DISCUSSION OF SOME SELECTED AIPS CALIBRATION ADVERBS

 

Here follows a description of some of the AIPS calibration adverbs. More details about adverbs can be found in the AIPS HELP files.

Sources can be selected in a variety of ways by referring to them by name, by frequency, and by coded strings and numbers. All of these methods can be useful, and you will probably evolve your own method which uses some subset of them.

• calsour is an array for selecting one or more calibration sources by name.
• qual is a number that qualifies which source you are selecting. For example, the same source name can have different source qualifiers for different setups in the observation. ATLOD makes use of this facility for different freqids.
• calcode is a very useful adverb, also used to make additional qualification of calibrator sources. It may be set by the online system and thus entered into the SU table when you load your data into AIPS. But whether it is or isn't set, you will be able to change it to whatever you like with SETJY (see § 5). Then, it is often convenient to select sources by just their calcode instead of having to type their names into the sources adverb list all the time.
• selband and selfreq allow you to select data by absolute bandwidth and frequency. I think they are cumbersome and suggest that the adverb freqid discussed below is more useful.
• Each different group of simultaneously observed frequencies (such as the two simultaneous IFs you get with the ATCA) is designated by an integer number called the freqid).

  The freqids and their associated frequencies can be seen in the output from LISTR discussed in § 5. freqid is useful because each collection of frequencies must be calibrated separately, so keeping all other source selection parameters the same, but changing freqid allows simple change of frequency. This is handy when your observation has been set up with continual cycling through the same groups of frequencies (e.g., to obtain separated frequencies for multi-frequency synthesis).

  freqid values are stored in the FQ table, and must be specified one at a time; there is no way to get all of them selected together (except in UVPLT). Note that the selband and selfreq adverbs overide the freqid selection. However, any ambiguity must be resolved by freqid.

  Each freqid has one logical row in the FQ table. For each simultaneous frequency (IF) contributing to the freqid, the logical row has an offset in Hz from the reference frequency in the header, a channel width, a total bandwidth and a sideband indicator. Thus, the logical row is multi-dimensional; there are as may actual rows in the logical row as there are simultaneous frequencies (IFs).

• bif and eif refer to the IF axis which is displayed in the header (use verb IMHEAD to see it) of the visibility data base. There is some confusion about information stored in FQ tables and the IF axis. The IF axis allows storage of one group of equally spaced frequency channels (i.e., spectral channels) per IF frequency, and unevenly spaced IF frequencies. The disadvantage of the IF axis is that, unlike the use of the FQ table, all IFs must have the same bandwidth and channel increments. The advantage is that it is an axis which the calibration software can loop over. The freqid is a random parameter which cannot be looped over (i.e. you cannot select all freqids together). AT data which contains simultaneous frequencies is put on the IF axis when possible.
• docalib specifies whether you want to apply a calibration table to the data before performing the operation of the current task. When you are calibrating, usually you want this set to -1 (false) because there is no calibration to apply yet. However, following successful calibration, you will want to apply it to the plotting and imaging routines, so you put docalib =1 then.
• gainuse specifies the calibration (CL) table containing the calibration corrections for use when docalib is true.
• gainver specifies the input CL table to which an SN table is to be applied thus producing an output CL table (pointed at with gainuse.
• flagver is used to specify which FG table contains the flagging information. Note that the null value (0) is interpreted inconsistently in AIPS. Some tasks will take 0 to mean the highest version FG table (use IMHEAD to see how many you have), and others assume 0 means 1. Always check the HELP file to clarify this. If flagver is negative, no flagging table is applied. Make sure you have flagver set whenever you have done some editing.
• doband specifies whether you want to apply the bandpass correction BP table (computed by the task BPASS) to the data before performing the operation of the current task. For example, you would turn this on when imaging every channel in a spectral-line data base. See HELP doband for some extra information about fine distinctions in what value to give to doband. This is only necessary if you retain all the channels across the band. In simple continuum observations, you may end up with just one channel (see § 6), in which case this is irrelevant.
• bpver specifies which version of the BP table to apply if doband > 0.
• smooth specifies what sort of spectral smoothing to apply to the data before performing the operation of the current task. For example, it may be advantageous to apply Hanning smoothing to reduce ringing. See HELP smooth for details.

Last update : 20/03/96

1