Next: Focal plane geometry Up: ACIS/HRMA Point Spread Function Previous: Comparing Models with Data
Some basic data products will be generated for each data set. These include an image, surface plots, the PSF centroid and its moments for different grades and energy bandwidths, cuts through the central pixel and Gaussian (or other functional) fits to those cuts, plots of radias surface brightness and encircled energy for the PSF wings (not applicable inside a radius of 2 pixels), and spectra to show the pileup fraction and a computed pileup fraction for each pileup peak. This list will be augmented as analysis proceeds, revealing other metrics necessary and useful for summarizing the data.
More sophisticated techniques are possible for comparing simulated to actual event lists. Images compiled from event lists can be cross-correlated. We can apply functional fits to these images and attempt to account for the complicated azimuthal structure. Direct comparison of the event lists may be possible and quite revealing, employing statistical techniques such as maximum entropy or maximum likelihood to assess the degree of similarity in the data sets. We hope to understand and use existing knowledge in these areas; for example, similar problems are faced by those modeling the PSF of the Hubble Space Telescope (http://scivax.stsci.edu/ krist/tinytim.html).
So far, only sample data sets have been explored. We are working on the geometry problem (Sect. 6.5) and coordinating with MST to get SAOsac output for an XRCF source of finite size and distance. Code development and testing of PSU's version of the CCD simulator is ongoing.