Eric> That's one possibility. But I have both a 0.4s frame-time
Eric> dataset and a 3.2s frame-time dataset, with the former
Eric> nearly unaffected by pile-up but only 1/4 the length. So I
Eric> have another possibility: to just ratio the two datasets and
Eric> then correct the longer, 3.2s frame-time dataset by the
Eric> fraction of photons lost to pile-up. Of course this won't
Eric> do anything spectrally, but it will be a try at getting the
Eric> deep data to have a PSF which agrees with what mkpsf gives
Eric> me.
Eric> Any thoughts on doing this?
The ratio may be too noisy to get an empirical PSF correction term.
But the 0.4s should be good at constraining the spectral model and
thus better determining the pileup fraction and modeling the piled
psf. (depending on the counts). Or, instead of ratioing the data
directly, you could ratio a models, where you fit 2D gaussian to
the sources. (or some 2D profile, lacking an empirical
parameterization of the psf).
Are all sources near the optical axis, or near the same off-axis
poisition? Off-axis psf nowhere near gaussian, so parametric fitting
might be out of the question.
-- Dave
This archive was generated by hypermail 2b29 : Thu Dec 13 2012 - 01:00:06 EST