For visibility data, the current version of fits can cope with both single- and multi-source files, with single or multiple antenna configurations (earlier versions of fits were more limited). Flagging (FG) tables are copied (if present), but the flagging specified by these tables is not performed. An additional MIRIAD task, fgflag
(see Section 9.5), must be invoked to apply the flagging tables. Task fits does not read or use AIPS calibration tables (SN, CL, BP, and BL). Consequently if you want to preserve your AIPS calibration, you will need to apply the relevant tables with AIPS SPLIT before writing the FITS file.
fits can handle linearly and circularly polarised correlations, as well as correlations that have been converted to Stokes parameters. In addition, it knows that the ATCA produces linearly polarised data. Because AIPS does not handle linear polarisations well, it is common practice to label the linear polarisations of ATCA data as if they were circular. If fits finds data from the ATCA which is nominally circularly polarised, it issues a warning, and relabels the correlations as linearly polarised.
Alternately, you could correct the polarisation labeling explicitly within AIPS by changing the reference value of the Stokes axis to -5 . Do this with PUTHEAD . It is necessary to do this after SPLIT because it puts them back to circulars !
By default the AIPS
task to write FITS files, FITTP
, writes to the
directory /DATA/FITS.
This directory has a time-destroy limit of 1 day. Alternately you
can set up an environment variable to direct FITTP
to write the
FITS file to any directory. For example, after setting the environment
variable MYAREA with the UNIX command
% setenv MYAREA /DATA/ATELES_1/miriad/rsaultyou can direct FITTP to write into this directory, by setting the OUTFILE adverb appropriately. For example:
> OUTFILE = 'MYAREA:MYUV.FITS'will direct FITTP to write a FITS file
MYUV.FITS in
/DATA/ATELES_1/miriad/rsault.
A quick way to set an environment variable to the current working directory is with the pwd command:
% setenv MYAREA `pwd`
Note that the environment variable name must be in upper case, though the directory name itself need not be.
MIRIAD handles only the new binary table format ( donewtab=1 in FITTP ). Typical inputs to AIPS FITTP are:
Using format=1 (16-bit integers) for visibility datasets with many channels will result in an output MIRIAD dataset in compressed format. It also produces a much smaller FITS file -- which may be important if disk space is low. Note that FITS data compression is different from MIRIAD and AIPS data compression. FITS compression uses one scale factor for the entire dataset, whereas MIRIAD and AIPS compression uses one scale factor per visibility record.
In most cases converting the FITS file to MIRIAD
format is straightforward.
Apart from the op parameter, which indicates the operation to
be performed by fits
(op=uvin is used to read in a visibility FITS
file), you need only supply the names of an input FITS and output
MIRIAD
dataset.
Note that MIRIAD
follows the case conventions of the host
operating system. Thus, under Unix, MIRIAD
is case sensitive, and so
you must give the FITS file name in the correct case. As it has most
likely come from AIPS
, the file name will usually be in upper case.
For spectral line observations, the velocity reference frame used in
MIRIAD
(and some other velocity information) must be set when the visibility
dataset is created. The default is to extract the appropriate information
from the FITS file. Alternatively, the velocity parameter can be
used to override or alter the velocity information given in the FITS file.
See Chapter 15 for more information.
Leave everything else unset. Once fits has run to completion, you should see the new directory containing the `items' describing your visibility data.