Return to Home Page
Taichi Kato wrote: > Re: (fwd) problems with AIP4win > > >>When using AIP4Win, or even MIRA, on a target with a close companion (making >>it difficult or impossible to exclude the companion from the background >>annulus) I often wish the annulus could be 'uncoupled'. This would allow >>the background measurement to be made from a nearby clear area. Does the >>fact that this option isn't available (at least in the software I'm familiar >>with) mean it is too dangerous to use, or just a limitation of the program? > > > This is often dangerous, because this background selection does not > compensate for the linear gradient (sometimes present, particularly if the > sky is bright), but may be a choice when other selections are unavoidable. > If you choose this background selection, you are recommended to take two > or several sets of areas located symmetric to the object and comparison > star. However, as is well-written in CCD photometry textbook, the "clearest" > area is not the best selection for the background estimate. A selection > of "average populated" area in the frame is a better representative, > since the object's diaphragm should contain similar faint (usually invisible) > background objects. If one chooses the "clearest" area, the brightness > of the object is overestimated. When manually choosing the background > location, please be careful to choose the same location with respect to > the object between different frames. > This shows some of the problems associated with working at low signal/noise or effectively high sky background. The concept of "average populated" area does not hold, since the stellar measuring aperture is always smaller than the sky measuring aperture/annulus. Therefore, it is less likely that there will be faint contaminating sources. It also depends on how the sky value is determined (median/mean/mode/etc.). If there are "faint" background sources inside of the stellar measurement aperture, you can get significant contamination of the stellar measurement when working at low signal/noise. For example, if the object to be measured is S/N=1, then an object a half-magnitude fainter would not be visible, yet would change the measurement by a large amount. This gets worse as you go fainter since you are more likely to find a population of faint things nearby even seemingly uncrowded sources out of the galactic plane, and those faint things are even more likely to be extended sources. Work with the bright stuff! There are a hundred million objects that are bright and uncrowded that a typical amateur can observe; why beat your head against the tough problems? Arne
Return to the Daisaku Nogami
vsnet-adm@kusastro.kyoto-u.ac.jp
