Hello,
I'm trying to make sure I understand how things are supposed to work so I'm hoping someone can confirm my understanding of the below and help me figure out to achieve certain things:

1) Under Mac and Linux unless applications are color managed, ICC Profiles don't do much. Under Mac OS specifically, 1D LUTs for VCGT that maybe included in ICC Profiles can affect black/white points and gamma, is this correct? (source DisplayCal and ArgyllCMS documentation)

2) Is there a way to reset/nullify the VCGT under Mac and Linux? Under Linux I used ArgyllCMS tools to verify the status of the loaded ICC profile as I can't seem to find any other tools to at least verify. No idea how to do that under Mac though, any thoughts?

3) Is there any easy way to make sure Mac or Linux are not messing up my colors/gamma for the GUI display (as opposed to dedicated monitor which wouldn't have ICC Profiles)? I have SpaceMan ICC, can I generate a Null ICC profile from there by exporting a Linear LUT from LightSpace?

4)Assuming an external dedicated monitor workflow(i.e. direct video output from Resolve or Nuke through Declink or similar), with LightSpace CMS, I'm profiling the capabilities of the monitor after manually bringing the gamma/white point/black point as close as possible to the output standard(e.g. Rec709 or Rec2020). With the main monitor(i.e. where Nuke or Resolve GUI's run), I'm doing the same thing but potentially "fighting" against the system color/gamma response implemented by the ICC profile(unless of course it's a null profile). Is my understanding correct?

5) Assuming profiling was done accurately and now I'm ready to create the calibrations LUTs for a dedicated monitoring workflow(e.g. through declink from Nuke or Resolve/Fusion): In LightSpace CMS I should be converting from the output color space of the application(i.e. Rec 709 Gamma 2.4 for the default project settings of Resolve, or Linear Colorspace for Nuke) to the profile generated by LightSpace for my monitor. Is this the correct approach?

5b) If I'm using ACEStt in Resolve or ACEScg in Nuke as the working/output color space, I should follow the same steps as above but set the source color space to ACEStt or ACEScg, and destination of course to monitor profile, is this correct?


6) For the profiling of the "GUI screen" I should do the same thing for color space conversion/LUT generation, although the OS's color management and would affect the ability of the LUT to do what it's supposed to do. Is this the right way of thinking about it?



Sorry for the lengthy questions and I really appreciate any feedback!


Regards,
IA

No, Mac applies ICCs OS wide.
Windows only applies the VCGT, unless the application is ICC aware.
Not sure on Linux...


No, Mac ICC use is a pain, as it cannot easily be disabled...
(We have never found a way to do it anyway...)
Again, not sure on Linux.


The only approach I can think of is as you say - a Null ICC.
But, will that actually make Mac colour null???
And a dedicated monitor will only Not use ICC if you are using a dedicated video output card (SDI connections).


Correct.


It is the colour space you want to calibrate the display to that matters, not the graphics program.
So Rec709 G2.4


You should be using an ODT to set the video output to the standard of the display - Rec709 G2.4


As above, all display calibration must be to the target colour space standard - Rec709 G2.4

Steve

Thanks Steve, things are much clearer now.

I'm experimenting with a few calibrations following your guides and using an i1 Display Pro. Everything works as expected and footage looks better to my eyes in Resolve when applying the resulting LUT to the display through Resolve's color management. What I can't figure out is why the gradients from the calibration images pack appear broken up/torn with my applied calibration LUT. The gradients display smooth and perfect when no calibration LUT is applied. Again normal footage from cameras included pre-graded footage looks better with the display calibration LUT, it's only the gradients that appear wrong.

I tried running the display characterization with wider cubes(10, 13, and 14) and even increased the integration time on the i1 probe to 2 seconds (also tried different probe calibration presets) but the gradients still appear broken/torn with the calibration applied. I also ran quick profiles(primaries and secondaries) through LightSpace with the calibration LUT loaded and I can see from the gammut view as well as the other curves that the display is behaving better than the original characterization.

Could this be related to poor probe performance or am I doing something wrong?

Would need to see the actual profile data...
You can email it to me.

Steve

Thank you so much Steve, email sent.