[daip] Fwd: Bandpass normalization

[egreisen]

On 2021-01-27 10:36, Frazer Owen wrote:
> The IF which is used, i.e. IF1, is on one extreme edge of the "good"
> spectral window and thus is not used later in the processing. Still it
> makes a good image just 10^3 too high. Whatever is happening is
> probably due to being on the edge. It also has been through MORIFs and
> thus is just the the extreme 1/2 of the 1st actual IF used. Originally
> the IFs were 5-8, each 4MHz wide. The corresponding 1/2 IF in the
> other side of the "good" spectral window does not make a useful image.
> 
> On 1/27/21 10:20 AM, Frazer Owen wrote:
>> I have put the uvdatabase I used to make the images at 
>> /home/pilabo/AIPS/scr/CYGAFRCYG1121.FITS
>> 
>> It is not very big. If you need the previous database before splitting 
>> that is pretty big.

That was a convenient database to plot since it is relatively small.  It 
shows GigaJy vis in IF 1 but also many MegaJy samples in the other IFs.  
I am not surprised that the imaging suffers some decorrelation in those 
IFs.  POSSM averaging over all times shows XX really bad in baselines to 
antenna 1 and YY really bad in other baselines in IF 1.  Averaging all 
times and baselines in POSSM shows YY in IF 1 still bad.  The other IFs 
give a clear and well behaved spectral index.  You could SPLAT say IFs 1 
and 2 and 3 for Cygnus only without calibration so that I could look at 
the input vis and the BP table.

Thanks

Eric