Horizontal black lines on stacked picture

[Stefan73]

Hi all,

I've just started guiding with PHD2 and there's appeared two black horizontal lines on the resulting stacked image (see 16_09_2022.fit).  They're most easily visible in blue with maximum stretching (like in the screenshot of siril I've attached).  Does anyone know why?  It doesn't look like the pictures of "walking noise" I've seen and before guiding I didn't seen anything like this.

The stack was created from 120 1 minute subs using a Canon 550D, 130PDS, OAG for guiding and no dithering.  I think the individual subs have a very faint black line there too but hard to be sure.

The 17_09_2022.fit was the same set up (although 90 subs I think) but I tried dithering 5 pixels every 4 frames (in a spiral) which has improved it a bit and the black lines look more smeared so it looks like more dithering is the answer but I've no idea what values I should be using and the ones I did use were just plucked out of the air.  What do people use?  Or am I barking up the wrong tree here and something else is amiss?

Any insight appreciated!

16_09_2022.fit 17_09_2022.fit

[ONIKKINEN]

Looks like typical Canon banding, but just a lot more aggressive than i ever saw with my 550D. I suppose these cameras have individual differences in their sensors so maybe not that strange.

But yes, dithering more will smear that out and make it essentially disappear in the end. You want to dither at least 20 pixels, in your case maybe 30 or more since the pattern is so bold. Try to dither at minimum once per 5 min.

[alacant]

3 hours ago, Stefan73 said:

insight appreciated

Hi

This is typical of the old d4 firmware. Make sure you're using the latest version.

Capture: Dither at least 10 pixels after every frame

Our method using Siril:

1. 550d so lose any dark frames you may have used.

2. Pre-process with a fixed bias offset of 2048

3. After pre-processing, apply horizontal banding reduction with highlight protection to the whole sequence before registration..

4. Stack using a clipping algorithm. Sigma 3,3 is a good starting point.

-> super clean image ready for processing

Cheers and HTH

Edited September 21, 2022 by alacant
ortografía spell

[Stefan73]

40 minutes ago, alacant said:

This is typical of the old d4 firmware. Make sure you're using the latest version.

Ok that's really interesting.  I'd downgraded ages ago to 1.09 to use magic lantern (which I don't need any more due the wonders of EKOS and INDI).  Time to upgrade again I think.

No dark frames and already using sigma 3,3 so that's OK.

The fixed bias offset is really useful as well, I was going to investigate what value to use there at some future point.

I'll see if the banding reduction can improve what I've got too but other than that it's really useful to have an idea of the minimum dither I should be doing for the future.  Thanks.