LAB curves control points picker bug

[JC-aff]

When using LAB curves either in an LAB document or via curve colour space option.

The test shows Bopp only but all three channels behave the same.

When setting control points on the curve using the picker, blue points should appear below halfway on the graph central cross, yellow points above and in a colour channel grey points always at 0(center -no colour).

Currently, as shown in the images, control points are spread across the entire curve-- behaving more like the tone, not the colour is being sampled.

Each image displays samples taken from either: blue, yellow or grey colour squares respectively.

I don't know if V2 has this same behaviour, but it would be something to continue putting me off upgrading.

Edited September 1, 2023 by JC-aff
Info

[NotMyFault]

Can you please upload the test document?

I cannot reproduce using the photograph of the screen - the coloured patches are not uniform, and in my tries the picker aways picket the correct La / Lb points. The patches on bottom have too much color noise.

The color picker seems to detect lightness only, and does not recognise colors at all. 

This applies to all color models (RGB and LAB).

In other words: 

if you use the curves adjustments, choose one color channel, the node position depends on the lightness value of the picked pixel, not on the selected color channel in the curves adjustment. But the node y position impacts only the selected color channel.

This becomes most obvious when you choose the alpha channel, see my older feature request which gained no traction at all.

I'm sure Affinity will rate this as "works as design", not a bug.

 

Edited September 2, 2023 by NotMyFault
revised

[JC-aff]

Sorry. haven't had time to get on the PC.

Here is the file. 

Current control points in the BOpp curve are random Blue, Yellow, Grey just to illustrate how they distribute across the entire curve.

LAB testing curve picker.afphoto

[NotMyFault]

Thanks. Using this file i can only confirms what have been said before. 
 

The color picker for curves adjustments uses lightness values only. It does not reflect the selected color channel. 
If you use the channel spaniel to show only lightness channel, you see what is happening.

Unfortunately the colors patches have some minor deviations, which are not shown in info panel or color panel in LAB mode with only 1% resolution.

It summarize, the color picker might not work as you expect, but for me it seems „works as designed“.

 

[JC-aff]

Yeah unfortunately devs either didn't consider how LAB channels work or just allowed the curves interface set to LAB to simply not work the only way it can... 

[JC-aff]

A very similar issue exists with CMYK curve picker also it appears...

Wrong place i know -sorry. As it's obviously not going to get "fixed" thought i'd put it here as this is another picker issue.

You can't pick a colour on an image and have it append to the correct position on the curve. Example Cyan 100%, 50, 10 all append in the top half of the curve; this is consistent across M and Y --in fact you could select Cyan in the Yellow curve and the points are appended in the same place as the Cyan channel...black and white both appear correctly as end-points...not unsurprisingly. Colour correcting accurately with curves is surely a fundamental need.

I'm truly amazed this hasn't been discussed to death (I can't find much relevant).

I don't suppose V2 behaves differently does it?

 

CMYK-picker test.afphoto

[NotMyFault]

I don’t understand why you keep raising „new“ reports covering the same topic. 

The behavior of the color picker for curves is consistent across LAB, RGB, GREY, CMYK. The same for V1 and V2. The same for Photo, Designer, Publisher.

It is not rated as bug by Affinity, just works as designed.