Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

Binning Ramifications

Astrophotography
  • Please log in to reply
9 replies to this topic

#1 GlendaleGuy

GlendaleGuy

    Vostok 1

  • -----
  • topic starter
  • Posts: 150
  • Joined: 26 Sep 2017
  • Loc: Phoenix, Arizona

Posted 16 April 2021 - 01:07 PM

    On my next run I'm going to bin my camera 2 x 2 or 3 x 3 to get the image scale in line for 1" to 2" per pixel. Does this mean I need to do the flats, darks, and bias with the same bin setting? I already have dark and bias libraries for 1 x 1.



#2 WadeH237

WadeH237

    Fly Me to the Moon

  • *****
  • Posts: 7,053
  • Joined: 24 Feb 2007
  • Loc: Ellensburg, WA

Posted 16 April 2021 - 03:30 PM

Does this mean I need to do the flats, darks, and bias with the same bin setting? I already have dark and bias libraries for 1 x 1.

Yes, you need to take a set of calibration frames that match the settings that you use for imaging - especially biases and darks.

 

In theory, you could doctor the master flats to work with 2x2 instead of 1x1, but you would need to know exactly what you are doing.  It is far better just to retake the flats as well.



#3 const

const

    Vostok 1

  • -----
  • Posts: 140
  • Joined: 11 May 2020
  • Loc: WA

Posted 16 April 2021 - 09:59 PM

What is your camera and why do you want to bin? Might be easier to continue capturing at native resolution and downsample some time after calibration.


  • ks__observer likes this

#4 GlendaleGuy

GlendaleGuy

    Vostok 1

  • -----
  • topic starter
  • Posts: 150
  • Joined: 26 Sep 2017
  • Loc: Phoenix, Arizona

Posted 17 April 2021 - 04:13 AM

What is your camera and why do you want to bin? Might be easier to continue capturing at native resolution and downsample some time after calibration.

I'm running a ZWO 071 on the C11 Edge HD. I'm constantly having to try and reduce star size in processing.  When running with no reducer the scale is .35 arc-seconds per pixel. To get to the 1 to 2 arc-second range I will need to bin around 3 x 3 to 4 x 4.



#5 Phishin_phool

Phishin_phool

    Messenger

  • *****
  • Posts: 478
  • Joined: 31 May 2020
  • Loc: Ohio

Posted 17 April 2021 - 07:37 AM

Forgive me if this is incorrect as i know somebody will chime in but ZWO 071 rez is 4944x3284 (16.2 mp)- binning 3x3 will reduce the resolution by 9 to 1648x1094 (1.8mp) and 4x4 will reduce it to 1236x821 (1.01 mp) - that is some serious image quality reduction.



#6 const

const

    Vostok 1

  • -----
  • Posts: 140
  • Joined: 11 May 2020
  • Loc: WA

Posted 17 April 2021 - 04:02 PM

I'm running a ZWO 071 on the C11 Edge HD. I'm constantly having to try and reduce star size in processing.  When running with no reducer the scale is .35 arc-seconds per pixel. To get to the 1 to 2 arc-second range I will need to bin around 3 x 3 to 4 x 4.

So, it is a CMOS camera, what I said applies. Where did you come from to 1-2"/pix? From my experience, processing goes better when stars FWHM is in 2-4 pixels range. In other words, to justify 3×3 binning I would need to see FWHM over 10 pixels. Is that your case? Whatever the good binning level is for you, you can apply it after calibration and don't worry about building dark/bias libraries for multiple levels.



#7 GlendaleGuy

GlendaleGuy

    Vostok 1

  • -----
  • topic starter
  • Posts: 150
  • Joined: 26 Sep 2017
  • Loc: Phoenix, Arizona

Posted 17 April 2021 - 05:47 PM

So, it is a CMOS camera, what I said applies. Where did you come from to 1-2"/pix? From my experience, processing goes better when stars FWHM is in 2-4 pixels range. In other words, to justify 3×3 binning I would need to see FWHM over 10 pixels. Is that your case? Whatever the good binning level is for you, you can apply it after calibration and don't worry about building dark/bias libraries for multiple levels.

    Thanks for your comments. When I'm running at prime focus the scale is .35 arc-seconds per pixel and my FWHM are typically in the 5 to 6 range. When I use the .7 reducer my scale improves to 0.5 arc-seconds per pixel and FWHM is around 4 to 5 I think without going back to look. I forgo the reducer sometimes due to the image size of the object. I should also mention my mount seems to fluctuate between 0.8 and 1.2 arc-seconds of corrections and my seeing on a good night is usually 0.7 arc-seconds, but often over 1.0. I'm on the list for a much better mount for that former part of the equation.

 

    From my readings on the subject it seems the best practice is to get a scale of 1 to 2 arc-seconds per pixel or to shoot for the number of your seeing and mount accuracy. Would that be correct? Here is one Astrophotographer's experiments on this and I've seen a few discussions here on Cloudy Nights saying the same thing and referring to the Nyguist theory of shooting for 1/3 of your seeing.

 

https://www.swagastr... microns square.

 

  That goes a little against the hard and fast rule I've seen quoted of trying to get your image scale between 1 and 2. I'm typing out loud here as I'm trying to figure it out. I'm really interested to hear your take on this. And thanks once again.

 

    When I tried two hours on M106 last night I was getting 1.8 to 2.5 FWHM per N.I.N.A. which should have been while using a calculated image scale of 1.06 arc-seconds per pixel. Also the drift in FWHM was in one direction over time, so probably due to temperature changing the focus. Maybe I should split the difference and try 2 x 2 binning?
 



#8 GlendaleGuy

GlendaleGuy

    Vostok 1

  • -----
  • topic starter
  • Posts: 150
  • Joined: 26 Sep 2017
  • Loc: Phoenix, Arizona

Posted 17 April 2021 - 05:49 PM

Forgive me if this is incorrect as i know somebody will chime in but ZWO 071 rez is 4944x3284 (16.2 mp)- binning 3x3 will reduce the resolution by 9 to 1648x1094 (1.8mp) and 4x4 will reduce it to 1236x821 (1.01 mp) - that is some serious image quality reduction.

I agree with you on first glance it looks like you are losing resolution. But my understanding on this is that you don't really have the resolving power to get to that resolution anyway unless you are matching your image scale correctly. But that's why I'm posting here trying to get it right in my mind.



#9 const

const

    Vostok 1

  • -----
  • Posts: 140
  • Joined: 11 May 2020
  • Loc: WA

Posted 17 April 2021 - 11:18 PM

When I'm running at prime focus the scale is .35 arc-seconds per pixel and my FWHM are typically in the 5 to 6 range. When I use the .7 reducer my scale improves to 0.5 arc-seconds per pixel and FWHM is around 4 to 5 I think without going back to look. I forgo the reducer sometimes due to the image size of the object. I should also mention my mount seems to fluctuate between 0.8 and 1.2 arc-seconds of corrections and my seeing on a good night is usually 0.7 arc-seconds, but often over 1.0.

Are you sure your seeing is 1"? With that I would keep all the pixels intact!

With all the other numbers, I would not look for 1" pixel scale, it is too low. Nyquist should be used carefully, and it sets the lower limit of resolution that 'works'. There is never too much pixels. The lower limit is 1/2 of what you want to be able to resolve. If your skies plus optics and mount give you 2" (a sensible number I would say) then 1" pixels is the biggest you should ever use. There is no harm in smaller pixels.

I often image under crappy conditions and get corresponding quality of data. I try to keep native resolution until after deconvolution step. In cases like FWHM=10 and low contrast because of Moon, registration fails to recognize large blurry spots as stars :) Then I downsample calibrated subs and try again.

When conditions are better, I keep original resolution (0.47"/pix with my 8" SCT) to the end. So, if you bin even 2× during capture, to me it really sounds like killing data that may very well be useful.



#10 GlendaleGuy

GlendaleGuy

    Vostok 1

  • -----
  • topic starter
  • Posts: 150
  • Joined: 26 Sep 2017
  • Loc: Phoenix, Arizona

Posted 18 April 2021 - 02:34 AM

Are you sure your seeing is 1"? With that I would keep all the pixels intact!

With all the other numbers, I would not look for 1" pixel scale, it is too low. Nyquist should be used carefully, and it sets the lower limit of resolution that 'works'. There is never too much pixels. The lower limit is 1/2 of what you want to be able to resolve. If your skies plus optics and mount give you 2" (a sensible number I would say) then 1" pixels is the biggest you should ever use. There is no harm in smaller pixels.

I often image under crappy conditions and get corresponding quality of data. I try to keep native resolution until after deconvolution step. In cases like FWHM=10 and low contrast because of Moon, registration fails to recognize large blurry spots as stars smile.gif Then I downsample calibrated subs and try again.

When conditions are better, I keep original resolution (0.47"/pix with my 8" SCT) to the end. So, if you bin even 2× during capture, to me it really sounds like killing data that may very well be useful.

I usually check Meteoblue for seeing before I set up. This is tonight's estimate. Though it's crappy and cloudy out it says about .65 arc-second seeing, unless I've been reading this wrong all along.

 

It was my understanding that if you use scaling of under 1 arc-second per pixel that it will bloat stars, which I definitely get. And that going to the 1 arc-second will get you more light on the binned pixels. And that guiding and focus cause less problems. I've been using 1 x 1 binning up until the other night, but am trying to figure out if there is a better way than I've been using.

Attached Thumbnails

  • MeteoBlue Seeing.jpg



CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics





Also tagged with one or more of these keywords: Astrophotography



Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics