The COMPUTE sub-panel will use the list of Wide Band UV tables to compute a (phase and amplitude) Self calibration solution, and apply it to all UV tables. It will identify which UV tables correspond to a given Wide Band one, so that the proper self calibration solution is applied.
In SPLIT, ALL and CONTINUUM mode, self-calibration is computed on continuum only data, after line filtering. In SURVEY mode, if no line catalog exists, the whole bandwidth is used. This may be inappropriate if a strong line dominates the integrated flux and is spatially extended.
The self calibrated UV tables are placed in another sub-directory ( ./SELF/ by default, controlled by all%self).
A prefix (controlled by
all%prefix_self) is automatically
added to the file names to remind the user that they have been self-calibrated.
Caveats: CURRENT LIMITATIONS
- No provision is made to attempt to use spectral line flux when
the continuum flux is too low. You may use the SURVEY mode in such a case.
- The filtering mechanism (based on
UV_PREVIEW) may leave too few channels for
sensitivity in case of strong line contamination and/or steep spectral index. You
can change that by using manually
UV_PREVIEW with different parameters
on one wide-band window prior to relaunching the Pipeline. As
UV_PREVIEW parameters
are sticky, your selected parameters will be used by the Pipeline.
- The pipeline evaluates to first order the quality of the self-calibration
to check whether it should be applied or not. However, there are
cases that escape such automatic estimate. Visual inspection is
recommended, and if the Self-calibration solution is not good, you should not apply
it. Also, the COLLECT step may help at this stage,
by improving the S/N on some bands.