MAG values reported by DAOPHOT/peak, DAOPHOT/nstar
ameisner wrote on Jan 30, 2009
I've noticed in fitting stars with peak and nstar that the values of FLUX reported by DAOPHOT/phot in general don't agree precisely with the MAG values reported by peak and nstar (peak and nstar appear to always give the same MAG value for a given star and psf). The disagreement is not huge, but is certainly beyond the error bars given by peak and nstar as MERR. I know that FLUX is just a sum of the pixel values within a circular aperture of radius given by the photpar/apertur parameter (of course having subtracted the background value from each pixel first). However, I would like to know what aperture radius is used to determine the value MAG (I can't find this in the documentation). My guess is that it is 2*fitrad where fitrad is the psf fitting radius used in the psf task, since when I use a psf with psfrad apertur/2 the disagreement between FLUX and MAG goes away, but when I use psfrad < apertur/2 I get MAG values indicating psf fluxes systematically lower than FLUX. I'm not doing crowded field photometry so I don't believe this discrepancy can be attributed to contamination from other sources within radius apertur of the point source of interest. Am I right on this, or is there any more specific information about the meaning of MAG that you could tell me? Thank you very much.
Francisco Valdes wrote on Jan 30, 2009
Hello,
The magnitude computed by peak is not an aperture magnitude so the concept of the aperture size is somewhat indirect. What is measured is the magnitude of a "psf star" (as set by the psf determine for a number of stars) which best fits a target star. The reason I said it is indirect is that there is an aperture associated with the psf fit, which is psfrad as you suspected. As noted in the documentation, psfrad should be large enough to sample most of the psf for bright stars. The reason it is ok to use a large value is that the fitting of the psf to a target is mathematically equivalent to a weighted aperture sum where the weignts are based on the psf. Therefore, pixels far from the core have much lower weight than those towards the center.
I hope this helps.
Frank Valdes
The magnitude computed by peak is not an aperture magnitude so the concept of the aperture size is somewhat indirect. What is measured is the magnitude of a "psf star" (as set by the psf determine for a number of stars) which best fits a target star. The reason I said it is indirect is that there is an aperture associated with the psf fit, which is psfrad as you suspected. As noted in the documentation, psfrad should be large enough to sample most of the psf for bright stars. The reason it is ok to use a large value is that the fitting of the psf to a target is mathematically equivalent to a weighted aperture sum where the weignts are based on the psf. Therefore, pixels far from the core have much lower weight than those towards the center.
I hope this helps.
Frank Valdes
Last post on Jan 30, 2009