Darktable and X-Rite ColorChecker Camera Calibration Software

the darktable way: https://pixls.us/articles/profiling-a-camera-with-darktable-chart/

An old way dealing with icc profiles: https://encrypted.pcode.nl/blog/index.html%3Fp=594.html

A more illustrated description: https://tomassobekphotography.co.nz/articles/create-camera-input-profile-with-ColorChecker-Passport.php

4 Likes

There’s quite a lot of material about input profiles / camera calibration in this forum. It’s often quite complicated! I’m fairly new to darktable and recently asked about this here - Camera input profiles in darktable

Thanks! After I did research the old way dealing with ICC profiles it seems that the workflow in general is similar to Capture One and X-Rite ColorChecker Camera Calibration Software workflow. Both approaches generate ICC profiles from flat white-balanced 16-bit TIF files. The difference is that the latter approach is more straight forward to create calibration profiles on the fly when developing RAWs for each shoot for constant color accuracy.
Wonder if there are any GUIs for ArgyllCMS to generate ICCs faster with few clicks so I can manage my workflow without a Windows machine.

I wrote a simple bash script to run the two ArgyllCMS programs needed to produce a camera icc file from a target shot of a ColorChecker:

#!/bin/bash

scanin -dipn -v -G1.0 -p $1.tif /usr/share/color/argyll/ref/ColorChecker.cht /usr/share/color/argyll/ref/ColorChecker.cie
colprof -v -am -u -C"No copyright, use freely." -O$2.icc -D$2 $1

Use it something like this:

$ ./cam.sh <infilename_withoutextension> <outfilename_without_extension>
1 Like

@MStraeten Martin thanks for sharing the links…all high quality…what is your opinion on processing the target to produce a DT Style vs an ICC file?? So I guess Darktable chart vs Argyll??

In my experience the icc way is better if you want to get ‚real‘ colors, the dt chart style approach is better to match camera presets. i never was able to get satisfying results using darktable chart with referencing the cie file …

The dt chart approach requires the use of tonecurve to get proper colors and so it‘s not the best way if used in a filmic workflow (out of the box RAW rendering not up to JPEGs)

Thanks for your perspective. I appreciate you taking the time to reply…

with Argyll you can use DCamProf http://rawtherapee.com/mirror/dcamprof/dcamprof.html to fine-tune or make a better ICC Profile

Did you ever just get the DCP files from adobe and try to convert them to icc…you can do this with dcamprof??? Might be worth seeing how that compares to the DT ones and you can add curves and lut etc to tweak the profile??

@priort, no, I didn’t try that. @Morgan_Hardwood kindly offered to make 5DSR DCP and ICC profiles, which would be available to all RT and dt users. Now Summer is pretty much here, I’ve done the outdoor shots and I’ve asked Morgan if he’s still ok to make these profiles, if yes then I’ll do the bulb shots and upload everything.

You can get the adobe ones free. You just install the DNG converter software and you have the whole Adobe library. You can delete the software after if you have not use…The work strait away in RT…conversion
of those to ICC is a two step process….dcamprof is the best tool I am aware of for that.

Best

Todd

I tried this for my two cameras, D7000 and Z6, made a LUT ICC from the Adobe Standard DCP for each. D7000 worked fine; Z6 not so much. I didn’t dig into the possible faults…

I must admit I played with it more in RT as you can use them directly and I could enable parts of the profile ie the tone curve and the lut. I think it gave better skin tones in the images I tried
but it wasn’t a comprehensive evaluation…

how you did that? Can you please explain the process

Explained here: Help converting dcp to icc using dcamprof

Note that, instead of -p matrix in the second dcamprof invocation, you would use -p xyzlut

dcamprof can be found here, source code: https://github.com/Beep6581/dcamprof. dcamprof is rather easy to compile, so I’ll leave that up to you…

2 Likes

etc. Thanks for the suggestion Todd but it seems it’s the adobe DNG data that’s used in darktable by default already. See first Aurelien P post on Feb 24 here -

Perhaps but likely just the matrix values…you can produce an ICC with several modifications including a LUT and tone curve version….

You over-estimate my capabilities!

Many experts here, can someone please make a detailed tutorial for mortals like us please!

Morning, @Rajkhand,

A good start is to search this forum.
Enter search word

torger

Have fun!
Claes in Lund, Sweden