Re: isis-1.0.8 is available

Craig Heinke <cheinke_at_email.domain.hidden>The print_kernel gives me, e.g., (similar for psffrac~0.95 or for the other 
>source)
>isis> print_kernel(1);                          
>1: 0.211913     0.929297
>2: 0.029725     0.0673295
>3: 0.00277968   0.00325211
>4: 0.000194952  0.000117811
>5: 1.09384e-05  3.41426e-06
>*** pileup fraction: 0.0707029

This says that 21% of the frames contained a single photon in the
pileup region, 3 percent were contained 2 photons, etc...  The 3rd
column indicates that 93 percent of the events were single photon
events, 7 percent were due to 2 photons, etc.


> I expected a pileup fraction of ~9% based on PIMMS, so this is pretty close

I do not know how accurate pimms is.  Since it knows nothing about
grade migration, it will always over-estimate the pileup fraction.
Actually pileup fraction is a poor indicator of pileup.  In fact, if
you set alpha to 0, you should find that print_kernel will indicate that
the pileup fraction is also 0!

>(not sure what to make of the other output).  Alpha does change in the 
>opposite direction to psffrac, but the source model parameters don't seem to 
>change significantly (of order 1% or less).

You would also expect this behavior.

>For one source, I get alpha=0.43 when psffrac=0.95, alpha=0.52 when 
>psffrac=0.9.  For the other, I get alpha=0.53 when psffrac=0.95, and 
>alpha=0.62 when psffrac=0.9. The fits are about equally good for either set of 
>values. Generally alpha seems to have (90%) errors of ~0.07 to 0.14 or so 
>either way for any fixed psffrac value.

These values seem reasonable.  In any case, I hope that the model is
working for you.

Thanks,
--John