Implement getting of uncertanties in PSF photometry

I see that the get_uncertainties() method for DAOPhotPSFPhotometry currently raises a NotImplementedError.

What’s the ETA on this?

[hamogu]

@typpo I don’t think there is a “typical” uncertainty. That depends a lot on what you are observing and what instrument you use. E.g. an object observed in an older, 1 m telescope with hardware that needs more love and software that was written by a grad-student in IDL long ago and has never been upgraded, might give you 50% error, while the same star observed with HST might have 1 % errors. But then again, if you are using HST anyway, you are probably not looking at bright stars but super ultra-faint galaxies at high redshifts - and those will be so faint that your errors are larger again.

Are your sources faint compared to the background? In that case there is not that much you can do. Do you allow many free parameters (e.g. size of the PSF variable over the field or a complext background shape)? Try reducing the number of free parameters, e.g. try a simpler background model, fix the size of the PSF if it’s the same in every image, …

I still find the DAOPhot manual by Peter Stetson a insightful, if not easy, introduction. Ignore the software part (since you are using astropy), but read the general discussion that is interspersed.

[larrybradley]

@Gabriel-p There is an open PR (#516) that will hopefully be merged soon.