[daip] Cell size

[Eric Greisen]

Johan van der Walt writes:
 > I've been experimenting with the cell size parameter in IMAGR when
 > cleaning images. There is only one continuum source on this image. 
 > 
 > I noticed that the peak flux increases significantly if I decrease
 > the cell size. The S/N remains more or less the same as the cell size
 > changes.

      This is rather odd - I just asked an experienced user and he
says he has never seen this.  Are you sure you imaging parameters are
okay?  If an image has been undersampled (too large CELLSIZE) then odd
things can happen.  The program (through GUARD et al) attempts to warn
you if points are too far out in the uv plane and discards points if
they are outside an acceptable range.  If points are discarded the FIT
beam can look like about 3 points / beam but when you use a smaller
cell size it stays at 3 points / beam.  It is a good idea to plot the
beam (with SLICE / TKSLICE perhaps) to  check the shape versus
sampling.

 > 
 > Since the aim of this project is to check the non-detection of radio
 > continuum emission by other authors it is seems important to me that
 > a reliable peak flux density can be derived. 

      Actually, fit total flux is much more important than apparent
peak.

 > 
 > How should I handle this increase in peak flux with decreasing cell
 > size? The Cookbook says that there should be at least three or four
 > cells across the main lobe of the dirty beam. Is there an upper limit
 > to the number of cells?

      People often use smaller cells than this - there is no limit
although the Cleaning gets odd eventually.  Note that changing the
CELLSIZE or IMSIZE (or ROBUST, UVSIZE, UVBOX, UVBXFN, UVWTFN) changes
the data weighting and the expected noise and possibly the apparent
peak flux.

Eric Greisen