Various programs or spreadsheets for reducing photometry are sometimes described as "based on" Henden and Kaitchuck or Hall and Genet. But these references are not software functional specifications. They leave implementation details open to interpretation (or imagination), and different implementors can come up with software that gives different answers. I would like to offer PEP data reduction as an example. For those not familiar with PEP, it is single-channel photometry. Only one star can be measured at a time, and the AAVSO standard sequence is: comparison star, variable star, comp, variable, comp, variable, comp, check star, and comp. The reduced variable magnitude is an average of the three variable samples.

The AAVSO web program for reducing PEP data is WEBPEP. It operates as follows:

1. Reduce all net star counts to instrumental magnitudes

2. For each variable sample, time-interpolate a comparison instrumental magnitude from the bracketing comparison magnitudes and compute the differential instrumental magnitude.

3. Correct the differential instrumental magnitude for extinction.

4. Adjust the extinction-corrected magnitude to standard magnitude

The single differential instrumental magnitude for the check star is computed the same way, except that no adjustments are made for extinction or standardization.

The three variable standard differential magnitudes are averaged and added to the comparison standard magnitude to give a reduced magnitude. The lone check star differential magnitude is added to the comparison standard magnitude to give the reduced check magnitude.

I have already seen examples of PEP data reduction programs that vary from this algorithm (eg: interpolation done on net counts, not magnitudes; extinction correction applied to the final averaged differential magnitude instead of individual sample magnitudes; extinction applied before interpolation). I'm not saying these variations are wrong, but they are different and will give different results. It would not surprise me if the various CCD observers are likewise using reduction software with different implementation details. As we try to squeeze the highest accuracy out of AAVSO photometry, we should keep in mind the existence of systematic effects introduced by software.

Tom

Hi Tom,

have time (3 years) write software for reduction of photometric data, to be used with color CCD camera (webcam, DSLR) etc.. but applies to any photometry. actually already finished, but I am writing the protocol for sending data AAVSO extended format, and protocol installation for various operating systems (though he still performed tests to verify that everything is working really well, what they're going to take something over time). I used a different atmospheric extinction method, I get two images one near the zenith and one low at 60 or more, the coefficients at each height are calculated and entered into the program with information from the mass of air each set of coefficients and calculation program in each set corresponding to the mass of our measure air and sets the corresponding coefficient based on the information entered. I did this for two reasons, first because the software can optionally enable the iterative function (Gary Bruce) which is allowing better bring johnson blue channel system, and what ocaciona that the slope of the relationship (especially in blue for DSLR) varieties with the air mass. some days ago ask here about that amount to report in format to star comparison, if the instrumental on raw instrumental and corrected color and atmospheric extinction, I decided to use what you described, using the set and corrected if use several comparison stars. Here you can see something about the program, the windows are somewhat different, which are some things that have changed. but can give you an idea.

http://olichris.jimdo.com/rgb-fotocalc-software/

At this very moment just saw the message about reducing software written by Gordon Myers, for ccd, obviously this will be better for me to filters standard ccd cameras, I wrote mine ccd color oriented. but it is good news.

As an exercise I decided to grind through the equations for establishing delta(B-V) from observed b and v counts. Please check my math, attached :-)

Tom

Hi Tom.

Derive the B-V index from b-v is traditional in the AAVSO observers, because the filters they use are no filters that approach to B and V Johnson (are filters B and V Johnson), so I have no problems with the equations of transformation they use, but for color CCD a transformation of this kind is insufficient (to be more accurate, the equations AAVSO are not insufficient, are the channels of the color CCD which are insufficient, for this treatment), particularly because the channel b is shifted toward the red, and R channel has no tail to the IR if you have R cousins. to get an idea of that is that with only 9 iterations the v channel is converted in V, while b is always between 85 and 95 iterations to converted in B. I belive if you are using filters johnson I do not think you have problems using the traditional transformations.

Good morning Tom.

I'm looking at your math. I have avoided the use of K' in the equation, no matter how algebra is performed in the equation, in my case is:

+ (- k ' Xi)

V = v

+ (- k ' Xi)- e * (B-V) - Zwhy I replaced: + (- k ' Xi), for derived the coefficients used all photometric calibration sequence two distinct air masses?, is because I consider it more accurate than determining K only one or two stars, I prefer the standard you are using full field.

I'm not saying that's the case you should use, you wish to download costs to a draft of what will be the manual for my software, as there and did what I did because the infrared dispersion relations have a huge, but only perform with testing purposes, then may determine the precise relationships. apology language, Spanish, but you can translate the document, I have tried to do more visual and explanatory as possible.

http://olichris.jimdo.com/app/download/5753131618/53bbf3ef%2F29945b03af8...

if the document does not open there, try the address you had already given in the first link: Manual de RGB FotoCalc

http://olichris.jimdo.com/rgb-fotocalc-software/

Tom,

Attached is the procedure I use for data reduction for two color photometry. Being a new PEP observer I may have missed something. Any comments or help from those more experienced would be appreciated. I have not compared this procedure to the one you posted. (It is on my 'do do' list). I will be calculating the sensitivities of the attached equations to provide an idea of how accurate each parameter must be to achieve a desired accuracy.

As I noted in a direct email to you I was suprised how sensitive the B magnitude is to the second order extinction coefficient. In my last observation the value of this correction was about -0.042 mag compared to about 0.0021 for the first order extinction. (My delta air mass between the comp and variable was quite small, about 0.01). Again, being a new observer I am not sure if this is typical.

Jim

hi tom.

I speak Spanish, and I write in this language and turn it into google translator, although check the syntax, confusing my writing is always when I pass by the translator.

these are the links where Bruce Gary explains the method, your explanation is very long, the end is where it speaks of the method itself.

http://reductionism.net.seanic.net/CCD_TE/cte_alternative.html#OtherFilt...

http://brucegary.net/AllSky/x.htm

I will summarize the concept. you generate your magnitudes transformed, and of these, you get a B-V index, this what you enter again into the equation, using your instrumental magnitudes b and v, and the new index obtained from the this second transformed magnitudes is even more precise, the process you repeat until index changes obtained no more. Always use your instrumental magnitudes because your coefficients were derived for these, which change with each iteration is B-V obtained.

I wrote the software applies the iterations taking into account the number of decimal places we have set for work (max 8). few decimal produces few number of iterations to stop changing the end result.