Camera Support #10868

Canon 760D noise profile

Added by Bruno Pagani about 2 years ago. Updated 5 months ago.

Patch attached
Start date:
Due date:
% Done:


Affected Version:
git development version



I recently acquired a Canon 760D camera and would like to know what exactly is needed in order to support it. According to this page:, it currently lacks WB preset, noise profile and colour matrix. I think that last one is going to be fixed by, but the other two remains, especially noise profile that currently makes darktable unusable for me. :(

There might be other things too, but the following page (referred from previous one) looks (quite normally after all) outdated:

Especially, I think that with 2.0 you removed libraw or something like that. So, my question is which steps should I follow exactly to bring all what is needed to support my camera in darktable and projects that darktable depends on (like capture this kind of pictures with those kinds of settings and provide us with them or run this script here). :) Also, it probably worth it to add a (wiki) page about this, or just directly to the aforementioned page. And if other projects are involved, links to similar pages for them. ;)

dt-noiseprofile-20160316.tar.gz (4.54 MB) Bruno Pagani, 03/16/2016 11:10 PM

dt-noiseprofile-20160430.tar.gz (3.94 MB) Bruno Pagani, 04/30/2016 05:30 PM

dt-noiseprofile-20160924.tar.gz (6.26 MB) Mandar Harshe, 09/25/2016 12:17 AM

dt-noiseprofile-20170628.tar.gz (2.15 MB) Bruno Pagani, 06/28/2017 01:21 PM

Related issues

Related to darktable - Camera Support #10613: Enhanced Color Matrix Canon EOS 760D New 08/24/2015


#1 Updated by Roman Lebedev about 2 years ago

#2 Updated by Bruno Pagani almost 2 years ago

  • % Done changed from 0 to 70
  • File dt-noiseprofile-20160316.tar.gz added
  • Status changed from New to Patch attached
  • Target version set to Candidate for next minor release

While waiting for more input from your side, I took the time to do some noise profiling. So here you go.

I’ve kept the original pictures if you need them too.

#3 Updated by Pedro Côrte-Real almost 2 years ago

  • % Done changed from 70 to 20
  • Status changed from Patch attached to Incomplete

The noise profile graphs look very strange. Could you please upload the raw files somewhere so we can have a look?

#4 Updated by Bruno Pagani almost 2 years ago

Yes, sure. Can’t upload the whole things from home, but will do from work on Tuesday afternoon.

Note that, possibly correlated to this strange look, I’m not happy with the results in DT. Picture look like pastel or something like that.

#5 Updated by Pedro Côrte-Real almost 2 years ago

Yeah, that's to be expected. The fits were not good at all.

#6 Updated by Bruno Pagani almost 2 years ago

Here it is (available for 60 days from now):

#7 Updated by Pedro Côrte-Real almost 2 years ago

  • Affected Version changed from 2.0.0 to git development version

Hey, redoing the profiling with the original images results in something decent but not ideal. We could probably get a better profile if you included more tones in between full white and full black. See these examples:

#8 Updated by Bruno Pagani over 1 year ago

Hey, sorry for the huge delay, been quite busy since. Please find attached a new attempt. ;)

#9 Updated by Bruno Pagani over 1 year ago

Hi, any news here?

#10 Updated by Roman Lebedev over 1 year ago

  • % Done changed from 20 to 70
  • Subject changed from Canon 760D support to Canon 760D noise profile
  • Status changed from Incomplete to Patch attached

#11 Updated by Mandar Harshe over 1 year ago

I just ran the noise profiling myself and generated this profile:

Hope it's useful.

#12 Updated by Bruno Pagani about 1 year ago

What’s the status here? Looks like it missed the train for 2.2, while noise profile for EOS 80D, submitted long after, seems to be in…

#13 Updated by Bruno Pagani 11 months ago

Bumping again, Redmine seems to get way less attention than GitHub, but they aren’t issues at the GitHub repo…

#14 Updated by Stefan Schöfegger 7 months ago

The submitted noise profile differ a lot. As there was an error in creating the noise profile in older darktable versions i think the profile should be calculated again with a recent version. Afterward it can be added (and wouldn't take as long as the old profile :-) ).

#15 Updated by Stefan Schöfegger 7 months ago

  • Assignee set to Stefan Schöfegger

#16 Updated by Bruno Pagani 7 months ago

Hi, sorry for the delay. I’m currently travelling, and I don’t have either the camera or the pictures with me. I’ll try to retrieve those from my online backup, else it will have to wait for the WE.

Thanks for coming back to us! :)

#17 Updated by Bruno Pagani 7 months ago

Here you go!

By taking a look at the PDFs, it looks way different indeed, especially for ISOs of 200, 400, 3200, 6400 and 12800. Tell me if that’s good for you, else I’ll do new shots when getting back home in ten days.

#18 Updated by Stefan Schöfegger 7 months ago


please redo the shots for ISO 3200, 6400, 12800. All other iso images look good, you can reuse them for the noise profile calculation.
Thanek you.

#19 Updated by Bruno Pagani 7 months ago

ACK, I should be able to do so on the 9th.

#20 Updated by Bruno Pagani 5 months ago

Sorry for the huge delay, I’ve made some unsuccessful attempts and then had to do a lot of other things.

I’ve made another hundred of attempts during the past week, to no avail. Curves look always completely wrong. Running the code against the previous series of pictures give the correct results, but it seems that I’m unable to choose a correct view/lighting… I’ll try again later but for now I have to let this aside again.

Also available in: Atom PDF