[daip] New Ticket - [!HBU-785265]: REBYTE interchanged real and imaginary, I think.

[R. Craig Walker]

New Ticket: REBYTE interchanged real and imaginary, I think.

I was processing data on my Mac at home that I had recently converted from PPC to Intel formats using REBYTE.  Most programs seemed happy.  I could plot and display the data.  But CALIB got almost no solutions.  After much thrashing around, I examined two data sets that should be identical.  One was converted by REBYTE.  The other had been written to a FITS file by the PPC Mac and read in with the Intel Mac.  CALIB is happy with the FITS converted file. Both are the same size to the byte and have all the same header parameters (at least after I compressed the one from FITS - note that they are compressed which is likely to be important).  I compared them using PRTUV.  The amplitudes, UV values, times etc are identical.  But the phases are very different.  After a bit of staring, I noticed that you could convert by inverting the sign of phase of the FITS version, then adding 90 degrees to get the REBYTE version.  I believe that this is exactly what would happen if the real an!
 d imaginary components, in which the data are actually stored, are swapped.

I can have the disk with me Monday if someone wants to look at it.  I am using it now on my desktop, but it should be the same on my new laptop.

Unfortunately, I don't think I still have copies of the PPC formatted data, and I no longer have a PPC Mac.  This may complicate testing.  But I would guess that the symptoms I described should make confirming and fixing the problem in the code fairly straightforward.

Cheers,

Craig


Ticket Details
===================
Ticket ID: HBU-785265
Department: AIPS Data Processing
Priority: Default
Status: Open
Link:  https://help.nrao.edu/staff/index.php?_m=tickets&_a=viewticket&ticketid=379