[Difx-users] Handling lin-pol data with difx2fits

[Walter Brisken]

On Thu, 11 Apr 2019, Leonid Petrov wrote:

> Walter,
>
>  I got your point. But once again, the current version 2.5.2
> just propagates hard-coded R and L babels no matter what is in the
> difx output.

It is true that the poltype is hard coded.  With some experimentation we 
might be able to change that to X/Y as aspropriate.  At one point I 
believe AIPS had trouble with other than R or L.  We can look into this 
for DiFX 2.7.

The polarization used downstream is not based on that table though.  See 
Table 12 and Table 6 in the AIPS Memo 114.  The "STOKES" parameter 
describes the visibility data.  You can see from here that FITS does not 
allow description of linear cross-correlated against circular.

>  Handling H and V is rather different than X and Y. And the data
> analysis software should have a way to distinguish between X/Y
> and H/V. How to handle experiments when ASKAP-29 records X and Y
> in compliance with IAU definition and YARRA12M records H and V?
> And some stations, say YARRA12M, for some period of time record  R/L
> and for for other periods H/V. What is your suggestion? Keeping in
> mind an environment when we will have to process every day several
> experiments with different setups in quasi-automatic fashion.

You should be using the antenna mount type keyword (Table 16 in AIPS Memo 
114).  This will distinguish the kind of native linear polarization you 
have.

>  Related question: any ideas what kind of changes need be done
> for difx2fits to handle mixed polarization case when H/V is correlated
> against R/L and H/V against R-only (or L-only)?

As noted above, the FITS format is unable to describe such data, so this 
is not possible.

-Walter

> Leonid
> 2019.04.11_09:59:10
>

-- 
-------------------------
Walter Brisken
NRAO
Deputy Assistant Director for VLBA Development
(575)-835-7133 (office)
(505)-234-5912 (cell)