Jitter Intermediate Data Products are Identical

[xcthulhu]

It appears that the different jitter data products in SPYFFIDATA are identical:

To reproduce:

# mkdir -p /tmp/spyffidata
# cd /tmp/spyffidata
# wget -qO- https://www.dropbox.com/s/0e4c2uk34phv4qx/SPyFFI_coreinputs.tar.gz | tar xz
# find . -type f -print0 | xargs -0 shasum

c4c751dee27b7047f32c980b8cf210ac0bfd36dc  ./inputs/AttErrTimeArcsec_80k.dat
364da0b684757ea59e27214f99f32fc54ca4ee23  ./inputs/cartoon.jitter
88264db54846de3ad885a3fa04765f98ce448bb2  ./inputs/covey_pickles.txt
04da11b315b2b1091d8edc8aa1c4b342b8f06f72  ./inputs/davenport_table1.txt
034c4911b6a2bc4a92d3421dee2f779071142f93  ./inputs/pickles_table2.txt
e8299c812c7353c72fc82fef0fed185956d520a4  ./intermediates/psfs/RRUasbuilt/focus0and10_stellartemp4350/originaldeblibrary.npy
1c0557c34b57abd0efe00b4100ca208d19b09efb  ./intermediates/psfs/RRUasbuilt/focus0and10_stellartemp4350/pixelizedlibrary_cartoon.jitter.cadence120s.unscaled_perfectpixels_11positions_11offsets.npy
1c0557c34b57abd0efe00b4100ca208d19b09efb  ./intermediates/psfs/RRUasbuilt/focus0and10_stellartemp4350/pixelizedlibrary_cartoon.jitter.cadence1800s.unscaled_perfectpixels_11positions_11offsets.npy
1c0557c34b57abd0efe00b4100ca208d19b09efb  ./intermediates/psfs/RRUasbuilt/focus0and10_stellartemp4350/pixelizedlibrary_cartoon.jitter.cadence2s.unscaled_perfectpixels_11positions_11offsets.npy

[zkbt]

The current code attempts to load a different PRF library depending on cadence and jitter. This was done to attempt to account for blurring of the PSF due to image motion within each exposure. To simulate this requires convolving high-resolution PSFs with an assumed kernel calculated from the average intra-exposure 2D jitter (including only translations, no rotations about the field center). It takes a long time, so at some point during development I must have copied the blurred PRF library from one cadence to all cadences. It's a kludge, and one that seems to have propagated far too long.

The effect of the jitter blurring is small. There are likely larger uncertainties left in going from Deb's optical models (as good as they are) to on-orbit performance. Also, due to the necessity to treat intra-exposure jitter as a single 2D convolution (e.g. that all stars everywhere in the field are translated identically, and with a jitter that is stationary across all exposures), the blurring is only crudely approximated. Not blurring at all is likely just as reasonable an approximation.

I suggest we simply stop trying to blur the PRFs in SPyFFI in a way that depends on cadence. There should be one PRF library. Maybe it's blurred to a two minute cadence for a typical jitter, but it should not try to adapt to different cadences.