Some of the first targets of my newly-acquired QHY294C OSC were galaxies like M66 and M81 and as I shoot from a light polluted location, I didn’t immediately notice that once processed in Siril, they came out black-and-white - I was shooting in bad conditions, didn’t get much time on target, etc. But one night I got over three hours on the colourful Whirlpool Galaxy and was rather surprised not to see a lick of colour, especially in the companion galaxy. I say this because when I would attach my Fujifilm X-T1 and shoot for even 30s it would pick up colour, so there should be no question that hours of imaging ought to be picking up colour.
So I looked through my data sets and discovered I had an early set of just ten usable 3’ images from the Horsehead nebula I had shot very early and hadn’t processed. So I processed it too… and it came out black and white as well.
Here’s how I was doing it: I figured, since the files were already FITS, that they didn’t need to be “converted”… to FITS, as I had been doing with my Fujifilm RAFs. I would point Siril at the directory, then on the Sequence tab press “Search sequences” and a .seq would be generated. If they were Darks or Flats, I would move on to the Stacking tab and stack them. When it got to the Lights, I would go next to the Pre-processing tab, set the dark and flat … and set “Debayer FITS images before saving” as I assumed this was needed to actually get colour. I’d then go on to Pre-process, Register and Stack.
Once it was done, there would indeed be three channels in the channels window and they even looked a bit different. Over on the RGB window the image would initially be green-dominated - so it looks like there’s “colour”. But once I removed the green noise and/or calibrate the colour, it would come out largely as black and white, with some amount of colour noise (but never colour signal).
Like this (10x180"):
(the image itself is pretty poor - it has a splotch top right which I didn’t get a flat to calibrate out, which probably explains why I hadn’t bothered processing it earlier)
This outcome didn’t make much sense as I could see even with the KStars FITS viewer that there was some colour in the Flame Nebula in each individual frame, so it’s not like there wasn’t colour. Initially I thought it might be related to the fact that KStars and Siril have a disagreement over correct image orientation (Siril renders them horizontally flipped compared to KStars), but I set the “Debayer FITS files in up-bottom orientation” in the Preferences to address this (even so, the images still come out flipped).
After a bit of head-scratching, I decided to carry out a test: “convert” my FITS as I would RAFs … to FITS, using the “Keep three channels” and “Debayer” options. Once converted, I proceeded as usual through the various steps.
Here’s what came out, after some basic colour calibrating and stretching (and horizontal flipping):
So what’s going on here? Why do FITS files need to be converted into other FITS files prior to further processing? I had rather hoped I’d be able to skip what seems a needless step by having a dedicated astrocamera.