I decided to give Darktable 4.0.1 a try as my raw editor. So far, I really like the UI.
I am most used to RawTherapee, and when I processed my photos in RT, my crucial first step was that I always would go to color management and specifically choose an input profile for the camera I shot the picture with.
It looks like DT has the same option in the “Technical → input color profile” module.
The profile or color matrix to apply. A number of matrices are provided along with an enhanced color matrix for some camera models. The enhanced matrices are designed to provide a look that is closer to that of the camera manufacturer.
You can also supply your own input ICC profiles and put them into $DARKTABLE/share/darktable/color/in or $HOME/.config/darktable/color/in (where $DARKTABLE is the darktable installation directory and $HOME is your home directory). One common source of ICC profiles is the software that is shipped with your camera, which often contains profiles specific to your camera model. You may need to activate the unbreak input profile module to use your own profiles.
If your input image is a low dynamic range file like JPEG, or a raw file in DNG format, it might already contain an embedded ICC profile, which darktable will use by default. You can restore this default by selecting “embedded icc profile”.
Questions:
I installed DT from flathub and can’t figure out where the Darktable installation directory is, so I can’t put the color profiles there if I can’t find it.
There was no ~/.config/darktable directory on my system, so I manually created the path specified in the docs and put a color profile there and it’s still not showing up.
All my color profiles are .dcp files and the docs are talking about .icc files. Could this be the problem?
I have been using RT for so long that I forget why I even started to care about input color profiles in the first place. I do remember that when I first started editing Raw, I struggled quite a bit until assigning correct DCP profiles to the pictures and that made everything work much much better.
After reading this topic I created an icc profile for my D7000 but in the input color profile drop-down menu I only see an “unknown camera” listed. I’ve tried many combinations of upper/lower case, space/no space file names but I can not get it to display with the fill name that I created. Was your file name displayed in the drop-down menu of input profiles?
I think you might need to edit it or do it again…the name displayed is not the file name of the icc but rather the description within the icc…depending on how you made it you may have left that blank??
After a few years with RT, I am now looking for the right setting for the camera color profile in dt 4.0. How do you do that? Do you need this at all, as in RT?
Or is “embedded Matirx” the right setting?
If you talk about ‘embedded matrix’, it seems to suggest you are opening a DNG file, and you would use whatever is in the DNG in this case. Seems fine.
If not, move it to ‘standard’ if you want to use DT’s own setting for that camera model. But if it works depends on how ‘true to the raw data’ your DNG is (if it’s processed all bets are off basically, depending on what wrote the DNG file).
Yes, I already thought that it is much better to use the appropriate camera profile in RT.
Not the Camera standard, but the specific.
Better the Auto-matched camera prifile
or the Custom
In my opinion, this makes much better colors and is so recommended by the RT instructions.
Am I wrong about this?
You are right. You don’t really want to use “Camera standard” in RT…
As for darktable: As far as I know, the “Standard color matrix” is (mostly?) based on Adobe’s DCPs. So, if your photos open and use this, you can assume to be fine. So mostly there shouldn’t be a difference between “standard” and “embedded”, since both are Adobe DCP. (if the DNG is from Adobe DNG Converter of course)
Thank you. I did run across your old post when I first started researching this. Dcamprof wanted me to compile, which tried to do, it complained about not having dependincies, I went to look at the dependencies wanted to be compiled also and at that point I felt I was in over my head.
I’ve tried to compile programs 5 or 6 times in my life and it never works smoothly. I don’t know how you guys do it so easily!
dcamprof is pretty easy to compile compared to the other software discussed here, but dependencies can be daunting. If you’re using windows, I have a .zip file:
I think you meant ~/.var/app/org.darktable.Darktable
Regardless, I poked around in those directories and didn’t find anything resembling the directories mentioned in the DT 3.8 docs.
But I just saw that 4.0.1 DT is available as a snap package. I generally hate flatpaks because I don’t understand how they work with Linux permissions and they sometimes give weird errors when updating. I will take off the flatpak DT and install the snap. Maybe the directory structure will be closer to how they describe it in the documentation.
I was able to grab the file before the VPS got hosed and I did convert one DCP file to ICC. dcamprof told me it was successful, but it didn’t show up in the DT input color menu, maybe because I did not put the ICC into the correct directory.
All this monkeying around has me wondering if I should even bother with input color profiles. Setting input profile was always my first step when I started editing pix in RT. 9/10 times, I used the profile’s built in tone curve as well and that was a great starting point for further editing.
Maybe DT is different and the input color is not so important.
So anyways, if anyone wants the compiled version of dcamprof that @ggbutcher posted a little while ago, I’m attaching it to this message.
Making profiles from target shots can be useful if you’re shooting under other-than-daylight and need good particular colors, such as for studio portraiture.
But for my money, there’s nothing like a camera profile made from measurement of its spectral sensitivity. With that dataset, you can make LUT profiles to control extreme colors, profiles for any color temperatures, and you can train them against various datasets such as the LIppmann_2000 skin tone dataset. Also, you’re not fighting glare issues as with target shots.
and I also have datasets from other sources where I couldn’t establish a license for publishing; let me know your camera make/model and I can tell you if I have a corresponding dataset.
.
