Weird Bandings in ZWO 1600

[ReZ]

hey folks
I have an OIII image and its integration time is 20X10min. As you can see I have some weird banding noise in there, it only appears in OIII and Blue filters, and it only appears once in a while. i have not been able to find a solution. maybe you guys can help me out?
thanks

P.S. It is not vertical nor Horizontal but diagonal

[IanL]

Are you dithering between frames? This looks like a classic case of pattern build up due to constant drift between frames. Check a number of frames before alignment and stacking and I suspect you will see the stars drifting in the same top-left bottom-right direction as the pattern.

It is a particular problem for CMOS cameras as it is next to impossible to calibrate out pattern noise. I notice the same issue in Blue and OIII probably due to the usually lower SNR compared to other filters.

Dithering doesn't eliminate the noise but by randomising the pattern, it no longer appears objectionable to the human eye.

[ReZ]

4 hours ago, IanL said:

Are you dithering between frames? This looks like a classic case of pattern build up due to constant drift between frames. Check a number of frames before alignment and stacking and I suspect you will see the stars drifting in the same top-left bottom-right direction as the pattern.

It is a particular problem for CMOS cameras as it is next to impossible to calibrate out pattern noise. I notice the same issue in Blue and OIII probably due to the usually lower SNR compared to other filters.

Dithering doesn't eliminate the noise but by randomising the pattern, it no longer appears objectionable to the human eye.

hi
i just checked the frames and it drifts in the same direction. I do belive you are right it is a dithering problem. i currently am using the small settings. will Spiral do it?
thanks

 

[tooth_dr]

Walking noise. I use random dither but I am lead to believe spiral works ok. For a DSLR CMOS I use 12-20 pix dither. 

[IanL]

My experience is that 15 -20 pixel random dither works well. Can't comment on spiral as never tried it.

[sharkmelley]

It's walking noise and walking noise is usually generated by bright pixels "walking" across the image.  But these are dark pixels "walking".  So what is causing those dark pixels in the first place?  Are you using darks and is it possible that the master dark is badly matched to the data?

Mark

 

[Stub Mandrel]

I'm not sure I buy 'walking noise' as if you are guiding without dither how can bright (or dark) pixels move across the image?

I suspect that it may have multiple causes and that one of them is high haze/thin cloud.

[wimvb]

On 25/08/2018 at 16:09, ReZ said:

will Spiral do it?

Yes.

Other solution include:

Cosmetic correction during image calibration.

Adjusting the limit for sigma rejection during image integration.

[tooth_dr]

2 hours ago, Stub Mandrel said:

I'm not sure I buy 'walking noise' as if you are guiding without dither how can bright (or dark) pixels move across the image?

I suspect that it may have multiple causes and that one of them is high haze/thin cloud.

I don’t know the cause. But. When I dither I don’t get those noise patterns. When I don’t dither I do get them. Same kit and exposure, many different nights or different targets. Easier just to dither and eliminate them. 

[sharkmelley]

2 hours ago, Stub Mandrel said:

I'm not sure I buy 'walking noise' as if you are guiding without dither how can bright (or dark) pixels move across the image?

Inaccurate polar alignment?  Flexure between main scope and guide scope?  Focus tube droop with gravity?  Mirror shift (on SCT scopes). Plenty of potential causes.

In any case, the original poster earlier confirmed that the star field was indeed drifting from sub to sub (see post #3 in this thread).

Mark

[wimvb]

4 hours ago, Stub Mandrel said:

I'm not sure I buy 'walking noise' as if you are guiding without dither how can bright (or dark) pixels move across the image?

I suspect that it may have multiple causes and that one of them is high haze/thin cloud.

 

1 hour ago, sharkmelley said:

Inaccurate polar alignment?  Flexure between main scope and guide scope?  Focus tube droop with gravity?  Mirror shift (on SCT scopes). Plenty of potential causes.

In any case, the original poster earlier confirmed that the star field was indeed drifting from sub to sub (see post #3 in this thread).

Mark

I agree that it seems strange that there is drift if you guide, because the shift occurs between frames, and never within a frame. Whenever I've had this, stars were (reasonably) round in each frame, yet there was a substantial shift between frames; as if the mount jumps between frames. A continuous movement should result in star trails. I always suspected the mirror of my dslr: it opens quite violently and could cause the mirror of my refkector to shift ever so slightly. Within an exposure, the mirror doesn't move due to "stiction". This is the best explanation I could come up with, but it doesn't explain in case of a mirrorless camera.

As the OP noted, the pattern dissapears if you dither.

Here's an example where careful calibration removed the streaks:

http://pixinsight.com.ar/en/info/processing-examples/19/crux-southern-cross.html

[IanL]

Part of the cause could be field rotation due to inaccurate polar alignment. It is hard to be sure, but looking at the background streaks there does appear to be a slight curvature with the centre of rotation located somewhere off to the bottom left of the image. The diagnostic test for this would be to stack the subframes without star aligning them first. If the resulting star trails are curved then I'd think polar alignment would be part of the cause. Flexure between guidescope and imaging scope or mirror flop (if SCT) are also likely to be part of the issue (no law that says you can't have more than one problem at the same time!)

15 hours ago, wimvb said:

I agree that it seems strange that there is drift if you guide, because the shift occurs between frames, and never within a frame. Whenever I've had this, stars were (reasonably) round in each frame, yet there was a substantial shift between frames; as if the mount jumps between frames. A continuous movement should result in star trails.

In this case if you take a close look at the stars in the image, you can see they are distinctly egg shaped, again in the direction of the streaks/drift. So this is a continuous drift, not a sudden jump from frame to frame. So in each sub, the stars are slightly trailed, but the alignment process obviously stops them from becoming trails. On longer subs they would form more obvious trails, which suggests some flexure or flop issue at play, since the guider would otherwise prevent trailing like this.  Camera mirror shake usually creates more 'dumbell' or erratic star shapes depending on the orientation of the main vibrational movement.

[ReZ]

Hey guys thanks for all your useful replies.

I managed to extract the pattern rejection map with PI. I believe the source is tracking related (dithering in one axis or a little image shift) and I have eliminated it by using spiral dithering.