Author Topic: Diamond shaped stars  (Read 2798 times)

Offline Les Tilly

  • Newcomer
  • Posts: 4
    • View Profile
Diamond shaped stars
« on: 2013 February 18 05:33:03 »
I have just started to use pixinsight I used the Batcg reprocessing Script and i get Diamond shaped stars ant suggestions to making them round

Les

Offline oldwexi

  • PixInsight Guru
  • ****
  • Posts: 606
    • View Profile
    • Astronomy Pages G.W.
Re: Diamond shaped stars
« Reply #1 on: 2013 February 18 07:10:15 »
Hi Les!
What optics have you used?
It looks like spikes introduced by obstructions
infront of the optics or in the optical path.
Gerald

Offline Les Tilly

  • Newcomer
  • Posts: 4
    • View Profile
Re: Diamond shaped stars
« Reply #2 on: 2013 February 18 09:57:52 »
I use a Tak 106 newQ and FLI 16803 I have perfectly round stars when i do it in other programs

Thanks for replying
Les

Offline Josh Lake

  • PixInsight Old Hand
  • ****
  • Posts: 424
    • View Profile
Re: Diamond shaped stars
« Reply #3 on: 2013 February 18 10:52:34 »
Can you post an example frame from PI and from another program to compare? I wonder if it's simply a case of PI's Screen Transfer Function stretching the data more aggressively than other programs and so you're seeing a structure that's in all of them. Have you opened the circular-star frames in PI to verify?

Offline Geoff

  • PixInsight Padawan
  • ****
  • Posts: 863
    • View Profile
Re: Diamond shaped stars
« Reply #4 on: 2013 February 18 14:42:46 »
You could try changing the registration algorithm in BPP.  I find that "Auto" (the default) often gives me strange artifacts.
« Last Edit: 2013 February 18 20:24:46 by Geoff »
Don't panic! (Douglas Adams)
Astrobin page at http://www.astrobin.com/users/Geoff/
Webpage (under construction) http://geoffsastro.smugmug.com/

Offline Alejandro Tombolini

  • PTeam Member
  • PixInsight Jedi
  • *****
  • Posts: 1260
    • View Profile
    • Próxima Sur
Re: Diamond shaped stars
« Reply #5 on: 2013 February 18 19:46:41 »
Hi Les, maybe it is not related to PI. This picture was taken with the same telescope and a Canon 5D.
The following screenshot has four image of the same star of 5, 15, 60 and 300 second, directly from raw, without process (only debayered), and besides the bayered ones.
It can be seen that when exposure time increase, the diamond appears.
Saludos, Alejandro.

Offline Juan Conejero

  • PTeam Member
  • PixInsight Jedi Grand Master
  • ********
  • Posts: 6650
    • View Profile
    • http://pixinsight.com/
Re: Diamond shaped stars
« Reply #6 on: 2013 February 19 00:56:09 »
As always happens in these cases, we really need to work with a sample of the data where the problem can be reproduced.  Otherwise all we can do is speculate.

I seriously doubt this can be caused by any of the interpolation algorithms in StarAlignment. All of these algorithms are isotropic by nature. They can cause some ringing, which can be controlled with the clamping parameter and, in extremely difficult cases, by choosing an inferior algorithm such as bicubic spline or even bilinear. But ringing cannot deform star shapes this way. Definitely, I need to see a sample of the original data.
Juan Conejero
PixInsight Development Team
http://pixinsight.com/

ruediger

  • Guest
Re: Diamond shaped stars
« Reply #7 on: 2013 February 19 02:13:20 »
Could it be related to spikes introduced by the micro lenses of the sensor?
At least the Kodak 8300 is known to produce similar looking star shapes.

But this doesn't explain why you get round stars in other programs.

Rüdiger

Offline vi100

  • Newcomer
  • Posts: 22
    • View Profile
Re: Diamond shaped stars
« Reply #8 on: 2013 February 20 05:47:41 »
Could it be related to spikes introduced by the micro lenses of the sensor?
At least the Kodak 8300 is known to produce similar looking star shapes.

But this doesn't explain why you get round stars in other programs.

Rüdiger

+1

Offline Les Tilly

  • Newcomer
  • Posts: 4
    • View Profile
Re: Diamond shaped stars
« Reply #9 on: 2013 March 10 06:50:16 »
Sorry for not getting back sooner. I went over a few more images and found it also in unrelated Pi images  Could this be a Under sampling Issue that Basically can only be seen by zooming in on the image?

Les

Offline lucchett

  • PixInsight Old Hand
  • ****
  • Posts: 449
    • View Profile
Re: Diamond shaped stars
« Reply #10 on: 2013 April 07 14:08:01 »
Hi,
i am seeing this only now..
I have the same issue, same chip.
The defect is either induced or revealed by the image registration algorithm.
All lanczos show the issue. Low clamping can mitigate the problem.
Changing the method does eliminates the issue, but with a loss in resolution.

Stars are slightly diamond shaped, they show something like subtle spikes.

Andrea