Closed dfm closed 12 years ago
That isn't a problem in the pixels-sample-the-PCPSF and well-sampled regimes.
Cool. Well if you're down then I can revert to the precomputed normalization that I was doing before... it's faster anyways!
I'm down; fix and close issue; we can discuss with BB tomorrow
done: 7f5835bb31e20d79c2416796d364a2113918e4b4
As Brendon pointed out, each PSF needs to be normalized differently (depending on how far the center of the PSF is from the center of the central pixel).
What is the best way to deal with this?