Comparing methods to correct gamut clipping in Darktable

[Neutral_Gray]

Gamut clipping has been discussed in several places here over the years, but this discussion went on across several versions of modules, workflows and DT iterations.

So I wonder:
What is the current “mainstream” recommendation to correct gamut clipping or to perform gamut-remapping in a best-practice way?

To be sure, I’m not making things up by only looking at the Gamut warning:

@anon41087856

But don’t overthing gamut issues, or make them up based on theoritical concerns you half-understand.

The standard case (instead of yet another sample RAW): highly saturated blue LED lights in a dark scene are re-mapped to deep blue by default.
(DT Version 4.8.1, input color profile: standard color matrix, CR2 from Canon EOS 7D)

Some options I could find:

• Limit the input color space, e.g. to “linear REC709” or “linear REC2020” (input color profile module)

@houz

I can confirm that the blue bridge goes all squirrely with the default matrix. With Rec2020 as input profile it looks way better.

• Activate gamut clipping (input color profile module)

@pk5dark

Maybe you can try different settings in the input color module like turn on gamut clipping or other RGB profiles?

• Use a camera-specific ICC profile (input color profile module)

@ggbutcher

you’ll make a linear gamma profile for your camera that’ll work just fine in a linear workflow.

@ggbutcher

Now, with my sunlight_calibrated camera profile

• Use a camera-specific LUT

@ggbutcher

A well-constructed LUT will provide the information to do just that, approximate a gradation in the destination colorspace.

• Activate and set gamut compression (CAT tab in color calibration module)

@garibaldi

Note that I have gamut compression set very high on color calibration

or in this AP Video (09:18)

[EN] darktable 3 tips and tricks : fixing the blue light problem

• RGB channel mixing (R, G, B tabs in color calibration module)

@anon41087856

There is a trick I took from theACES , which involves applying a custom RGB matrix to force those blues back.

• Lower the blue channel value in the “colorfulness” tab (color calibration module)
  (comment underneath the AP video above)
• Corrections in “Color Balance RGB” module

@anon41087856

Color balance : […] push colors around to reach new parts of the color spectrum to create an atmosphere and bring life to images, while ensuring colors stay in gamut

• Sigmoid-specific: Use the “smooth” preset together with “sRGB” as the working profile in the “primaries” tab (Sigmoid module - which, of course, rules out Filmic RGB)

Perhaps you’ll also have other options in hand.
Still, the question is: Do we have any consensus on the issue?

[ggbutcher (Glenn Butcher)]

My personal opinion is that there’s always a gamut transform in the display/export transform, so you need to get that good first. And, the only real place you have to influence that is in the initial transform from the camera space. Everything else downstream of that are pre-defined matrix transforms. That’s why I recommend a LUT-based camera profile.

Any of the other in-program tools are working on top of that, so you need to keep that in mind.

Probably not a consensus builder, but FWIW…

[priort (Todd Prior)]

How are you assessing gamut…The gamut checker icon will use the gamut of the soft proofing profile and depending on the settings the overexposure icon will give it to you based on the working profile.

You could take the approach to use the defaults and keep the gamut of the working profile in bounds and then let the display and output profiles manage the conversation to a smaller gamut or you could decide on another strategy keeping in mind how your profile setting influence what you see for out of gamut…