Flats from triband filter problem

This is the first time I’m trying to photograph with the Antlia Triband RGB Ultra filter 2”, in fact it’s the first time I’ve tried using any filter.

I started by taking calibration photos. After shooting biases and flats, doing master-bias I tried to do master-flat. Unfortunately, after loading the sequence, there is no way to do normalization. I follow the instructions from Siril - Manual pre-processing

The flats despite being illuminated correctly are completely blue. I am guessing that this is because the filter is just letting the blue light through. Unfortunately the blue channel is basically white. Also when I underexpose the flats photos the effect is the same.

Important information (due to the inability to set ecos for my camera) I use jpgs, dslr (OM-1), white light source laptop or iPad.

Lights are also blue and the blue channel is also all white. The red and green channels are ok. It seems that in this case the exposure may have been too long. In the case of flats there is no such concern, the histogram in the camera shows 1/3 and +/- 0.0

Is the problem in processing the flats or in making them?

14:12:02: Reading FITS: file pp_flats_00001.fit, 3 layer(s), 5184x3888 pixels, 32 bits
14:12:02: Reading FITS: file pp_flats_00008.fit, 3 layer(s), 5184x3888 pixels, 32 bits
14:12:02: Reading FITS: file pp_flats_00004.fit, 3 layer(s), 5184x3888 pixels, 32 bits
14:12:02: Reading FITS: file pp_flats_00006.fit, 3 layer(s), 5184x3888 pixels, 32 bits
14:12:06: MAD is null. Statistics cannot be computed.
14:12:06: Normalizacja nie powiodła się. Check image 1 first.
14:12:07: MAD is null. Statistics cannot be computed.
14:12:07: Normalizacja nie powiodła się. Check image 8 first.
14:12:07: MAD is null. Statistics cannot be computed.
14:12:07: MAD is null. Statistics cannot be computed.
14:12:07: Normalizacja nie powiodła się. Check image 6 first.
14:12:07: Normalizacja nie powiodła się. Check image 4 first.
14:12:07: Stacking failed, please check the log to fix your issue.

Hystogram view of the frame:

FIT Header:

Statistics:

Histogram view of the blue channel:

I don’t know where to look for logs for Siril on MacOS.

Hello. I see many issues here.

  • First, your flat is demosaiced. You never should debayer a flat/master-flat. You must keep it as a CFA image (i.e monochrome).
  • Your statistics show negative numbers, and as you said, saturated blue channel. This may come from the demosaicing process. I don’t know, I don’t have enough information. However, most important, use your RAW flats and come back to give feedback ;).

Thank you for looking into my question.

As I mentioned: I am forced to work in jpegs. I know it’s bad, but I have no choice. OM-1 is not supported on kStars/Ekos.

Of course, I can take calibration images in RAW, but does it make sense for the calibration frames to be in a format other than lights?

I don’t really know when demosaic takes place - is it already in the camera (probably not, though) or in Siril, and if so, where does one turn it off?

And one more thing. When there is no filter, everything goes ok - in the sense manual and/or scripts processing in Siril.

If you are in jpg, then do not calibrate it.
This is nonsense.
You can divide by flat if you really need to remove vignetting, but don’t remove bias to the flat because bias has already be removed during jpg processing. However it seems your jpg processing normalize the image to the max which is a bad idea for flat processing.

And of course do not remove darks. It is useless.

Thank you very much! Your answers are prompt and very helpful.

I’ll ask still: did I switch something somewhere, because when I convert RAW files to FITS and don’t turn on “Debayer” the files convert to B&W.

I decided to check how the RAW files convert when I do flats with the filter. Also they are blue, although they are not burned out. But is this the way it is supposed to be?

If I had been working in RAW would the flats ultimately allow this excess blue to be removed? Am I completely confusing the threads and only color calibration would help with editing?

Is this an intended action?

Yep, this is normal.

Yes.

No problem with the excess blue. Why not? Because Siril converts the flat file to a monochrome flat by equalizing all CFA flux. This is done with the “equalize_cfa” option, which is enabled by default in the built-in scripts (you have to check the corresponding toggle button in the UI if you process it manually.

I don’t understand your question here.

Don’t worry about that. Now I understand. Thank you very much.