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

Checking SNR prior to calibration

  • Please log in to reply
9 replies to this topic

#1 SeymoreStars

SeymoreStars

    Skylab

  • *****
  • topic starter
  • Posts: 4,374
  • Joined: 08 May 2014
  • Loc: Pennsyltucky

Posted 11 March 2024 - 08:43 PM

I have some exoplanet work which requires evaluating the SNR of a particular star during the image capture process.

 

Are there other tools which do this?

 

Thanks

Steve



#2 robin_astro

robin_astro

    Mercury-Atlas

  • *****
  • Posts: 2,567
  • Joined: 18 Dec 2005
  • Loc: UK

Posted 12 March 2024 - 07:28 AM

If you can measure the integrated counts in an aperture then sqrt (counts * gain in e-/ADU) will give a good estimate on bright stars


  • SeymoreStars likes this

#3 SeymoreStars

SeymoreStars

    Skylab

  • *****
  • topic starter
  • Posts: 4,374
  • Joined: 08 May 2014
  • Loc: Pennsyltucky

Posted 12 March 2024 - 11:27 AM

AIJ is what I used. Grabbed a single image, plate solved with ASTAP, set the apertures and let AIJ choose Comps. This produced the Measurements table, where the SNRs were found.

 

Thanks

Steve



#4 StupendousMan

StupendousMan

    Mariner 2

  • -----
  • Posts: 295
  • Joined: 21 Aug 2005

Posted 17 March 2024 - 07:01 PM

If you can do a bit of extra processing, it might be a good idea to compute the standard deviation from the mean of the difference in magnitude between two stars in several images.  Pick two stars, the target A and some other star of similar brightness B.  Use aperture photometry to calculate an instrumental magnitude for each star in each image.  Then compute the difference

 

            d1   =   (instr mag A) - (instr mag B)    in image 1

            d2   =   (instr mag A) - (instr mag B)    in image 2

    ...

            dN  =    (instr mag A) - (instr mag B)    in image N

 

The scatter in the values of these differential magnitudes give you a pretty good idea of the uncertainty in the measurements of the stars; yes, it should be a bit bigger, but I like conservative estimates.

 

This method is a robust one that includes source of uncertainty not included in the simpler and quicker method mentioned earlier.



#5 robin_astro

robin_astro

    Mercury-Atlas

  • *****
  • Posts: 2,567
  • Joined: 18 Dec 2005
  • Loc: UK

Posted 18 March 2024 - 06:58 AM

 

 

This method is a robust one that includes source of uncertainty not included in the simpler and quicker method mentioned earlier.

Agree 100%   SNR is only one factor in the overall uncertainty and often not even the main one, particularly when you get into the systematic effects

 

Cheers

Robin



#6 Ed Wiley

Ed Wiley

    Vanguard

  • *****
  • Posts: 2,460
  • Joined: 18 May 2005
  • Loc: Texas, USA

Posted 18 March 2024 - 10:16 AM

The fact that SNR is only one factor in determining a reasonable uncertainty is one of the reasons Ken Menzies and I advocated for ensemble photometry (JAAVSO, 2022, v50). In ensemble the uncertainty is the standard deviation of the magnitude estimates of the comp stars estimates of the target star magnitude reported by VPhot. You really need at least three comp stars to obtain a "valid" standard deviation.

 

This does not work for time series analysis where the target is varying continuously over a short time span and the overall STD is naturally high. Another check is the difference between the reported magnitude of the check star and its standard magnitude. This is the O-K protocol. If the O-K is within the errors of the check star (a secondary standard like the comp stars), then you are doing OK. Using several check stars would duplicate StupendousMan's method, but you do need 3+ stars to obtain a "valid" standard deviation and those stars need to be secondary standards. This would work for time series analyses.

 

Precision is another matter. For slowly changing targets like Miras the precision can be estimated with a short time series (say N=4). The standard deviation of the mean magnitude of the time series yields precision. Of course, this is not possible for time series analysis where precision would be some sort of average rms of the fit to the light curve model.

 

In some cases (e.g. exoplanets transit timing) your main objective is to maximize precision. In other cases you need to maximize both precision and accuracy (e.g., building physical models of stars or systems or providing data for that activity). 

 

Another problem is consistency among observers. Transforming data to the standard system is a major boost in promoting consistency. Of course, this requires at least two photometric filters.

 

Ed


  • Xilman likes this

#7 SeymoreStars

SeymoreStars

    Skylab

  • *****
  • topic starter
  • Posts: 4,374
  • Joined: 08 May 2014
  • Loc: Pennsyltucky

Posted 18 March 2024 - 10:35 AM

I am simply trying to determine if my exposure time is adequate to satisfy Exoplanet capture while at the telescope.

 

Steve



#8 StupendousMan

StupendousMan

    Mariner 2

  • -----
  • Posts: 295
  • Joined: 21 Aug 2005

Posted 20 March 2024 - 06:19 PM

Suppose the transit depth is 0.020 mag. Take 5 images, and perform aperture photometry on two stars (A, B) in the field of brightness comparable to the target.  Compute the mean and stdev of differences in the instrumental magnitudes of the pair: mag(A)-mag(B).   If the stdev is of order 0.020 mag, you have a chance to detect the transit (especially if you average a number of consecutive measurements).  If the stdev is much smaller, you're in good shape.  If the stdev is much larger, you should probably pick a different target.



#9 mrm6656

mrm6656

    Lift Off

  • -----
  • Posts: 23
  • Joined: 18 Mar 2019
  • Loc: Boulder County, Colorado USA

Posted 21 March 2024 - 09:36 AM

The simplest thing to do may be to use ASTAP. If you use the ASTAP UI to plate solve (rather than using the command line), then you're almost home. Once an image has been solved, you can mouse-over any of the image's stars, and its HFD, FWHM and S/N will be shown in the data listed at the bottom of the UI. The values listed are approximate, but my experience suggests that those values are probably plenty accurate for your current needs. ASTAP can also show magnitude estimates for the stars in solved images via its menu: Tools/Magnitudes(measured)annotation. The magnitudes are derived from the database that ASTAP uses to plate solve

 

      --- MIke


  • Ed Wiley and SeymoreStars like this

#10 GaryShaw

GaryShaw

    Gemini

  • *****
  • Posts: 3,090
  • Joined: 28 Nov 2017
  • Loc: Massachusetts / Wyoming

Posted 29 April 2024 - 05:15 PM

AIJ is what I used. Grabbed a single image, plate solved with ASTAP, set the apertures and let AIJ choose Comps. This produced the Measurements table, where the SNRs were found.

 

Thanks

Steve

Steve

… or you can just hover on the star to be measured and click it using the scroll wheel. That will produce a single line measurement table for that star …. Including peak pixel value, the SNR  of the star and its FWHM.


Edited by GaryShaw, 29 April 2024 - 05:17 PM.

  • SeymoreStars likes this


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






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics