I've been using VPhot for just over a year, and as I use it have come up with a couple questions on "Best Practices" and one suggestion for upgrade:
1. When critically examining my images in VPhot I note that some stars with a bit of color variation do not nicely fill the aperture using 1 aperture size across all colors. Do folks generally use a single aperture size for all colors or vary them to optimally fill the aperture for each color? For now, I go with one size aperture (the largest) for all colors, taking care not to overfill the aperture but on the other hand, underfilling for one or more colors.
2. Reporting multiple variables per field: is this generally done or avoided?
I may be over analyzing things, but that's my nature.
Suggestion: In "sequence" include aperture settings. As it is now, I need to be careful what the aperture is set at because the last aperture used from another star is what comes up as the aperture on the next star. It seems to me that aperture should track sequence...
Peter
BPEC
Peter:
1. As discussed in the vphot course, each image requires 'one' aperture for all stars on that image. If not, you would include different fractions of the stellar profile and get inconsistent magnitudes for different stars on that image. Your eyes are actually deceiving you a bit. If you look at the stellar profile and the calculated FWHM for each star, unless there is an optical distortion, all stars on an image nominally have the same FWHM. Thus, you should select one aperture based on the mean FWHM for all stars on the single image. This is hard to convince your eyes of due to screen stretching and brightness. Bright stars look bigger and you jump to the conclusion that they should have larger apertures. This is not true. Look at the stellar profile for bright and faint stars on the image and determine if the shape and pixel radius size is actually the same! You want to pick one aperture radius so that the same 'fraction' of the profile is used for all stars. It should not really be the radius that 'looks' best at including 'all' of the brightest stars. If you do pick the brightest aperture, you are just adding more background sky noise, especially to the fainter stars. The confusing visual appearance is that fainter stars underfill the aperture while brighter stars overfill the aperture. If you look at the stellar profile for both, they occupy the same 'fraction' for each star despite magnitude.
Obviously, the aperture may/will be changed from one image to another image, to address changes in seeing, focus, filter.
2. Your second question is actually interesting! First, why would you not want to analyze 'all' the variable stars present on your images? (As long as they are not too bright or too faint to give reliable magnitudes.) However, the qualifier is that IF you are the only person measuring these other stars and you only do it once, no researcher will have enough data to be useful at characterizing this star. IF you put personal effort into analysing them on a regular basis, the data would potentially be useful. Do what you want to do. Ask for an AUID and do it!
Ken
I'm fully with you on…
Ken,
I'm fully with you on item #2. For instance, in my BU TAU images I report 10 variables. That said, BU is 2 or 3 magnitudes brighter than the others, so I actually do "2 passes": I have one sequence for BU using comp/check stars close to its magnitude and another sequence (with a different name) for the other variables using comp/check stars closer to their magnitudes. Perhaps overkill, IDK, but I can't see how it will hurt. I suppose I should do a small study on 2 pass vs 1 pass.
Regarding #1: Perhaps I misspoke, but I was looking at the case of changing apertures between filters on the same stars or keeping the apertures constant for all filters on a set of images. Unless I can make a set of sequences (each storing the apertures used) for a given star but varying filter, it is just too onerous to vary the aperture for each image. That to me is too complicated to do manually and get it correct each time.
Peter
Peter:
For item 1, I thought you might be asking that but based on how you worded the question, I gave my answer as I did.
IF you really are talking about multiple images taken in different filters, the most efficient procedure is to run a time series of all the filter/images at once and use the variable aperture based on FWHM method. This method is adjusted to seeing/focusing. This addresses all of your concerns and it requires no manual change of aperture. There is no reason to keep a fixed aperture for all the images from different filters and in fact that is a poor practice. There are always a variety of reasons why the FWHM will change from one image to another. These reasons include mostly focus and seeing. Note that a filter change often produces a slight focus shift even for 'parfocal' filters. These may change even over a short period of time but certainly over days. Actually, it is for this reason, that it is not a good idea for vphot to maintain a single fixed aperture with the sequence. The FWHM may change somewhat from one day to another. I actually am willing to admit that my system/seeing are reasonably stable and I almost always use only one fixed aperture (4 pixels) for all my single image photometry. But, that is not a reasonable choice without practice and knowledge of your system.
Comments?
Ken
Peter,
"it is just too onerous to vary the aperture for each image. That to me is too complicated to do manually and get it correct each time."
How do you set the measuring aperture? I've never considered this onerous or complicated.
Do you look at each image for defects, signs of saturation, gradients, etc. before doing the measurements? That's the time to adjust the measuring aperture, if necessary. Also, if you use the Two Color transformed tool, the default 1.5 X radius of FWHM usually works well.
Phil