Softproof layer of TIFF file in Adobe RGB (1998) to sRGB does not work

[Wendelin]

The softproof adjustment layer does not change the colors in the image viewer, if the proof profile is sRGB. 

THis is always reproducible.

Steps:

1. Open a TIFF file which has Adobe RGB (1998) embedded.
2. The file should have out-of-sRGB-gamut colors. Alternatively bump the saturation in an adjustment layer.
3. Add a softproof adjustment layer:
4. Select Proof Profile sRGB  IEC61966-2.1
5. To see this bug, Rendering Intent and Black point compensation do not matter.

Actual behavior:

•    Enabling / disabling the softproof layer don't change any color in the image viewer (though it should, because most of them are out of sRGB gamut)

Expected behavior:

•    Enabling the softproof layer should show the image with converted colors in the sRGB color space, conversion method according to rendering intent and black point compensation.
•    Disabling the softproof layer should turn the image colors in the viewer back again to the Adobe colors.

 

Note, my calibrated monitor is set to Adobe RGB (1998). Also, the softproof layer is the topmost layer in the stack. So I would be able to see color changes if Affinity Photo would work as it should. 

 

Cross checks that I did not do anything wrong:

• If you change the proof profile from sRGB to Adobe RGB (1998) and switch between them in this "Soft Proof" dialog box list of profiles, no color change can be observed either.
• Enabling / disabling the Gamut check does show all the colors in grey which are in the Adobe RGB file but not in the sRGB gamut/color space.
• Soft proof seems to work for CYMK print paper profiles (at least reduced contrast)
• If the TIFF file is converted to sRGB,  Affinity Photo does adjust the colors in the image viewer (i.e. desaturates them very noticably).

 

Both, 8 bit and 16 bit TIFF files have the same issue, here I show only 8 bit.

Note, screenshots below are probably depicted with sRGB colors because you ar looking probably to a browser which only supports sRGB, and the Windows clipboard possibly too. On my screen the colors are equal regardless of softproof layer checked/unchecked, and they are very vibrant Adobe RGB colors, the out of (sRGB)gamut check shows where they are:

 

softproof disabled, same colors, same histogram:

 

Gamut check:

 

With this image, it is even a problem without pushing the saturation (vibrance layer):

 

Regards

 

[walt.farrell]

Welcome to the Serif Affinity forums.

In your Windows Color Management settings, what display profile do you have set for your monitor?

My guess (and it is a guess) for what you're seeing: You have the monitor display profile set to sRGB. And Affinity is already transforming what you see in your Adobe RGB image to sRGB for display. Therefore, because the image is already being shown to you in sRGB, when you use an sRGB soft-proof layer there's nothing different to see.

[Wendelin]

Thanks Walt for welcoming and answering me.

Note, my calibrated monitor is set to Adobe RGB (1998), using Eizo Color Navigator, which also sets the Windows system profile. And I see drastic color differences when softproofing this tiff file with other color managed software e.g. Capture One.

 

 

[walt.farrell]

OK, if the monitor profile is set to Adobe RGB (1998), and it can actually display Adobe RGB, then my guess is probably wrong. Thanks for the added info.

[lepr]

23 hours ago, Wendelin said:

The softproof adjustment layer does not change the colors in the image viewer, if the proof profile is sRGB. 

I reported the same for the macOS Affinity apps a couple of years or more ago. It's probably trapped near the bottom of a prioritised list of bugs.

My workaround has been to install an additional sRGB profile with a different name (for example, sRGB-elle-V2-srgbtrc.icc from https://github.com/ellelstone/elles_icc_profiles) for soft proofing in Affinity.

[Wendelin]

Thanks Walt.

Thanks Anon2.

I think I will try this. Is there a specific reason for the specific profile you have linked? Is sRGB the only known profile which is not supported by Affinity? 

How confident can we be that Affinity Photo work correctly with other profiles e.g. the CYMK profile I have mentioned. It does adjust the contrast but does it also correctly convert the colors? 

 

[lepr]

3 hours ago, Wendelin said:

Is there a specific reason for the specific profile you have linked.

I've always found soft proofing with it in Affinity reliably shows what I'll get when exporting an sRGB version of a wider gamut document. I also recently had a Nikon sRGB profile installed when installing Nikon software, and it is probably just as good for soft proofing.

3 hours ago, Wendelin said:

Is sRGB the only known profile which is not supported by Affinity? 

I don't know. It's the only one which I've noticed to incorrectly have no effect in a Soft Proof Adjustment, but I've only had reason to try a small number of profiles.

3 hours ago, Wendelin said:

How confident can we be that Affinity Photo work correctly with other profiles e.g. the CYMK profile I have mentioned. It does adjust the contrast but does it also correctly convert the colors?

My workflows are RGB only, so I can't advise on the reliability of soft proofing with CMYK profiles.

[Wendelin]

Anon2,

Thanks a lot for your answers.

It seems that the  sRGB  IEC61966-2.1 profile is "sRGB Color Space Profile.icm" in my Windows system folder, maybe there is an issue with .icm instead of .icc files.

I downloaded the sRGB-elle-V4-g22.icc profile and voila, Affinity Photo can softproof the reduced gamut of sRGB (elle v4 g22). 

I seems odd that this issue exists for years.

 

Dear Affinity Team, if you read this: Color management and workflow is probably THE most complex thing in photography (at least for photographers) and because of this fact, photographers deserve the best possible support from software to cope with color management with ease. 

 

Thanks and regards

 

 

[lepr]

On 5/1/2020 at 7:11 PM, Wendelin said:

I downloaded the sRGB-elle-V4-g22.icc profile and voila, Affinity Photo can softproof the reduced gamut of sRGB (elle v4 g22).

Why did you choose one with gamma 2.2 tone response curve (TRC) when you can have one with the real sRGB TRC which is linear at low values (shadows) and then approximately gamma 2.4?

You should have downloaded the file I recommended or the V4 alternative of it. sRGB-elle-V2-srgbtrc.icc

[Wendelin]

1 hour ago, anon2 said:

Why did you choose one with gamma 2.2 tone response curve (TRC) when you can have one with the real sRGB TRC which is linear at low values (shadows) and then approximately gamma 2.4?

You could have downloaded the file I recommended or the V4 alternative of it.

I spend quite some time on Elle's website trying to understand what these self-made (linear) profiles are about. My conclusion was (from the what I understood) that I should use the linear gamma profile only if I understand exactly the use case and what I am doing, and I don't feel I understand enough at this point in time. Though the standard gamma might have disadvantages, it seems that this is the profile which is closest to what most people would have installed, and if I export to JPEG 8-bit (obviously) in sRGB then I do this for web use of photographic images, i.e. for other people to see my images. (Having said this I should have probably downloaded the v2 profile for browser compatibility...?)

I did not notice gamma 2.4 in the parts of her article which I have read, but might have overlooked it, a "hard" proof (pun intended) that I understood probably not even a fraction. 

I need to do more research.

Thanks

[lepr]

11 hours ago, Wendelin said:

I spend quite some time on Elle's website trying to understand what these self-made (linear) profiles are about. My conclusion was (from the what I understood) that I should use the linear gamma profile only if I understand exactly the use case and what I am doing, and I don't feel I understand enough at this point in time. Though the standard gamma might have disadvantages, it seems that this is the profile which is closest to what most people would have installed, and if I export to JPEG 8-bit (obviously) in sRGB then I do this for web use of photographic images, i.e. for other people to see my images. (Having said this I should have probably downloaded the v2 profile for browser compatibility...?)

I did not notice gamma 2.4 in the parts of her article which I have read, but might have overlooked it, a "hard" proof (pun intended) that I understood probably not even a fraction. 

I need to do more research.

Thanks

You seem to have misunderstood both me and some information at that site.

I never said to use a linear profile. I said to use a profile with the correct sRGB TRC, which happens to be linear at low values only and approximately gamma 2.4 for the remainder.

Elle also recommends a profile with that sRGB TRC, not one of the ones with sRGB primaries and alternative curve:

"sRGB-elle-V2-srgbtrc.icc: This sRGB profile can be assigned to DCF R03 camera-generated jpegs, and also can be used for editing 8-bit images. For maximum compatibility with color-managed browsers, use this variant for images posted to the internet.

sRGB-elle-V4-srgbtrc.icc: This sRGB profile can be assigned to DCF R03 camera-generated jpegs. It can be used for editing 8-bit and also high bit depth images. 

For finished images, it's better to assign the "sRGB-elle-V2-srgbtrc.icc" profile before uploading the image to the internet. This extra step probably only benefits Firefox users who have calibrated and profiled their monitors, but haven't changed the default Firefox color management settings, which is to say, probably not very many people."

In any case, if colour management is working correctly on your machine, there should be a visible difference between true sRGB and your chosen ICC profile with gamma 2.2 curve.

[Wendelin]

On 5/2/2020 at 9:03 AM, anon2 said:

For finished images, it's better to assign the "sRGB-elle-V2-srgbtrc.icc" profile before uploading the image to the internet. This extra step probably only benefits Firefox users who have calibrated and profiled their monitors, but haven't changed the default Firefox color management settings, which is to say, probably not very many people."

In any case, if colour management is working correctly on your machine, there should be a visible difference between true sRGB and your chosen ICC profile with gamma 2.2 curve.

 

Hi Anon2,

Thank you. Yes, I have misunderstood. I now downloaded and installed the profile sRGB-elle-V2-srgbtrc.icc.

I've also read more from her website and I am really astonished that the sRGB profile which was meant to standardize and ease color management seems to have diverged into so many differing profiles in the real world.

https://ninedegreesbelow.com/photography/srgb-profile-comparison.html

 

I made a simple test with one of my JPG images from an Olympus camera. The file has an EXIF tag sRGB (not the full profile) as revealed by exiftool used as described by Elle's article:

https://ninedegreesbelow.com/photography/embedded-color-space-information.html

The image is colorful especially in the red department.

Now I opened this file in Capture One 20.0, Gimp 2.10.18, Affinity 1.8.6.641.

The colors in Capture One and Gimp look quite similar, if not even identical. Affinity shows the red slightly muted and some blues are a bit shifted and desaturated.

• Capture One possibly uses the "sRGB Color Space Profile.icm" in my Windows system folder, but maybe its own "sRGB Color Space Profile.icm" profile in the app folder.
• Gimp says it uses the "GIMP built-in sRGB" profile
• Affinity it probably using sRGB  IEC61966-2.1, as configured in the preferences.

C1 = GIMP <> Affinity

I assigned the profile "sRGB-elle-V2-srgbtrc.icc to this image in Gimp and Affinity (C1 does not have this option). I cannot observe any color change. Still:

C1 = GIMP <> Affinity

When assigning sRGB-elle-V4-g22.icc to Gimp and Affinity, an can see ever so slight color change with this image. Not much so. Still:

C1 = GIMP <> Affinity

 

Changing Monitor setting sRGB or Adobe RGB does not influence this discrepancy.

 

As if color management between devices and color spaces wouldn't be complex enough for the normal photographer, industry standard profiles and /  or different applications behave differently and or have bugs.

Mankind has flown to and walked on the moon some 50+ years ago, and even built compatible space station parts even across political systems, but this digital image industry can't get a few colors right???

🙂

Thanks and regards