please see my channel coefficients
The values donāt matter. āInvalidā only applies to the CCT and means it is meaningless in terms of temperature because the current illuminant is too far from a daylight spectrum. Whatever coefficients you have are ok as long as they make greys achromatic at the output.
if I select the second grey patch from bottom in color calibration module it says
CCT: 4223 K (invalid)
illuminant custom
hue 7.3
chroma 21.9
gamut compression 1.00
Invalid means ādisregard the 4223 K figure and donāt even try using a daylight modelā. Everything else will behave as expected. Itās an information, not an error.
I really appreciate your and all the other great developers work on darktable. Thanks all for this precious piece of software.
Regarding your latest gadget of integrating the support of a color checker I have the following question:
As far as I understood, it would be no problem to add support for other color cards (like @priort already did) as long as they are built as an n*m matrice.
Well, I have one of these nice IT 8.7 C1 color cards from coloraid by Mr. Wolf Faust. Those are build as a 12x22 matrice plus the 24 patches large grey scale below the color matrice itself.
After looking at the colorchecker_DT Mod.txt file, I think that such structure wonāt be compatible with the current implementation? But perhaps Iām just wrong and you may enlight my mindā¦ 
Thanks a lot in advance.
you might try to include it yourself in an individual build. Since the measurement values differs between batches of the target it doesnāt make sense to bake one of them into the code.
1 Like
I have one of those Faust target and it is mostly unusable in real-life conditions (because itās nearly impossible to avoid reflections). Also you wouldnāt carry one in your photo bag to shoot a quick reference shot, since itās big and fragile.
That color calibration tool is intended to be a quick way to extract the color calibration values from a color checker, on location. Getting a glossy IT 8 chart properly lit with even light and without reflection is a job for a real metrology lab, with diffusers, and derivating a proper color profile from that is better fit for Argyll CMS or such.
Finally, as Martin said, the batches of targets are inconsistent, which is not a problem if you use them with Argyll CMS since Wolf provides the custom .cie values in a file with the chart and Argyll can extract them, but itās really not the scope of this tool (in which more patches does not necessarily decrease the final delta E of the fitting).
4 Likes
Does this module work in linearRGB?
If you hover over the module headers you can see which color space they work in. For color calibration you get:
1 Like
Does this module work in linearRGB?
yep, for some definition of āRGBā.
1 Like
thanks
I have a question about this module. Here and in the documentation it is said that colors in the gui are projected to sRGB. My monitor is not calibrated to sRGB, does this change anything?
There are also some small errors in the documentation, not sure this is the right place to point them out. Most are typos but one is incorrect math terminology: The neutral element of matrix multiplication is called identity matrix, not identify matrix.
If you want to suggest edits to the documentation you can do so by raising an issue or a pull request against the dtdocs repository here.
That was just my fat fingers, and it got copied over from the old channel mixer page 
Probably there is a better way to explain it than matrices ā I tried drawing one diagram that never got merged. I think AP also drew a different diagram to depict color mixing, but I donāt know what happened to that. I guess weāll eventually arrive at something that explains clearly without over-complicating itā¦
I think if it is a math operation, it should be explained as such. Anything else like diagrams would be an addition for people not familliar with the maths.
Some people are scarred/allergic to math. And some do not have necessary math imagination developed. So for those a visual explanation would be better suited than a math lesson
1 Like
Great pointā¦Not all people are hardwired to learn the same. It doesnāt mean they canāt learn or understand they just do better with a different deliveryā¦
2 Likes
Very much looking forward to @anon41087856 video in English about the module. Not pushing, as it is a huge amount of work to do this in two languages!
I have briefly watched but not studied the French video. Gave me a good workflow. In the meantime I might go through it, take notes, then ask some questions here. The English version would help greatly, as has the ones on filmic and tone equalizer. The videos do teach in the way I learn 
I have a become a huge fan boy of this scene referred way of working. It is faster for me with more pleasing and, more importantly for me, more predictable results.
@anon41087856 do you have something like a patreon yet? You could have kept all of this knowledge and tech to yourself to make your photos just better, and I appreciate your gift.
1 Like
There is an english video explaining the module in some detailā¦Gamut freak rants videoā¦
https://www.youtube.com/watch?v=vVoIx3eMxwA and
Do you perhaps mean the rbg colorbalance module. I believe he has not done an English presentation for that oneā¦
@priort:
Yes. I meant the color balance rgb module.
My apologies, and my fault for reading/posting on my phone.
Butā¦ I did mean the heartfelt thanks to @anon41087856