Measures and files the readnoise for UIST from a set of dark frames
The first dark frame is used to ‘clean up’ the array before statistics are done on the remaining frames.
The recipe calculates the readnoise as follows. It first derives the population variance estimate (PVE) of the dark frames, calculated on a per-pixel basis. It then finds the square root of the mean of this PVE image. If the images were taken before UT 2002 December 2, the readnoise is calculated as the product of the square root of the mean of the PVE, the number of reads minus one, the read interval, and the the gain, divided by the number of multiple reads. Otherwise, the readnoise is formed by multiplying the gain by the square root of the mean.
The readnoise is nominal if it falls between 38 and 45 e-/second.
The engineering log $ORAC_DATA_OUT/uist_array_tests.log.
The processing engines are from the Starlink packages KAPPA and CCDPACK.
Uses the Starlink NDF format.
History is recorded within the data files.
Error propagation is not used.