[Aavso-photometry] Ensemble photometry - revised

[Michael Newberry]

Arne,

Thanks for the link to that paper---I had not seen it before. The author has 
brought together a great deal of relevant information in one place. It is 
really helpful to see all the noise sources listed there. But she missed 3 
of them. These are:

1. Quantization Noise, caused by losing resolution in the number of 
electrons that are chopped into frewer numbers of "counts". Measured in 
electrons, this is equal to 0.29 x the square root of the gain. My 1991 
paper in PASP describes this and you can get the reprint from the CFA 
astronomy Abstract service at 
http://adsabs.harvard.edu/abs/1991PASP..103..122N

2. Centroiding noise, caused by uncertainties in centering the aperture 
precisely on the stars. There is a little experiment showing the effect of 
this on our website at http://www.mirametrics.com in the "Tech Notes" 
section, note # 9.

3. Sky Extraction Noise: Call this whatever you want, but it is caused by 
determining the sky background underneath the star image. Various methods 
for estimating sky background bave their little peculiarities in response to 
the presence of cosmic rays and "abnormal pixels". This means that sometimes 
they bias the background a little high, occasionally a little low. Averaged 
over many stars on many frames, this adds a little noise component to the 
measurements. I have talked about this in a couple of photometry talks at 
the AAS (see for example the proceedings of the 1999
Precision Photometry" workshop in San Diego at 
http://adsabs.harvard.edu/abs/1999ASPC..189...74N ). I also recollect 
mentioning something about this in the Mira docs, but I don;t recall which 
version---maybe in the current version.

Michael Newberry
----- Original Message ----- 
From: "arne" <arne at aavso.org>
To: "Jim Roe" <jroe at jamesroe.com>
Cc: "Aavso-Photometry" <aavso-photometry at mira.aavso.org>
Sent: Friday, November 23, 2007 9:22 AM
Subject: Re: [Aavso-photometry] Ensemble photometry - revised


> Jim Roe wrote:
>> I'm trying to understand and learn how to do "ensemble photometry"
>> because I have come to believe that it will produce more accurate and
>> precise results.  Right?  I have re-visited threads on the list,
>> Googled, read several papers but I have yet to find a clear, concise
>> definition of the technique.  E.g.,
>>
>> Honeycutt (PASP 104:435-440, June 1992) states: "A simple (and often
>> quite adequate) technique for CCD ensemble photometry is to calculate
>> the difference between the instrumental magnitude for the program star
>> and a comparison magnitude obtained from the sum of the intensities of
>> perhaps a dozen of the brighter stars which appear in each exposure of
>> the series."  This doesn't seem right to me, perhaps it is a typo that
>> was mentioned in an earlier post.
>>
>> In another paper on the web by Hayes-Gehrke
>> <http://www.astro.umd.edu/~avondale/extra/Variability.html>
>> it is stated "Ensemble averages make this technique even more reliable.
>> Given at least a few dozen stars in the image, the assumption is made
>> that the average magnitude of these stars would remain constant from
>> frame to frame, regardless of the type or behavior of the individual
>> stars. "  This makes more sense to me.
>>
>> Finally, The MaximDl manual says: "If you tag more than one reference
>> star, you must set the magnitude for each; the results will be
>> averaged."  Does this, then, constitute "ensemble photometry?"
>>
>> If so, and I would just as soon continue using Maxim given my $400
>> investment therein, several questions arise.
>>
>> Which extra reference stars should I include?  Presumably only from
>> those for which photometry exists, but sometimes (often) there are
>> photometry stars that are much fainter than the program star and, unless
>> I expose for good S/N on them, their measurement errors will be greater
>> and it would seem adding them to the "ensemble" will degrade the whole
>> system.
>>
>> Would I consider all of the individual reference stars to be check stars
>> to be compared to the ensemble average?
>>
> Dang - one *important* word got left out.  Here is a revised reply.
> ------------------------------
> We are not recommending ensemble photometry submission until the
> new upload formats are announced (shortly; how shortly depends on
> the conclusion of our beta testing period, so I can't give a
> definitive date).
>
> The basic concept is that you have multiple comparison stars in your
> image, so why not use all of them to derive a magnitude for the
> target object?  You can think of it as (V-C1), (V-C2), etc., where you
> average these values together.  That reduces the effect of some bad
> measurement of any comparison star, variability of the comparison stars,
> systematic effects like transformation, etc.  So it is good, right?
> The devil lies in the details.  There are multiple methods of
> creating such an ensemble solution, from using flux space or magnitude
> space, averaging (V-C) or forming a master comparison star, how
> many stars form the ensemble, how do you weight the individual
> measurements, etc.  In addition, what magnitudes you use for the
> comparison stars impacts the final solution, and we cannot deconvolve
> this effect sometime later like we can if you use just a single
> comparison star.
>
> The new format gives a method for submitting ensemble results, and 
> includes
> a way in which we can adjust the photometry later.  I hope it works!  In
> addition, the new comparison star database photometry is precise enough
> that choice of ensemble stars should not affect solutions dramatically. 
> So
> we are getting closer to accepting ensemble results.
> Arne
> _______________________________________________
>
> Aavso-photometry mailing list
> Aavso-photometry at mira.aavso.org
> http://www.aavso.org/mailman/listinfo/aavso-photometry
>
>