AIPS Frequency Bugs

There are a number of bugs in AIPS which result in inconsistent or incorrect frequency information in the FITS file. These frequency bugs tend to become apparent when using two simultaneous IFs, particularly when the IFs are in different sidebands, or if one of the IFs is discarded at some stage. After reading with fits, it is strongly encouraged that you examine the frequencies that Miriad determines from the FITS file. Indeed it would be best if you checked these before writing a FITS file in AIPS, though this is more difficult because the frequency information is spread rather widely in AIPS. In Miriad, the easiest way to examine the frequency information is with task uvindex (see Section 10.3).

Errors in the FITS file frequency definition may affect the u-v coordinates, and so may affect the RA/DEC scales on any images that might be produced from faulty visibility datasets.

One fairly innocuous AIPS frequency bug is that the bandwidth of `channel 0' datasets can be wrong. Leaving this uncorrected will invalidate error estimates generated by Miriad tasks - which is not a major problem if you ignore these anyway!

Another, more serious, problem is that two locations in the FITS file give the ``reference frequency'' - the header and the antenna table. In principle these should agree. However there are instances when they will differ. Task fits will issue a warning when this happens, and will use the antenna table reference frequency as the true reference frequency. In some cases this will be the correct frequency. In others it will not.

There are two approaches to fixing the errors in the frequency definitions - either correct them in Miriad (using puthd and uvputhd) after reading the FITS file, or in AIPS (using PUTHEAD and TABED) before writing the FITS file. Which you will use will depend on your familiarity with the AIPS and Miriad visibility dataset structures and the complexity of your dataset (how many frequencies, etc), and where the bug occurred (are the u-v coordinates correct?). It is best to seek help, both to fix your data, and to aid tracking down the AIPS bugs.