Next: Imaging multiple channels
Up: IMAGING
Previous: Imaging
Let us make the problem a little more complicated, and discuss data for
which all the channels have been averaged to one (channel 0) and for
which there is more than one time switched observing frequency. For
example, in the 6 cm band, you may have data at 3 pairs of frequencies,
spaced throughout the band, which you wish to combine to benefit from
the enhanced (u,v) coverage (assuming the spectral gradients are
sufficiently small).
- Now, these data will reside in three
freqids and two IFs (the simultaneous data go on the IF axis) within
the one multi-source file. However, you can only access one
freqid at a time. Thus, you must first split off the data
associated with each freqid and IF into a separate
calibrated single-source file. Apply SPLIT (See § 12) to
the channel 0 file (or use the channel-averaging option in SPLIT).
If you combined data from different observing runs, you will already
have done this.
- The next step is to join the files back up again ! Do this with the
task DBCON. I have already discussed DBCON in
§ 14, and suggest you read that section. However, this time you
must disable the frequency check.
DBCON |
reweight=0 | Leave weights alone |
dopos(1,1)=1 | Check phase centres |
dopos(2,1)=-1 | Don't check for frequency agreement |
doarray=-1 | Separate input files in time by 5 days |
- Note that u and v are generally stored in the file in
wavelengths at the reference frequency. This reference frequency will
vary between the files that you care going to concatenate. However,
DBCON does not recompute u and v for the new reference frequency
of the output file. But all is not lost, for single channel data, the
imaging tasks will not look at the reference frequency. They will
simply take u and v as they are in the file so that the the image
scales will come out correctly.
Similarly, you must split the IFs separately because the
concatenated file only has one FQ table, so that the second IF would not
be correctly described if the 2 IFs were left together. As an example,
if you had 3 freqids and 2 IFs and one channel, following SPLIT\
you should have 6 separate files, one for each freqid and IF
combination. Then you join them all up again with DBCON.
- Following all this messing about, you can now image the data with
UVMAP, MX, or HORUS in the same way as described in § 15.2
above.
Again I point out that MIRIAD handles multi-frequency data
in a much better fashion than AIPS.
Next: Imaging multiple channels
Up: IMAGING
Previous: Imaging
nkilleen@atnf.csiro.au