Stacking rotated subs !

[halli]

Hi

I've taken two nights of Ha data for the image below but noticed that  the diffraction spikes on the stars are not coincident ie there are two sets of 4 spikes slightly rotated wrt each other see image below .  The image below comprises 2 nights of stacked data.

 

When I examined the stacked image from the data set from each night I found that there was a rotational difference between them - the diffraction spikes for each night looked ok and the focus was also OK.     I guess DSS therefore rotated one to get the alignment during the stacking process of the complete 2 night set  - only trouble is the stars now have weird diffraction spikes which must have occurred because of the introduced rotation resulting in the two sets of rotated diffraction spikes.  

i would appreciate if anyone could let me know  a way around this under these circumstances.

Thanks for any help

[Zakalwe]

DSS will use the star centroids to stack. You could try playing around with different stacking algorithms in DSS to get a different result- a kappa-Sigma *might" reduce the diffraction spikes (the same way that it takes out satellite trails).

 

Failing that, import the final image into PS. Then create another stacked image just using one night's data. Layer that on top of the first image and use a layer mask to mask everything out except the stars and diffraction spikes. That'd give you an image with both night's data for the nebulosity and just the stars and spikes from one night.

[halli]

Thanks for your comments Zakalwe.

You have confirmed my thoughts as to the reason for the problem.  Unfortunately I do not have PS only PI.  I think there are similar techniques in Pi that can be used I will need to research this.

 I might also try to do the stacking in PI to see if there is any difference.  

Cheers

Ian

[ollypenrice]

It's an odd effect to have, though. Did the main tube get rotated in the rings, perhaps, between nights? Rotating the camera won't do it because, unless the tube is rotated, the relationship between the spider and the sky will be unchanged.

Olly

[Zakalwe]

7 minutes ago, ollypenrice said:

It's an odd effect to have, though. Did the main tube get rotated in the rings, perhaps, between nights? Rotating the camera won't do it because, unless the tube is rotated, the relationship between the spider and the sky will be unchanged.

Olly

Are you sure Olly? I would have thought that rotating the camera will change where the diffraction spikes will fall on the sensor. The stacking process will register the stars, but the spike will rotate as the image is registered and cropped?

[ollypenrice]

Just now, Zakalwe said:

Are you sure Olly? I would have thought that rotating the camera will change where the diffraction spikes will fall on the sensor. The stacking process will register the stars, but the spike will rotate as the image is registered and cropped?

I'm never sure about anything related to spatial awareness! But here's my train of thought: say your spider is precisely aligned along RA and Dec (so one set of vanes is parallel with the counterwieght bar.) This will create spikes which run E-W and N-S on the image (irrespective of how the frame of the image is orientated.) If you don't rotate the tube but only the camera, then you'll rotate the frame of the picture but the spider will still trace Ra and Dec.

Olly

[Zakalwe]

I've got a headache now!

[The Admiral]

17 minutes ago, Zakalwe said:

I've got a headache now!

It's the same as your Eggy Stars problem Zakalwe. When you rotated the camera, the eggy-ness stayed the same in relation to the stars. It would be the same with spikes. Rotating the spider in relation to the stars would move the direction of the spikes relative to the stars.

Ian

[Zakalwe]

23 minutes ago, The Admiral said:

It's the same as your Eggy Stars problem Zakalwe. When you rotated the camera, the eggy-ness stayed the same in relation to the stars. It would be the same with spikes. Rotating the spider in relation to the stars would move the direction of the spikes relative to the stars.

Ian

That makes sense....I am a bit thick when it comes to this!

[The Admiral]

18 minutes ago, Zakalwe said:

That makes sense....I am a bit thick when it comes to this!

Not at all, join the club! When I first looked at your eggy stars I thought that the elongation had rotated, until I realised that the stars had moved as well

Ian

[ollypenrice]

3 minutes ago, The Admiral said:

Not at all, join the club! When I first looked at your eggy stars I thought that the elongation had rotated, until I realised that the stars had moved as well

Ian

I always struggle with this and need a pen and paper to sort it out. Regarding Zakalwe's stars I was wondering what we might conclude from a test in which he rotated the whole tube in its rings but, since the rotator is forward of the focuser on his scope, I concluded that it would have no effect. I might be wrong about that, though!

Olly

[halli]

Well guys - I will put you out of your misery - I now remember that I believe I did rotate the tube between nights !

I spotted that the camera and filter wheel were offsett to where they should have been for the best DEC balance and hence I rotated the tube a bit.  It never occurred to me at the time  or subsequently that it would have affected the image.  

It all now makes sense as the angle between the spikes looks about the same as the amount I rotated the tube.  I was thinking about how I might have  rotated the camera accidentally and couldn't understand it as I was sure I didn't. 

Also I have stacked plenty of subs previously with slightly different camera/drawtube  rotational positions which had never affected the diffraction spikes only the image framing.  I also thought it may have been a focus issue but looking at one nights worth of data I was sure it wasn't.   What a weird affect and something not to repeat in the future. I guess this is living proof of the impact of tube rotation.  It was when I was looking at the subs between nights that I picked the rotation up because the relative star positions were different between the nights which could only be explained due to image rotation so it wasn't just an affect on the diffraction spikes due to the spider orientation !

Looks like camera rotation wrt drawtube  affects frame orientation only after stacking ie the angular displacement of the image rectangle changes wrt the image centre  and tube rotation is pure image rotation with the consequential impact but the frame rectangle orientation remains the same.   Wow - how interesting !

Anyhow now we have experimental proof of the hypotheses you guys had proposed if anyone can help me remove the darn affect in PI it would be gratefully appreciated.............

Thanks for your input

[Astrofriend]

Maybe you can find some clue from this discussing about PI and diffraction spikes:

https://pixinsight.com/forum/index.php?topic=4002.0

 

/Lars