Behaviour of Curves tool in LAB space (Affinity 1)

[Zygodon]

I have been working with Affinity 1 in LAB colour space, with generally satisfactory results, but my confidence is undermined by seemingly strange behaviour of the Curves tool in the LAB space. To illustrate my problem, see the enclosed pdf. To generate this, I opened a RAW photo file and changed the colour profile to ROMM RGB, developed it, and converted the Document format to LAB/16. I added a pixel layer over the original photo, before deleting the photo pixel layer. So I had just one transparent pixel layer in LAB format. Then I used a brush to add swatches to the pixel layer. Then I opened a Curves tool, and located the input (X-axis) coordinates using the curves tool picker. I noted their values by did not change the output (Y axis) values. I reset the tool, and repeated for the A opponent and B opponent channels. Screen shots of the swatch panel and Lightness, A opponent and B opponent channels of the Curves tool are shown in the .pdf.

The swatches contain three values of L, A and B in various combinations. L : 0, 100, 50. A and B: -128, 0, 127. I expect, therefore, three peaks in the curves channel histograms - and indeed, that is what I see. My problem is that the points identified by the Curves picker are strange in two respects:

1. The X (input) values of the points are the same for all three channels, and none of the points lie over the histogram peaks. The grey50 point is at X=0.47, not X=0.5.

2. On the swatches, the text overlay shows the L,A,B values of the brush used for each swatch, and beside (for the black, grey50 and white swatches), or below (for the colour swatches), the Curves X value. The the brush values and Curves values agree for the black, white and grey channels (taking into account the origin shift and re-scaling of the Curves axes compared to the brush colours). The Curves values for the colour swatches, however, follow no relationship that I can discern to the brush colours.

My expectation was that the values found by the Curves colour picker for each channel would lie on the appropriate peaks of the Curves histogram for that channel. Do I have a misunderstanding? Can you explain this behaviour?

Thanks,

John

LAB test 2.pdf

[Callum]

Hi Zygodon,
Welcome to the forums

I have logged this with our developers to be improved in the future. I will update this thread once I have more info.

Thanks
C