Expert mode provides several ways of controlling how PSS uses the source part of its model, the point spread function. The psf may be evaluated only once over the image, or at every point of interest.
PSFCON - Assume constant PSF across field /YES/ >The default offered for this parameter is
YES unless the user has
specified a psf model at the PSF prompt.
The radius of PSS box used in the statistics is now selected. The case where the psf is constant is displayed thus:
Energy fraction Radius (pixels)
50% 2.1
68% 3.9
90% 15.2
95% 43.1
PSFPIX - Radius of psf box in pixels /4/ >
The default offered is 68% enclosed energy radius. The table is quite
different when the psf is allowed to vary. The example below is from a
ROSAT XRT image with the PSPC detector, with arcminute bins.
Energy fraction Off-axis angle
0 30 90 arcmin
50% 0.3 1.2 8.1 pixels
68% 0.4 1.6 10.4 pixels
90% 0.6 2.3 14.8 pixels
95% 0.7 2.6 17.4 pixels
PSFPIX - Radius of PSF box in pixels /1,2,10/ >
The reason for using a varying psf in this instrument is obvious! PSS
offers a default of the 68% enclosed energy radii at the three off-axis
angles quoted. Selecting this default means that PSS will vary the size
of the box across the field. The box size is found by linear interpolation
from the three radii given -- bound by the lower and upper values.
Alternative responses to this prompt could be,
PSFPIX > 1,2,13 ; Use 50% enclosed energy PSFPIX > 5 ; Use a constant radius of 5 pixelsAny three values may be specified, as long as they increase left to right. It is best to use radii which enclose the same fraction of psf energy with radius, as the fraction of the psf used affects the sensitivity of the parameterisation.
It should be noted that when a varying psf is used in SEARCH mode, PSS is performing a call to the psf system on every pixel in the input image slice. For this to be a practicable method for large image areas the user is advised to use the psf model facility (Appendix A.4).