What Have I Forgotten About Greys?

[TestTools]

Back in my youth, I remember something about 50% grey not being 50% grey and a trove of mystical perceptual knowledge things. Is this one of those?

You will notice on this "all for education, not a copyright infringement, I promise" 2020 color space version of the MacBeth chart that I am fumbling around with (attached). I put 18% grey behind all the swatches. How did I do that? By racking up the R in the 16-bit RGB picker and typing 65,535*.18, then copy pasting the result into G and B as well. But, when I pull down to Greyness, that turns out to be 25%.

In this iteration, the swatch marked 18% is "Grayness" 18% so you can see the difference between the two colors.

The numbers are different for 40% (a 5% difference) and 20% (a smaller difference, 2 or 3% but I didn't write the numbers down to say for certain.)

Background: The concept is to use this slide in a cinema projector that can correctly portray the 2020 gamut, and I am expecting that I can do the primaries and secondaries and greys by numbers. The little boxes in the arbitrary top two rows are the original Hex numbers (unfairly since they were made for D50 and a much smaller gamut and reflective, no emmisive), while most all of the larger sections are just my putzing around trying to use combinations of 25% increments to somewhat imitate what I understand to have been the original intention.

But it is this grey issue that I am wondering about – shouldn't I expect that Greyness 18% be equal to 18% of the 3 color values?

Thanks! 

MacBethStyle-16RGB-2020-ColorChart.afdesign

[NotMyFault]

1 hour ago, TestTools said:

 

But it is this grey issue that I am wondering about – shouldn't I expect that Greyness 18% be equal to 18% of the 3 color values?

Only in SRGB profile. Not when using other color profiles.

 

Below a screenshot of a grey gradient from 0 to 100% interpreted in a bunch of colour profiles. all files RGB/16.

Rec.2020 differs the most in darker tones.

[NotMyFault]

The main difference is the gamma curve. Use the scope panel to inspect.

If you use sRGB as reference: scope (Intensity waveform) shows linear response. (and Display P3)

If you import (place) others:

Adobe RGB: 

All remaining:

[TestTools]

Wow! Thank you. This is going to take me a bit of study to wrap my head around.

Many thanks for thinking it through and the details of explaining with graphics.  I guess I have to make in-situ tests to find out what gets through the entire system most accurately.

 

[TestTools]

@NotMyFault Thanks again. 

I think the issue you are describing may not apply to what I am describing. (…though I could be wrong.) I can see how different spaces may interpret with different ramps when placing something – But I am not 'placing' anything. 

I am creating 2 squares, then filling in the color.

I don't understand why 65535 x .18=11796 in each of RGB doesn't equal the pulldown of 18% Greyness. I would presume them to be the same.

But now I have tried even the most elementary example, using 256 color picker and even 8-bit Grey or 8-bit RGB in 709 space, and many in between my original 2020 16-bit space.

50% greyness and 256/2 (…or 256*.5) don't equal the same color. This happens on an iPad as well. And, I can't get help on "Grayness" in the manual.

 

2SquaresGrayness.afdesign

[NotMyFault]

We may cross-talking a bit. The most important part: we have at least 2 color pickers, and they work differently in case the display or document is not plain sRGB. One is bound to sample inside the document window, the other (in color panel) can sample outside the window, anywhere at the display. 

The difference is: one samples the color values inside the document, the other the result after applying the color profile and converting to the display profile. This may explain your observations.

 

 

[lacerto]

I am not sure if you could have asked a question involving more complexities than when asking how grays are handled in color conversions (and not just in Affinity apps).

I think that color-to-gray conversions in Affinity Photo are handled as kinds of pigment conversions, K inks. That means that they are profile-dependent and vary according to the working (document) RGB and CMYK profiles. So, e.g., when you have R128, G128, B128, you might get a gray value 36 when viewing the color definition in grayscale, which is about mid gray in perceptually handled print-oriented standard grayscale ramp in context of European-oriented print -- basically the same as when you would get Gray value 64 in Photoshop in an identical environment, but in Photoshop would then have symmetry when converting back to color-based profiles.

In Designer and Publisher there is symmetry, so entering R128, G128, B128 in either app, using any color profile and color mode, it seems that you will get 50% Grayscale value, and entering 50% gray will give you R128, G128, B128 when switching the color model. On the other hand, in Designer and Publisher, people coming from Adobe environment expect gray values to be handled and exported as K inks, not as RGB values (possibly converted to four-color CMYK values).

So, I’m not sure what exactly happens here, and why, and hope someone wiser will step in and shed some light on this.

EDIT: Part of the confusion was resolved by realizing the Affinity apps do not use the same color preferences across the apps, so the different behavior between the apps described above was explained by using Dot Gain 15% as the grayscale profile in Photo, and having the default D50 (basically the same as Gamma 2.2 in PS) active in the other two apps. But the problem with non-symmetry exists in all apps, so you whenever you have a profile where e.g. RGB 128, 128, 128 does not give Gray value 50, you get constant conversion of colors when switching between the Gra and RGB color models in the Colors panel (or constantly changing and misleading color values for the same actual colors, in case you have the lock turned on).