Since Noel is back with us, I thought this header might catch his attention. He was the one who first reported it six or seven years ago.
The reason I'm bringing this up now, is that someone in the Lightroom forum had this problem just yesterday (he was sending ProPhoto files to Photoshop and couldn't understand the cyan cast in his dark grays). There's no doubt it's the same issue.
This made me dig up my old ProPhoto test gradient file. I don't use ProPhoto all that much, so I haven't really looked for it in a while. To recap, this is an example of what it could look like (this is an old screenshot that I've kept). Open and view against a darkish background:
But when I opened the test gradient now, there was no trace of it, zero. The gradient is smooth as silk. That surprised me, because while the effect can vary, it always seemed to be there in some form.
Technically, this is an inaccuracy in the conversion from ProPhoto into the monitor profile. This conversion is performed by OpenGL in the GPU when you have the PS preference set to Normal and Advanced modes. In Basic mode it's shifted back to the CPU, which is more accurate so the problem disappears. And you only see it in ProPhoto files because ProPhoto is very compressed in the shadow values compared to other color spaces. The stratospheric gamut has a price, and that's it. With this compression, inaccuracies get amplified.
So what has changed on my system over the last 12 to 24 months? Not much. I still use Eizo Colornavigator as I did then, still producing similar matrix monitor profiles. But one thing has changed - going from GeForce to Quadro, and with that, 10 bits per ch output.
So Noel, if you're reading this - I understand you have Quadro GPU too, and running displays at 10 bit. Do you still see it?
Anyone else?
Explore related tutorials & articles
80
Replies
At this point I'm convinced there's something special about the ProPhoto profile itself.
A (small) new piece of info. I had thought that my own software was rendering the grayscale completely as gray, but lo and behold (and this is something I see in @rob day's post above as well) with extreme saturation there are tiny color errors in single pixels right at the transitions between adjacent levels (in the 8 bit screen grab world). That probably made no sense. This hyper saturation-enhanced comparison may help (Photoshop display on top, what I *thought* was a correct rendition on the bottom, but look closely):
Those tiny lines show up in @rob day's screen grabs above if you push the saturation way up.
Dag, please do a screen grab of the gray gradient as displayed on your system, hyper-saturate it, and post it here.
I wonder if these color inconsistencies are just a matter of magnitude, always there (because of something about the ProPhoto profile) but being (mostly) covered up by roundoff error in some cases and not in others.
-Noel
I had thought that my own software was rendering the grayscale completely as gray, but lo and behold (and this is something I see in @rob day's post above as well)
My (OSX) capture is monitor RGB with my monitor profile assigned. Before I uploaded it I converted it to sRGB, so the values are no longer perfectly equal after the multiple conversions.
https://forums.adobe.com/people/Noel+Carboni wrote
Dag, please do a screen grab of the gray gradient as displayed on your system, hyper-saturate it, and post it here.
OK, this is a screenshot of the original ProPhoto file as displayed.
Then I put 6 hue/sat layers on top, cranked them all up to 11 on the dial, and flattened.
To minimize random errors in post-conversions, this is a 10-bit screenshot pasted into a 16-bit file, monitor profile assigned > converted to sRGB, then converted down to 8 bit and saved as PNG. Phew.
My own sneaking suspicion, echoing Noel's, is that ProPhoto is simply too large for its own good. Insert a dust grain of a tiny imperfection, and bam - it blows up.
https://forums.adobe.com/people/D+Fosse wrote
https://forums.adobe.com/people/Noel+Carboni wrote
Dag, please do a screen grab of the gray gradient as displayed on your system, hyper-saturate it, and post it here.
OK, this is a screenshot of the original ProPhoto file as displayed.
Then I put 6 hue/sat layers on top, cranked them all up to 11 on the dial, and flattened.
To minimize random errors in post-conversions, this is a 10-bit screenshot pasted into a 16-bit file, monitor profile assigned > converted to sRGB, then converted down to 8 bit and saved as PNG. Phew.
I followed the above procedure and see no artifacts other than normal banding due to 8 bit display path.
WIndows 7, PS 19.1.6, Quadro 600 GPU, No 30 bit mode
https://forums.adobe.com/people/rob+day wrote
I get similar capture artifacts no matter what the profile assignment is, but they never show in the file that's being captured, either in appearance or the numbers. I'm not doubting that it's a problem, I just can't reproduce it.
Curiouser and curioser... Yes - I also see that with any profiles that don't match my reference sRGB monitor profile only when CPU-based color-management is being done, not when GPU-based color-management is being done. This is a surprise to me; I've always assumed the grayscale was pure, because without so much extra saturation the errors are not perceptible.
Each of these are gray gradient fill layers expressed in the listed profiles, displayed on my monitor, screen grabbed, and saturation of the screen grab increased to +100 twice.
This forum botches up image attachments, but note the major color inaccuracy in 1 of the 4 images in the first set, and tiny pixel-width lines in 3 of the 4 images in the second set.
As displayed in Photoshop CC 2019 with GPU color management:
As displayed in Photoshop CC 2019 with CPU color management:
-Noel
Further experimentation reveals the results are changed by the choice of color engine one makes in the Edit > Color Settings... If for example I choose the Microsoft ICM color engine instead of Adobe (ACE), I see a noticeable shift in the color inaccuracy and brightness of the dark parts (they're brighter with the Microsoft ICM). I don't see a difference by tweaking the 3 checkboxes, but I haven't done an exhaustive test.
Here's what I've got for color settings:
Note that closing and restarting Photoshop is required to see the change take effect.
-Noel
This is the one I have; looks to match yours, Todd, though of course we're not seeing all the detail in the tables.
-Noel
it would introduce inaccuracies near black
Shouldn't that show up in the print output numbers? I'm not seeing any problems in my #45 test. With six color samples in the shadow end. Bottom is ProPhoto, top is the ProPhoto file converted to Adobe RGB, the info panel is showing GRACol CMYK:
You're not seeing it on screen either, so no reason to assume it automatically shows up in print output.
The whole thing seems unpredictable, that's why it's so puzzling. Clearly there is some marginal "butterfly effect" that goes this way or that depending on some unknown initial circumstances.
Maybe we need to dive into chaos theory to explain it...
AdChoices