[daip] Re: POSSM with BPARM normalisation (fwd)

[Eric Greisen]

Thank you - we do have an error.

The DATGET routine smooths the data BEFORE applying the BP
calibration.  Therefore, the edge channel would be very enhanced by
adding it partly to channel 14 (in your case) and then scaled way up
by the BP cal which would have been correct for channel 15 with no
smoothing but which applies about a factor of >2 correction.

Smoothing the BP cal in BPASS produces a BP table that is incorrect
fpr unsmoothed data and that is not an aips error.

But this thought makes it clear that in the second case one does want
to smooth the data before the BP cal so that the smoothed BP cal
applies to the smoothed data, while in the former case one wants to
apply the unsmoothed BP cal and then smooth.  I see new options in
SMOOTH coming (uh oh!)  This will take a while since it is pretty
fundamental.

Eric Greisen