×

INDI Library v2.0.6 is Released (02 Feb 2024)

Bi-monthly release with minor bug fixes and improvements

Fixed pattern noise due to dithering

  • Posts: 984
  • Thank you received: 160
Hi, I've been taking a few pictures lately and one thing that bothers me is a pattern (diagonal bands) that does show up in my final results. IMO it is due to sensor pixels that are more or less sensitive. When INDI dithers a certain amount of pixels and does so always in the same direction, something like this will show in your pictures (see attachment). Is my assumption correct that INDI will dither in one direction all the time? If so, dithering in a random direction should help avoid such patterns. As an alternative, dithering in a spiral pattern could also improve results. In case something like this has already been implemented, please let me know where to find it. Thanks a lot!
Last edit: 5 years 10 months ago by Alfred.
5 years 10 months ago #25919
Attachments:

Please Log in or Create an account to join the conversation.

  • Posts: 39
  • Thank you received: 11
Hi Alfred,

"...Ekos will move the guide star in a random direction for the amount of pixels specified in the dither input box..."
Seems like it didn't dither at all. The pattern looks like drift due to imperfect polaralignment. I seen this in undithered captures I did myself, too.

Andreas
The following user(s) said Thank You: Alfred
5 years 10 months ago #25920

Please Log in or Create an account to join the conversation.

  • Posts: 984
  • Thank you received: 160
OK, when Ekos uses random directions it must be misalignment. When I think about it... I did align pretty well but at the end of the night I noticed I had fogotten to tighten the azimuth screws. So it might have moved during the night. Bummer.

Thanks for your help!
5 years 10 months ago #25923

Please Log in or Create an account to join the conversation.

  • Posts: 39
  • Thank you received: 11
You should not see this misalignment (showing in noiserain) at all, if dithering went well. Maybe dithering was just not wide enough. How much pixel did you dither and what is the ratio between guide scope focal length and main scope focal length?
The following user(s) said Thank You: Alfred, Alan Mason
5 years 10 months ago #25925

Please Log in or Create an account to join the conversation.

  • Posts: 984
  • Thank you received: 160
Main scope is 1200mm and guide scope is 1000mm. I never got to use the internal guider with my Alt-5. But this time I used a G-11 and internal guiding did perform very well. It reported dithering was performed successfully. I used 2,8 pixels every 1 frame.
5 years 10 months ago #25926

Please Log in or Create an account to join the conversation.

  • Posts: 39
  • Thank you received: 11
This is the cause of your "undithered" looking capture. Asuming pixelsize is nearly identical on both cams you dithered your main cam 2.8*1000/1200=2.33 pixel.
This is near nothing. Hot pixel or "ununiform behaving" ones are neighboring themselfs in the stack, this makes the rain fall.
Dither at least 10-20Pix with this configuration to separate them for sure.
Andreas

edit its 2.8*1200/1000=3.3 but thats not enough either
The following user(s) said Thank You: Alfred, Tarun, Alan Mason
Last edit: 5 years 10 months ago by Andreas.
5 years 10 months ago #25927

Please Log in or Create an account to join the conversation.

  • Posts: 984
  • Thank you received: 160
Ok, will do that and report back. Thanks an awful lot!!!
5 years 10 months ago #25928

Please Log in or Create an account to join the conversation.

  • Posts: 984
  • Thank you received: 160
I promised to report back... A few days ago I took another photo of M51. This time dithering was set to 9,7 pixels (10 pixels is the max in Ekos). The result is marvellous, noise is perfectly random IMO. Problem solved! Thanks a lot for your help, andreas.d!

P.S.: I had to abort the session after taking 13x3min exposures due to bad weather.
5 years 10 months ago #26059
Attachments:

Please Log in or Create an account to join the conversation.

Time to create page: 1.147 seconds