The option to specify preferences for directly debayering raw DSLR files has been removed from Siril 1.2.0 beta. I didn’t find a reference to this in the release notes but I assume it’s by intention and not a bug.
It’s not easy to find but the corresponding line in the changelog is “Replaced Libraw with RT debayering”.
The reason is that we had inconsistencies between the two approaches, and the demoisaicing algorithms provided by RawTherapee were already used in the main processing workflow (after calibration). You can use those now to debayer you RAW images, I don’t quite remember what it changes in the GUI, but it should still be possible during conversion.
1 Like
Got it, thanks. Just for your information, the change removes the DSLR RAW Debayer UI from the preferences window. I used this quite often because it had a couple of features I found useful mainly for troubleshooting, especially the ability to easily set a custom white balance and display TRC.
Yes. I did the changes and I assume that ^^
Unfortunately, a side-effect of this change is that I have a number of older images where I used the “DSLR RAW Debayer” option for my calibration frames. This is bad practise, but it was also years ago, and I didn’t know better 
Since I kept only the calibration masters and not the RAW calibration frames, I will probably always need to keep a pre-1.2 version on my system just to be able to work with these old data again, unless you folks can put the function back in 
If you did this, then calibration frames (i.e darks and biases) don’t apply. Even with old Siril. Indeed, if your files are already debayerd then it is nonsense to calibrate the data.
They absolutely calibrated badly, and it was an important step in my learning process a couple of years back to realise that and stop using the DSLR RAW debayer option.
But I think that my results with this old data, once debayered, are still better calibrated than not.
I understand the reason the function was removed, and I’ll just keep a pre-1.2 version around in the event I want to reprocess this older data.
And you’re probably wrong.
First, debayerd data have the bias that is already subtracted.
Second, subtracting dark will introduce artifact instead of removing hot pixels due to the interpolation algorithm.