I've been analyzing data for the M67 standard star field. Whether using VPHOT or AIJ for the analysis, a number of the standard stars are close enough to other stars that measurements are contaminated by the flux coming from the second star.
How do folks deal with this?
I have eliminated the offending stars but it does make me wonder why these stars are in a 'standard' list in the first place? Would it not be better to address this at the list level, rather than have everyone do their own, different eliminations?
For my setup,…
For my setup, photometric results for the following standard stars in M67 get contaminated (determined with VPHOT and AIJ) due to the proximity of a nearby star:
The pixel scale for my C8 SCT/0.63FR/ASI533MC setup was approx. 0.6 arc sec per pixel for these measurements.
As Ken mentioned, the contamination is going to be a function of the resolving power of your setup. So, this list is likely only applicable to systems with a similar pixel scale.
In my analysis, I also removed photometric results where the SNR of calculated results was determined as < 50dB. Thus, in total I excluded results for 75 stars (21 for contamination and 54 for low SNR) out of the 178 possible ones for M67.
Arne generated the photometry using a large scope (~1m) at USNO Flagstaff. Thus, the stars were resolved in that scope.
So in your smaller scope, some do overlap. You just need to ignore those. Using TG, the overlapping stars show clearly as outliers and you can delete them by simply clicking on them.
IMHO, it is not a big deal to do this based on the size of your scope. I'm sure others agree with you but I prefer observers to understand what they are doing and use their own judgement to generate their transform coeffs.
If someone wants to curate some of the standard field sequences, please share them with myself and Ken for review and we can make them available to everyone.
We might want a couple sequences for each field, optimized for different fields of view. Say, 15', 60' and 180' maybe?
And please make sure the color spread is a wide as possible. This becomes the most important factor in terms of the quality of the coefficients developed.
As has been mentioned several times in this discussion, the blending of close stars for any particular subset of standard stars will still depend on the unique circumstances of each observer.
If we provide curated subsets of standard fields based on field size I think some new observers may get the (obviously mistaken) idea that they can safely just use the sequence that most closely matches their own FOV. That said, I do like the idea of giving new observers a head start in the right direct direction.
How about this slight modification of George's idea of curated standard field sequences:
On the AAVSO website there would be a collection of VPhot sequences, subsets of standard fields, which are contributed by experienced observers. Each sequence would have a description of scope, camera, pixel size, FOV size, typical seeing conditions, etc., that apply to that sequence . This archive would also have a brief explanation of the desirable characteristics of a standard sequence for calculating transforms. New observers could browse this collection, find a sequence with situations similar to their own, and have some advice about how to evaluate or modify their choice if this is needed.
Phil, it seems like the most we would need to do for these subset sequences is identify seeing conditions. For instance, we might have separate subset sequences in M67 (and the other standard clusters in VPhot) that only have stars that are well separated for a FWHM of 1", 2", 3", 4", and 5". You would choose a sequence based on the FWHM of your images. FOV, limiting S/N and thus depth of the sequence, and everything else is nicely handled in TG already. Really, I'm not even sure I see a purpose for subset sequences. TG nicely identifies outliers for elimination from the curve fit.
That makes sense…
That makes sense relating to a larger scope and its respective resolving power.
I did not want to just remove the outliers, as some of these outliers are outliers for reasons other than being near doubles. And, I did not want to lose that information in case it was relevant to the analysis I'm trying to undertake. Thus, in the end, I went through each outlier, checked whether it was overlapping, and removed only those that were.
I agree it is important to understand what your doing, and it does not take too long to do this for your equipment. That being said, it might be worth putting some sort of warning on the standard star field page because you can't just assume the data derived from a 'standard' star is going to be correct.
Data from standard stars are always "correct." But which standard stars one uses as a database for determining transformation coefficients is very dependent on the resolution of your rig. Small scopes are different than large scopes. I do agree that it is useful to discuss this in CCD Guide, and, in fact it is covered:
p. 53:"As with any crowded field, be careful not to measure any stars that are so close together that their images are 'blended'."
and: "Also be careful with star identification and in the case of multiple stars with the same identification, check the RA an Dec to make sure you know which are which."
There seems to be two basic approaches.
(1) Simply use all the stars and then eliminate the stinkers. This is useful for TG6.8 where there is a neat editing step that can be used for each filter, but not, IMO, for programs such as LesvePhotometry where the ensemble is specified directly and the resulting coefficients are automatically harvested.
(2) Select an ensemble based on insuring that you have a good spread of colors and magnitudes that fits your particular rig in terms of SNR over all filters. The "Magic 18 (Guide: p. 55) makes a good start, but watch out for saturation.
Both approaches seem to work.
Sorry, not a precise…
Sorry, not a precise choice of words. The magnitudes are indeed correct, but their proximity to other stars opens the possibility for contamination based on the resolution of the rig being used.