The previous post about D65 n°4 starting was false alert.
A lot of being human is about being stupid, and what happens then is failed works.
This is the ‘new’ or ‘real’ D65 n°4
CLICK ON THE PICTURE TO ZOOM!
I used to hate the adjectives ‘warm’ or ‘cold’ (in French couleurs chaudes et froides) when used to characterize colors. But often I think about them when I look to little hue variations, I try, and I cannot help thinking that these adjectives can be quite useful. It’s funny, I think everybody can understand what’s meant when a color is qualified as being ‘cold’ or ‘warm,’ especially when colors are compared: like if I say ‘this color is colder than that one.’ The series of the Reds here can be seen as getting ‘colder’ when Zinc White is added (the Reds that become pinkish.) Same for the Oranges on the left, they can be said ‘warmer’ than the Reds, no?
I am expanding my color “gamut” by making “subtractive” pigment mixtures. Everything is “plotted” in a CIE u’v’ diagram (see below).
The blue dots are measured colors, the two red squares are very pure cadmium Reds, the green triangle and the light blue cross are my new best Blues (one being very very dark: 3% of light diffusion), the cross in the middle is the white point. The Orange dots represent the limits of the sRGB gamut – I’m getting very close!
I figured out that even my best mineral pigments can become better by “purification” using transparent organic pigments.
CLICK ON THE PICTURE TO ZOOM!
On the left: the colors of the planckian blackbody locus from 4000 to 9500° K with the same luminance as the background
On the right: the same colors with the maximum luminance value possible inside the gamut of my colorsystem.
Note that the colors on the left are the average of 4 colors, on the right of 3 colors, but their hue is precisely the same.
Have a nice day!
CLICK! on the picture TO ZOOM!
Average colors of the Blackbody radiation colors (from 3500°K to 9500°K) and the color of the background. 50% of the paper is covered by paint. (This actually equals a desaturated picture of these colors, i.e. the mixture of blackbody colors and white D65 light at equal power)
Doing optical greys with colors might become history in my work, I start doing colors with colors – let me say something about my color system here:
I use colors following “Grassmann laws,” that means that I measure the lights coming from my pigments to your eyes, and by varying the proportions of these pigments I change the average color of the light that comes to you.
That works exactly in the same way as you computer’s screen except that I have more primary colors and usually less light.
Here, I used 7 “primaries”: Ultramarine Blue greenish extra, Cobalt Turkish light, a Green mixture of the former and of the next, Cadmium Yellow light, Cadmium Red “Cinnaber” shade and Titanium White (see below)
The “primaries” of my system (didn’t use the Cobalt Pink here yet!)
To create a color, I will make the average of 4 of these primaries, so a “possible” color will always be located in a “pyramidal space” with a primary color at each peak.
If I add the Black primary, the main pyramids will be:
ReGrBlu-White and ReGrBlu-Black. Around them you’ll have ReGrYe-Whi/Bk ; GrBluTu-W/Bk ; ReBluPink-W/Bk
Which makes a total of 8 pyramids that touch each other and define the gamut of my system! (see below)
8 pyramids = gamut of 8 colors
The best solution to get the amounts a, b, c, d of the colors A,B,C,D required to get a given color M is to use matrix calculus:
with: a+b+c+d=1,
the colors coordinates are in the CIE 1931 XYZ color space, so we create a square matrix with the color coordinates for each color, multiply by the solution, and equal it with the color we want:
Then we solve this:
with P being the inversion of our matrix:
That’s it!
Thank you very much PA5CAL from this forum to have helped out!
* * * *
Today could be a great day for painting: Georges Seurat’s dream-project of a “scientific painting” advanced a little. I’m sure he would be (very) happy to see this. I wish I had the email of the painter’s paradise, I would send this to him and all his friends…

Hi-resolution picture: CLICK ON IT TO ZOOM!
The two largest rectangles have the same color average: same colors in the same proportions.
The one on the top is made of 4 layers, one color per layer, one spatial frequency per color, and the angles between the layers are “optimized” in order to get rid of moiré effects.
On the other one there’s also 4 layers, but with a unique frequency, a 45° rotation and the colors are randomly diffused.
Both are trying to “match” the average color of the background and in both paint is covering 60% of the paper.
The 1st one is almost invisible, the second appears to contain intense colors.
I’m gonna do a “lamp simulation” (a circular gradient with constant tint and varying brightness, based on the abstract model of a “perfect lamp,”) based on 11 colors: 3 times RGB with 3 levels of brightness (I dont enter into the detail of how I choosed them) and Black/White.