Is there a risk involved in coping icc monitor profile from Window to Linux when it comes to correct color management?
Background:
I have a wide gamut EIZO CG279X monitor.
I process images using GIMP and Darktable using Linux Debian 12.1.
My monitor is calibrated and GNOME and applications are color managed using the monitor icc profile.
I use NVIDIA drivers for Quadro RTX4000 GPU.
GPU and Monitor are connected using Display Port cable.
The monitor can execute self calibration on its own using a monitor built in calibration device and store calibration result in semi persistent monitor configuration memory. From this point of view nothing else is needed besides a powered on monitor. The operative system and the GPU are not involved here.
However, I generate a monitor icc output profile with EIZO Color Navigator 7 application, which I execute on Windows 10 (in a dual boot system). I then copy the monitor icc-profile result from Windows 10 to Linux Debian and set up GNOME, GIMP, Darktable and Geeqie Color Management using this profile.
My issue:
Is the generated monitor icc profile from Windows 10 and EIZO ColorNavigator 7 a faithful representation of the same monitor when I use the icc profile also for Linux Debian GNOME, GIMP and Darktable?
I can run calibration in Linux, but EIZO only support Red Hat Enterprise. I can also stay in Windows 10, but I prefer Linux.
I will forward the same question to EIZO as well, but before I do that I am interested in your experience and feedback.
It feels safe, insofar that the EIZO CG279X stores calibrated state in monitor, GPU not involved and at least ideally, the monitor icc profile should characterize the monitor output device and not GPU + Driver + Monitor, but I do not know.
Another alternative, staying in Linux, is to skip the EIZO built in calibration device and use an external colorimeter on Linux, using open source tools. That can be done.
However, what happens in the case listed in the issue?
Thank you for your input. My answer was delayed because I have been on a business trip.
Yes, to do profiling under Linux is a way, but then I have to use an external color calibration device when I develop the icc profile on the Linux side. Displaycal can not control the EIZO monitor built in calibration unit.
It is very convenient to have a built in device in the monitor. I use the default and let it self-calibrate the monitor every 200 hours of use. This has greatly improved the chances of me actually adhering to periodic calibration.
I am still very interested in the original question:
Is there an issue when I develop icc-profile on Windows side (EIZO Color Navigator 7) and then copy it over to Linux side (Debian Gnome)?
Is it wrong and erroneous, or working in theory but risky, or is it fine (why)?
I think it should be safe. Why don’t you just test it? Do the pics look different on Linux if you use the copied profile?
I mean afterall color navigator should do profiling only without calibration after hardware calibration.
No, that does not work. The screen is profiled and the profile reduces or increases saturation depending on the color space of the screen. I know from experience that it does not work.
I suppose, but I do not know, as stated above, that since I can self calibrate the CG279X monitor ( with its internal calibration device) in a completely stand alone way, with only monitor powered on, that my calibrated state depends only on that, i.e., the monitor. I also suppose that a correctly generated icc profile characterizes the monitor, (4) in chain, and that (2,3), GPU and Nvida driver and DP are not involved here.
However I can only verify calibrated state and create monitor icc-profile using EIZO Color Navigator 7 on Windows.
In the end only EIZO knows how Color Navigator 7 is coded, I understand that, but I am also interested in the theoretical question of the feasibility of moving icc-profiles from Windows to Linux with different Nvidia driver versions on Windows and Linux.
Will (2) in systen chain, that is Nvidia driver versions and Nvidia driver differences between Linux and Windows driver versions mix things up here?
Feedback so far seem to indicate that it is OK to move icc-profiles from Windows to Linux with different Nvidia driver versions on Windows and Linux.
If I open GIMP and toggle win-win and linux-win I see no difference. I interpret this as good news when it comes to my practice of creating icc on Windows side and then using the same profile on Linux.
If I toggle win-win and linux-linux I do notice a visual difference comparing rendition in GIMP main window. This is perhaps logical, since we have used different tools and profiled at different times under different circumstances, to create the Windows profile and the Linux profile, is it not? (In theory it should be possible to get them to be the same, though).
Edit: @Claes, which tool do you use to display the profile information?
which tool do you use to display the profile information?
“ICC Profile Info”, which came with DisplayCal, I think.
toggle win-win and linux-win I see no difference. I interpret this as good news
W-e-l-l… Since this is a jungle/abyss/quagmire, let us tread carefully.
Spontaneously, I would say that it seems as if you can use your Win .icm
as a profile in for instance The Gimp.
But, because of the differing calibration curves, I would say that your
Win profile can not be used to “re-program” your graphics card in Linux.
On the other hand — I do not have an Eizo, so I do not know how it behaves.
Why not play around with it and see how it behaves?
I agree when it comes to GPU card reprogramming. However, I have one single calibrated state, stored in EIZO monitor persistent memory, after EIZO Color Navigator 7 calibration. The GPU is not reprogrammed in my case.
I do. I am fine, so far. I have not noticed that I have broken anything. But I am open to advice.
