@troy_s correct me on this if I’m wrong but as far as I undertand it:
This is not what Troy was saying at all.
This is about OCIO having the concept of “Looks” which Blender’s color managing does support but many other softwares do not.
AFAIK you can just basically bake in a “look” into the base transform, bypassing the looks concept to get this to work as intended and be perfectly matching in other software. That’s more or what has been done on the DaVinci Resolve version of the AgX OCIO, as I understand it.
Instead of a single AgX with a bunch of look variants, you’d then need a bunch of separate “Agx Look1, AgX Look2” or whatever.
Now whether looks are a good idea in the first place is a different story. But either way, it has nothing at all to do with AgX as such.
On that note I should mention the way to generate the version of AgX that landed in Blender is completely open source. (In fact, that was kind of a requirement)
If anybody wants to “roll their own”, that starting point is there for y’all to try.
I mean, color understanding was more bound to stuff like particular materials. For instance, you can have an “absolute blue” if you talk about a very specific, say, Lapis Lazuli Blue. In that case, you refer to a very specific compound, and it has its own “texture” and behavior when it is lit in different ways or mixed with other pigments or what not, and all of that would then be included in what it means to be “Lapis Lazuli blue”.
And in this scenario, the “absolute” color also has effects to worry about such as changes with age! Physical pigment colors have a half life. Classic paintings such as the Mona Lisa used to look very very different in the past. Classic Greek marble statues were painted and not just plain white.
Of course even this has caveats such as there are different quality levels of that kind of blue, i.e. how deep blue was the lapis lazuli you crushed up to make it?
Was it very high purity, or did it have a lot of white material or specks of pyrite?
Like, it’s gonna depend on which of these you crush up to make your “absolute” color.
So how about using precisely spectrally defined light source then?
Sure, in a completely black room with an exact scene setup, you can then reproduce hyper exact stimuli in that particular environment.
But when is that lab setup ever real?
Your room is not black so you automatically have to worry about ambient. reflections. The lighting also changes on whether the sun shines or the LED-based light bulbs start to shift or what have you.
And the ambient reflections thing is not a trivial matter. A colorist working at PIXAR for The Incredibles kept posting work that was too cool. It was a big mystery why, until the director paid a visit and found that the room had red curtains.
The curtains kept spilling onto the screen, completely skewing the result! They had to change to grey curtains to fix it.
I even noticed this myself while helping test AgX: It mattered a lot whether I looked at my screen in the morning with sunlight streaming in, or late at night. The kinds of issues I’d see or not notice completely shifted. And it’s not just a matter of “oh this situation is better than that situation for testing.” The issues were just in completely different parts of the image.
For the same image.
Rendered on the same screen.
And notably, color perception shifts a lot even with things like language. To the Ancient Greeks, the ocean and their own red wine had the same color category. For ages nobody even had a word for blue. That’s an incredibly recent invention, in fact.
One place where this sort of thing is in action today is the notion of “pink” (not “hot pink”) which is just a light red. If it’s blue, you see that as “light blue”, if it’s green, you see that as “light green” or “lind green” or whatever. But with red? Pink.
Why? Not because your eyes process this differently somehow, but because your brain has learned to divide up the category differently simply because there is a category in the first place.
None of this stuff is objective or absolute. None of it can be. It’s intrinsically linked to how we see the world.
If you were a goldfish or a duck or a dog, colors would look wildly different to you simply from seeing a different spectrum. If you have any of the several variants of colorblindness, same thing.
Heck, even the color of your own eyes and, apparently, the degree of yellowness of your retinal nerves have a measurable impact on how you perceive color, that some of the studies I have read attempt to correct for.
All the mathematical models (such as the “hue” value proclaimed by a painting program based on sRGB) are just that. Models, meant to make sense of perception.
But unfortunately, perception comes first. The math is completely arbitrary and, quite often, a rather poor fit to “what’s really going on.”
Certainly, historically we have not dealt with a hardcoded sRGB for “thousands of years”.