While you are working on those tests. I did some of my own on a different front. I did this test/processing in Photoshop
- I took your image as ProPhoto RGB 16 bit and duplicated that Layer
- On that second Layer, I coverted that image bits for just this second Layer to sRGB relative intent and then back to ProPhoto RGB via a Smart Object approach
- I took the difference between the original and the the second layer that went through color space conversion via the difference blend
(Note when changing the image from sRGB back to ProPhoto, if it is in Gamut (which it would be since sRGB is totally within ProPhotoRGB - it does leaves the colors the same)
- I extremely amplied the resulting difference result to show any differences from the original, and place the original above the result for reference
Note that the woman's face has just a couple bits that were touched and there so going through the color space change did not change the color numbers. So changing color spaces is not the reason you are seeing / perceiving color differences.
I used the approach above to numerical see the differences and not depend on my own screen being a narrower gamut monitor.
The few changes that are seen are in the exact locations that the color clipping window noted that would be out of gamut colors that would be adjusted.
So the search for you root cause continues. I have my suspicions of the source of the problem yet will let the data and the results of your test help guide the search.
John Wheeler
[ATTACH=full]107095[/ATTACH]
