How to get correct values for L, not averaged ones

[Didge]

Hello,

I've searched the forums here and couldn't find an answer to this question ... so I hope someone can help me ...

I'm doing some experiments relating to image resolution, and I want to turn an RGB image into 4 separate L, R, G and B layers, then recombine them to re-create the original image - basically like an LRGB astrophotography image.

I expected the HSL filter with saturation set to zero to give me a layer with the **true** grey values for each pixel, but it doesn't do that - it actually calculates the average of the RGB values, then assigns that as a grey value.

So for example, if a pixel is RGB = 55, 67, 82, I want the L value for that pixel to be equal to the base gray level for that pixel  :-

L = 55

and the corresponding values in the RGB layers to be the difference between the original RGB value and the base gray level :-

R = 55 - 55 = 00

G = 67 - 55 = 12

B = 82 - 55 = 27

This way, I can then add the RGB values (00, 12, 27) back to the L value (55) to reconstruct the original RGB value (55, 67, 82).

This would give me a luminance layer, and individual R G and B layers which, when added together, are the same as the original image.

But AP doesn't work that way.

If I set the HSL saturation to zero for the same pixel, the value of L that AP gives me for that pixel is the average of the RGB values :-

L = (55 + 67 + 82)/3 = 68

rather than the minimum value (55) - which is what I expected.

Can anyone help me create a layer that is a grayscale image, with the value of each pixel equal to the minimum value (not the average) of the RGB values for each pixel in the original RGB image ?

Many thanks

[David in Яuislip]

This procedural texture will take the minimum RGB value and apply to all channels - check a random pixel with the effect off and on

 

This one subtracts the minimum grey values

 

Now do a merge visible using both procedural textures in turn, put the grey one at the top, change its blend mode to add and the original image is rebuilt. Hope that's of help

[NotMyFault]

Affinity uses the Rec.709 coefficients to calculate luma in HSL

https://en.wikipedia.org/wiki/Rec._709

When encoding Y’C_BC_R video, BT.709 creates gamma-encoded luma (Y’) using matrix coefficients 0.2126, 0.7152, and 0.0722 (together they add to 1)

[Didge]

Hi David,

Many thanks - that is **exactly** what I needed - thank you so much.

Surprisingly simple when one knows what they're doing

Cheers,

Gary