Best fwhm determinator
Frank Valdes wrote on Jul 25, 1997
Q: I have a question for advice on which is the best IRAF task to use to determine the FWHM of some point sources. I know this is a little like asking you which car should I buy, it probably depends on what one wants etc., but maybe if I give you some background to the type of data for which I want to determine this it will help. I want to determine the FWHM of images of point sources taken with the Yerkes 60-channel Far Infrared Camera on board the Kuiper Airborne Observatory. I have been using "radprof" up until now, but do not like the fact that it fits a spline rather than a gaussian, and also how sensitive it can be to the chosen outer radius of the fit. I would imagine our images are fairly close to being gaussian, sometimes elliptical as there can be an elongation in the direction that we chop the secondary mirror. Therefore I feel a task which determines the simga or FWHM of a gaussian fit would be better. I can see there is the task "fitpsf", but as far as I can tell there is very little control over things like how the sky is determined, unlike with radprof. Also, although fitpsf produces a plot of the radial profile with the "d" command, it does not seem to fit a gaussian line on this plot, something which would be nice for inspection purposes, as well as for recording in one's notebook. The other task I have come across which fits a gaussian is the "r" command in "imexamine". A: There are several possibilities. This is an important issue particularly at the telescope and observatories where they try and quote some kind of seeing measurement. The usual first tool is IMEXAMINE using the 'r' (or 'a') keys. In earlier versions of IRAF only a Gaussian fit was allowed. This was sensitive also to the fitting radius. In the V2.11beta release and also available as an add-on in the NMISC package / NEWIMEXAMINE there is the option to fit Moffat functions and to iterate on the fitting radius. The single best change is using a Moffat function because it handles the wings better and so lets the FWHM be less dependent on the details of the wing of the PSF. There is some discussion of this stuff in http://iraf.noao.edu/iraf/web/new_stuff/newpsf.html. The new IMEXAMINE also includes a direct measure of the FWHM and the enclosed flux estimate. The task PSFMEASURE in the NMISC add-on provides other PSF measurements. This has somewhat more sophisticated algorithms and is based on the enclosed flux curve rather than radial profile points. IMEXAMINE and PSFMEASURE still fit circular functions but also produce ellipticity and position angle information based on moments of the light distribution. Finally there are the tools for fitting PSF's in the photometry software (such as you mention). I don't know the details of this software though it is more sophisticated than RADPROF. Lindsey Davis might comment on this. So as you see there are a number of choices and deciding what is the "best" way has no simple answer. I suspect from your message that IMEXAMINE is actually what you should be using.
Last post on Jul 25, 1997