Whitebalance of Canon camera in darktable

Hi all,

for a photo stacking I took several pictures and set the WB of my Canon 550d to sunny (5200 K). When I checked WB in darktable it showed 5322K for temperature (and 1,013). I expected 5200K and 0.

Have I been wrong with my expectations?

Thanks for your comments!

Different apps calculate temp-tint differently. Doesn’t mean that they are being read wrong. If you want to check, try switching the WB preset to daylight. If the values of red, green and blue change, it means that your camera default has different values from what darktable has. The colour values are what we call multipliers. The presets should match your camera’s.

Thanks for your response!
I have prepared a comparative table with all the presets of the camera and dt.

grafik

You can see that the values differ. I have not noted all the multipliers, but of course they also differ slightly.

Concerning pictures the influence is minimal, but you can notice it, if you look closely.

You are still comparing the temp-tint values. This is what I meant by multipliers.

exiftool output (2nd and 3rd values are green)

image

dt output (preset to “daylight”)

image

dt daylight camera daylight
red 2.146 2236 / 1024 = 2.183594
green 1 1024 / 1024 = 1
blue 1.526 1577 / 1024 = 1.540039

They are slight different as well.

PS There are other stuff to consider with respect to appearance such as colour management, base curves, etc., when we are concerned with appearance. Many people question why does dt look different from Lightroom, etc. That is another topic all together that has been addressed elsewhere in this forum.

1 Like

Both dt and RawTherapee make use of the RPU, which is a repository of raw files from many cameras. If the multipliers of the file corresponding to your camera has the same multipliers, then it either means that dt devs have a good reason why the WB values in app are the way they are, or maybe they made a mistake and you may want to report about it.

Hello @afre,
Thanks for your explanations and your patience. I already knew that the multipliers are somehow normalised to the green values, but I never looked deeper into the calculation from the raw- or exif-data.

I have compared the behaviour of my raws with the RPU-raw of Canon 550d. In dt they produce the same multipliers.

Just the WB presets of dt are different, they have differing multipliers. I guess the dt WB module does not react camera specifically, and the presets are fixed independently from the camera model.

I found an answer and a solution in the dt wiki:
https://redmine.darktable.org/projects/darktable/wiki/White_balance_presets

However, this kind of code tweaking is above my skills, unfortunately. Particularly that I use dt under ubuntu as well as windows.

But for people with better skills there is a solution to obtain camera specific presets for the WB in dt!

This is important. If you report it, someone will look at it.

Now, you can either just post this info into a redmine issue:

https://redmine.darktable.org/projects/darktable