No problem, I’m glad you find this interesting!
That’s pretty slick. I suppose it does the same thing to the brush, and hopefully the smudge buffer and anything else that might be affecting painting on that layer?
Not at all, it is probably the most basic approximation. I think if you looked into it, it might be a Beer Lambert equation. Basically the attenuation of light as it goes through a substance.
I’m not so sure, I think the response is dependent on the particular colors, too. So if you adjust the curves just right it might feel the same for white+black, but that would mess up the way it feels for other colors. I know if you do linear gradient with this method darker colors will dominate lighter colors across most of the gradient. Since that is how real paint work I think that’s a key component to why it feels more natural and expected. But you may be right, I didn’t try adjusting the curves.
Also, did you notice blue+white skew towards violet w/ sRGB, but skews slightly towards cyan with the spectral? I think most would agree the latter is more pleasing. The interesting thing is that the skew depends on the shape of the spectral curves, which can vary a lot depending on how they were calculated. So, there is opportunity for customizing how your colors behave, if one was so inclined.
I don’t think there is a robust way without having more primaries. I suppose they don’t even need to be on the spectral locus, but no reason not to place them there. I’ve tried HCY/HSV and CMY but they don’t seem natural, or they have weird issues.