Xtrans and ST

[Misange]

Hi Ivo and other astro afficinados,
I recently changed my camera for a fujifilm XT4 and tried yesterday my first astro with it shooting Andromede M31. 167 pictures at 3200 iso f6.8, 300mm (eq 450 mm full frame), + DOF.
I stacked in DSS directly the raw files which took sooooooo looooong ! here is the link to the file :
https://drive.google.com/file/d/10KLZcs ... sp=sharing
I later read that users usually convert into DNG with Adobe DNG converter or other soft. Although Andromede is clearly visible in each of the raw files, the DSS file is almost black. I loaded the fts in ST as if it was a linear Bayer eventhough its not. In mono the image seems rather good actually. Untill I get to the color image ! all green. I managed to correct a bit but the colors remain Ugly. I read a previous topic in this forum on Xtrans but it dates from 2016 and I hope that others faced this question and can help with the best process to follow.
Help much apreciated !

[fmeireso]

The image came in orange. I think that the iso setting 3200 is way to high.
For my Canon 800D i use 400 iso. Some DSRL better iso 800.
Higher iso setting causes the LP to lit up in the background as orange. Maybe this the main reason; But see what others have to say.
There is also vignetting and some severe dust spots. I assume you did not incorporate flats/bias in the stack.
Anyhow the picture is better then my first try at the time...

M31 (Large).jpg (352.33 KiB) Viewed 4823 times

[Mike in Rancho]

Hi Misange,

I found the same as Freddy did, and maybe a bit more? I did a deep crop to get away from the vignetting and the dust spots. Those were strange, almost unsure if all of them actually are dust? Kind of seemed like hot/cold pixels gone wrong perhaps. I also used a Wipe mask which helped a good deal, but when I got to color I finally gave in.

From what I could tell (and I double checked opening in both DSS and Gimp), you have no blue channel here. Or it's very very attenuated, such that you effectively have no blue channel.

Here's all I could quickly find as a DSS discussion on X-Trans, perhaps you can find more though? https://groups.io/g/DeepSkyStacker/topi ... s/79249311

For my final try at it, I figured hey, ST can do some kind of missing channel interpolation, right? So I loaded the file into R and G using compose, leaving B missing (since it really is missing!), then when I got to color I went to the matrix area, but it only seemed to offer me options for "interpolate G." But I wanted to interpolate B. Maybe I need to go read the manual?

[Misange]

Hi Freddy and Mike,
Thank you for your help and for having downloaded the file and tried.
I dont think that iso is the problem for rotted colors. It could generate some noise but not get rid of the blue channel as seen here. Yes the image includes darks offsets and flats. There was an horrible large spot on the sensor ! (now cleaned).
I saw the conversation you mentioned Mike in the DSS forum and I will try to convert the raws into DNG before stacking the pictures. I am surprised not to find more infos about xtrans and astrophoto. I'll post results if I can improve Thanks again.

[hixx]

Hi, the problem with x-trans is, it does not have a regular Bayer matrix, it employs a 2R:5G:2 ratioinstead of 1R:2G:1B
Hence color shift is to be expected. You will need a Stacker capable of X-trans demosaicing - I think Astro Pixel Processor supports that format but it's not free

cheers,
Jochen

[LuckyEddie]

Hi Misange,

I use a X-T2 for my astrophotography. I recommend Siril for stacking - it's fast, free and handles the X-Trans sensors. It will work with Fuji .raf raw files - it's converts them to .fits for stacking. One thing to watch for is that Siril requires a lot of free disk space for temp files and that needs to be on a SSD to get fast processing times.
Bringing the Siril default .fit output into StarTools the colours looks fine - the only thing missing is that there are no Matrices for Fuji cameras in the colour module but that's minor.
Happy to provide more details on my workflow if you require.

[Misange]

Hi Misange,

I use a X-T2 for my astrophotography. I recommend Siril for stacking - it's fast, free and handles the X-Trans sensors. It will work with Fuji .raf raw files - it's converts them to .fits for stacking. One thing to watch for is that Siril requires a lot of free disk space for temp files and that needs to be on a SSD to get fast processing times.
Bringing the Siril default .fit output into StarTools the colours looks fine - the only thing missing is that there are no Matrices for Fuji cameras in the colour module but that's minor.
Happy to provide more details on my workflow if you require.

Sorry for the late reply, I was out for a while. Thank you ++, this is exactly what I was looking for : a user I never used Siril and I think that there is quite a bit to learn to use it correctly but I will try. Yes I will be very grateful if you can qhare more details on your workflow.

[Misange]

Hi, the problem with x-trans is, it does not have a regular Bayer matrix, it employs a 2R:5G:2 ratioinstead of 1R:2G:1B
Hence color shift is to be expected. You will need a Stacker capable of X-trans demosaicing - I think Astro Pixel Processor supports that format but it's not free

cheers,
Jochen

Thank you Jochen, I understand now that DSS is not the good choice. I will try astro pixel if there is a free trial or will try Siril as recommendend by Lucky Eddie.