Next: Photometric Analysis
Up: Precision Photometry at the
Previous: Introduction
To thoroughly study the reduction problems involved, simulated PC images
have been made. The point spread functions (PSFs) expected after
optical correction were derived for each filter using version 2.4 of the
Tiny Tim software (Krist, 1993). The images were initially created with
four times oversampling, then scaled down to 800
800 pixels, and
finally background and noise were added. These were estimated,
following the recommendations in the WFPC II Handbook, to correspond to
a signal to noise ratio of about 100 at magnitude V = 21.0. The star
density was derived using an extrapolated luminosity function determined
from earlier observations (Ardeberg et al. 1985), where a field in
the LMC bar was studied in detail. An example of a simulated image is
given in Fig. 2, which represents the field seen through the
Strömgren 
filter (F547M). This image contains about 40,000 stars
down to magnitude V = 29. Fig. 3 shows an expanded part from Fig. 2,
and in Fig. 4 the Tiny Tim PSF is seen, both with and without
oversampling. The FWHM of this filter PSF is 0.9 pixels. The
irregular shape of the poorly sampled stellar images is evident as well
as ``fixed pattern noise,'' resulting from the uneven, patchy, wings of
the PSF.
In addition to the original images, several modified versions were made to
test the response of the photometric programs:
- To test for PSF shape dependence, images were generated identically
to the original images, except that a Moffat function with FWHM = 1.4 pixels
was used instead of the Tiny Tim PSF.
- To test for the effect of the undersampling, a rebinned and
expanded image was made with twice the pixel resolution of the original
image (using pixel replication) and convolved with a Moffat function
with a FWHM of 2 pixels. This increased the FWHM of the PSF
from 0.9 pixels to 4.3 pixels.
Next: Photometric Analysis
Up: Precision Photometry at the
Previous: Introduction