it is not a wavelength dependency but a channel per channel one. At the end there is not the nice integral of the parameters, but a simple channel computation. This is not physical - well the sampling is insufficient - but this is obligatory to keep reasonable calculation times.
I'm having a 40 pages report about scattering phenomenon in Iray in DS (I have written it after several days analysing this), I think I'll include it as an annex part of a video tutorial later on in 2016 about Iray materials.
So to be brief, the ratio of the RGB you have will change regarding at least the 5 parameters mentionned above, plus others. There is also a dependency on the source, not on absorption, but on the resulting color, still channel per channel.
i must not take care about this.. when i send the FINAL COLOR which i want with ZERO absorbation in the scatter..
Our goal is that it looks.. well like a Albedo which includes Scatter...and that' is what we get....
And for sure absorbation is at the end simple color mixing.... otherwise i could not see a result on my screen
heck what are you making it so diffuclt?
every wavelenght i can calcualte in RGB as long as it is in RGB range...
and when i know that at thickness x absorbation is Y... i can simple set the RGB!
IN A SINGLE SCATTER WITH ONE THICKNESS - that's all we have here
we know the final result which we want allready! this is OUR albedo map with backed in SSS
he get a one to one projection, on the surface...which includes translucency.(because it still scatters).. not 100 right i know.. BUT NOT TO SEE.. if somebody can calculate that backwards- and the test from the correct side - aka setting RGB absorbation AND a color corrected map in translucency..
And NOW.. someone of you please calculate and give me the RGB for transmitted color.. so that that i get the same result.. as i have allready.. with a red color.. or any other color map in transulcency.
this is what Jaq does - he takes the final result.. because it is allready there - in a single scatter with one thickness
There she is. And you seem to have telepathic abilities...
You didn't tell me you also teach. 
Please don't use the term "SINGLE SCATTER™" anymore... it makes me want to strangle you...
(Ha, now I get it: you want me to become DARTH ARNOLD). 
You can calculate this TMC easily for one given thickness (well one given propagation length). Just take the final R,G,B you want to reach, then fix the TMD you want (expressed in cm in DS, which is common for the 3 channels) and then fix also for instance the translucency color (per channel - it does not really matter as long as you have no 0 in channels). Then for each channel, you have to revert the beers law for the value of transmitted color on this channel. Do this for the 3 channels and you have your TMD, since the 3 times - for each channel - you have one solvable equation with a single unkown variable. You can even do this on excel if you want to reuse afterwards (excel has neperian logarithm I hope)
Yes, maybe I had not told to you that after working for exo planet detection, I changed a bit my profession, and then I became an expert in optical radiometry, photometry and colorimetry (plus human visual perception) for my company, and this is why some engineering schools hired me to teach that (also optical sources and detectors) to their students. Not easy to teach this to students on the friday morning after their "thursday evening parties". This field of physics is made of plenty of intregrals over space (over surfaces and solid angles) and over wavelengths (source, detectors, surfaces and volumes), and this is better to be awake with a fresh brain (which did not seem to be the case of my students!!!!)
It is not in my inteend to degrade sience
..
But the difference from which side i start - absorbation -> final result.. or .. final result -> no absorbation....where is it ?
And even MORE important - if there is a difference - which one is more correct?
NP. You seem to look for (if I understood well) the TMC (transmission measurement color) and you have a final result in mind. In this case you have to put your 3 equations for find this RGG.
For the rest or your question I don't clearly understand. The best for you is to define exactly what you know and what you need.
So in order to understand what you want to do the real questions are :
- what is exactly the "expected" result you want? Translucency?
- What are exactly the "known" and "fixed" variables you have?
- What is the "unkown" variable you have and you are looking for?
a light particle X enters our model surface - snaps the color of translucency..red as example.. scatters and lands finaly back on the sruface with a lower lumninace based on R in our transmitted color...
a light particle Y with the same low lumincane as particle X after scattering - starts from our surface does the same way - and lands on the surface not changed because there is no absorabtion...
must be the same result - or what do i overlook?
@V3Digitimes
what i want to know...
1. WHy do you not use a Albedo Texture with backed in SSS in translucency and scatter without absorbation?
2. Why do you make it so difficult to set a X color or map.. and Y transmitted color.. when the final results we want is what we have allready - the SSS which is in our albedo! which includes ALL absorbation paramaters allready!
The light particle does not enter a surface but a volume. It does not "land " on surfaces. It is either reflected, or transmitted on surfaces.
To be sure are you observing the back-scattered part of translucency beam (the one sent back in the half space the light is coming from) or the part of the translucent beam in the propagation direction (the one going "'through)?
the ONLY difference i expect is backlight translucency color .. but this difference is minimal.. and even better from the albedo, then from a single color.which is not thickness mapped..
1. Scatter without absorption is not physical. Every volume is absorbing. Physicist do not recognise that and Nvidia people say it would be a mistake to do so. If you use scatter without absorption on a volume with thickness variations, would would obtain a constant translucency, whatever the distance the light travelled into the volume. It means that if you use this on figure they will become "like a candle". It means that light is able to go through the torso of your figure as well as in the ears, with the same amount of light going through for all. You can multiply this light by any translucency map at the end, the "amount" of light will be the same everywhere, just a bit patched by the translucency color. You will light going through all the volumes equally if you do so. Not only this is not physical but it is also bad at renders!
2. It is not a matter of difficult, but of sense of doing so. I agree a wide part of the scattering information is already in the aldebo - well base color. This is the way images are built presently, especially when they are made with photo reference. The photo reference does not exclude SSS or translucency. But what is not in the maps is the way scattering will react to lights in your scene. The SSS information you would bake in your map is "constant" and depends only on the lighting conditions you baked them from. In this case, they are perfect if you rerender with the same lighting conditions. Now, change the light conditions in the scene, the information in the map, baked with another type of light source, do not correspond any more to the new light condition. They are not the convenient ones for your case. You have to re-bake.
@V3Digitimes thanks for detailed answer...
What is the average color for translucency of skin ? or what is recommended in teaching models ? why use so many pure red 100% ..any reason for this ?..
Which leads then to TMC... we know the values of skin.. or the RGB penetration deepness for each color, red, green, blue... is my approach correct to use this ratio for TMC ?
That's an interesting question... I'm currently trying to find out. The most puzzling is, from my observation, in Iray the Translucency Color seems to act as some sort of multiplicator for a corresponding Transmitted Color. Need to test more colors, though...
that's what i tought it IS so - arnold..... that's wh i said there is ONLY 100% luminance.. if you use 255 red in translucency and a high luminance in red Tmc.. then you clip.. red... that's why very pale close to white (grey) is not possible anymore...
You're right Arnold. It is a multiplicator.
Why 100? I think it is because is easier to use 100% red, easier to handle. Because one began to do this and the other ones followed. This is the only reason I see. But take care that by doing so, you totally kill the GB component which is in my opinion not reasonable. I personnally use "what I estimate would be the transmitted color for the figure at the transmission distance I impose", and then I use the translucency color as a slight correction more or less to the red. And this is never pure red, even if red is less absorbed than the rest. Human translucency should in my opinion not be 100% red. Blood could be. Flesh or cartilages are not.
Take a pure green backlight (light coming from a back of a figure). Should there be translucency "through the ears" in real world? Yes.
Now, set a pure red translucency or tranmitted color in DS. Is there any translucency through the ears? No! because you multiply the red of the source (0) by the value of translucency (255). At the end you reach a 0,0,0 translucency.
Lazy method to put pure red. Easier to handle in a lot of cases, but not rigorous.
you mentioned that we can backcalculate.. aka whenn i have a redish/brown in translucency.. and i want a x color on the surfacce reflected? (landed is wrong
)..and i know the base color....
so - the correct way in our limitated Iray iwould be... to take the average color of a translucency map ? then from the base color... then we say we want they final color X..... and use TMC formulas....and here is our RGB for Skin..
does this make some sense ? i mean one argument to not use backed in SSS in translucency is different light....
even more wrong seems to be to use wrong TMC colors... should we start to calculate TMC ?
YES!!!!!! TMC is the base of all. But you have calculate TMC for a given TMD. More preciscely, the couple TMC/TMD - plus take care at then end of SMD when you include SSS (and of SSS reflectance tint if you opted for a scatter transmit mode). Actually real tranmitted color is the base of all. Real tranmitted color is for each channel : Io.exp(-a.x) where Io is the initial value of the channel (you can here already include light color, or include light color at the end), a the coeff of abs, and x the propagation length in the volume (the thicker the most absorption in a decr exp way). You have one a for each channel, obtained via : ITCM = IoTMC.exp(-aTMD). Invert with an neperian (natural) log to get a for each channel, and you are able to calculate the real transmitted color for each "volume thickness". This for each channel and of course for the channel ITMC/IoTMC *255 is your color on the channel. At the end multiply by translucency color. Then take care, this will give your only the color of the translucent part of the beam. Then you have to include the base color influence, and if your gloss and SSS are null, you will have the resulting color. (with SSS things go more complex).
But I cannot really say more, or otherwise I'm gonna finish to put all my tutorial here, which won't pay the hard days of work I had on it.
PS : the solution TMC/TMD is not unique. They is an infinite (well finite because color are sampled on 255) number of solutions for a given RGB color at a given distance (a fixed distance which is not the TMD, but "any" distance). Well never mind, several TMC/TMD can lead, for a given distance to the same translucency color. But for all the other distances, the resulting color will differ. TMD distances of the order of magnitude of the cm will give you the widest/finest range of 255 sampled RBG values for a coherent result on human skin.
@V3Digitimes.. ok..thanks.. that's a lot of new stuff for Arnold to learn
.
can you point us on ressources ?
Crrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrt... pling. That's the € 200 bill dropping by a cent...
Thanks a lot! And a lot lot more. That'll mean I'd to develop a formula for each Translucency Color RGB in connection with the Translucency Weight parameter to get a target Transmitted Color RGB (which must be known) at a 100%. We definately might need a greyish dead looking Diffuse/Albedo texture for that. Or else a common Diffuse Texture will be overblown by the Transmission.
that's why i started to narrow the range of my diffuse maps..and used colors out of the albedo range for translucency - . first because i saw cliping of translucency colors... and some dissapeared... congrats finally you understand scattering and transmited color...in irays SINGLE SCATTER
so what is now the name for irays scatter? monochrom scatter it is not.. i tought the correct name is single scatter (one thickness, one TMC in RGB... but not sure
AHHHHHHHHHHHHHHHHHHHHHHHHHHHHH... you did it!
https://youtu.be/W175S0ageVI?t=22s