DSS seems to vertically flip images when stacking

[Andy56]

Hi,

I've just started using Siril to batch remove Canon banding but I've noticed some strange behaviour in DSS when stacking the files.

I'm basically evaluating my 600D I have just started to re-use. It has gained some lines across the image and I'm experimenting with different processes to find the best way to remove them (or buy another camera).

I have converted 10 *.CR2 files to FITs and de-bayered them.

I then run the Siril batch Banding Reduction on these files to get a new set of FITs.

I have then run DSS to stack the fits files without calibration files.

The DSS stacked FITs are both inverted vertically but the CR2 files are not.

This does not happen if I stack them in Pixinsight.

Can anyone shed light on this.

I guess it's something I'm doing but it's seem to be a Siril/DSS thing.

TIA

Andy

[Elp]

Maybe there's a checkbox in the setting that enables DSS to flip, haven't checked it? But I don't see why you don't manually flip it afterwards in any number of software, you usually have to do this anyway with your images depending on the optics of your imaging train.

[alacant]

2 hours ago, Andy56 said:

Siril batch Banding Reduction on these files to get a new set of FITs

Hi

Siril flips the images.

Is there any reason for not using Siril to process your cr2s from start to finish? 

Siril's banding algorithm is best applied after pre processing the fit sequence. You need only a flat frame stack which has had the offset subtracted.  

Details here.

You are free of course to flip back or choose any orientation you like. Probably easier though to do it after processing.

HTH

Edited March 8 by alacant

[symmetal]

Some astro image processing software flip the images vertically while others don't. It's to do with how rigidly they conform to the fits standard. Fits was initially designed to store graphical type scientific data where the origin (0,0) is at the bottom left. Pixel image data has the origin at the top left. There is nothing in the fits file header, afaik, that states which orientation the data should be presented so it's up to whoever's processing the fits files to decide. Programs dealing with only astro images should decide that 0,0 is top left, while others stick to the original fits convention that 0,0 is bottom left.

Alan

Edited March 8 by symmetal

[Gfamily]

I can't say it's this for sure, but as a camera designed for domestic use, your Canon would automatically be presenting the image as inverted relative to how it was detected on the sensor (because that's what we expect with a camera).

It's possible that when the raw data is packaged as a FITS format file, the inversion is not applied. 

[Andy56]

Many thanks for your responses

Elp: Yes, I can flip it afterwards, and have done so in Pixinsight. I was more wondering why, it's something I don't understand.

Alacant, symmetal, Gfamily: This make sense that the origin is bottom left and not top left. This would cause the flip. 

Alacant: I haven't used Siril for processing because I have not learnt how to do so yet. I'm retired but still can't find enough time for it all. I have gained good familiarity with PixInsight and use it most of the time. I'm trying to resolve the balance between the use of software to correct lines from the Canon camera vs get a new(newish) one. Siril was a quick way to allow me to debayer images for examination in PixInsight. You have replied to my query Am I looking at a faulty sensor? with the good suggestions to overcome this issue, ie the  Banding Algorithm in Siril. I hadn't realised that these existed until you suggested them, many thanks. I found the equivalent in PixInsight, CanonBandingReduction. Up to now my camera did not exhibit this issue so this is all new to me. Thanks for the link about Siril's banding.

As and aside a member of our local astronomy group has offered to give some training on the use of Siril, I've signed up for this. 

 

Many thanks

Andy

Edited March 8 by Andy56

[happy-kat]

Just as an off shot have you tried a different cable from the camera in case it's noise interference causing your banding

[symmetal]

I just found out that a new header keyword of ROWORDER was added to the FITS specification about 4 years ago to help resolve this issue, but unless software is rewritten to look for this keyword it will of course be ignored. Here's a couple of topics about it.

https://pixinsight.com/forum/index.php?threads/support-for-fits-tag-roworder.14906/

https://www.indilib.org/forum/ccds-dslrs/7695-introducing-a-new-roworder-keyword-in-fits-headers.html

Alan

[Andy56]

I'm currently using it straight to the sd card so I don't have to use a laptop or similar so no transfer cables. But having said that reminds me that I don't think I've used it like this before. So I will try a different SD card because many people have suggested it although I could never see the logic of it.

Andy

[Andy56]

Symmetal: ah maybe. I'll have a read in the morning

Thanks

Andy