task description

The gencal task provides a means of specifying antenna-based calibration values manually. The values are put in designated tables and applied to the data using other tasks (applycal, gaincal, bandpass, etc.). Several specialized calibrations are also generated with gencal.


Calibration types: caltype

The gencal task supports many different calibration types via the caltype parameter. These are listed here in two groups. Many of these options are part of Preparing for Calibration and more information about how they work can be found in that section.

Manual caltypes

The following enable directly specifying calibration factors for each specified pol, antenna, spw. Except where noted, each expects one real parameter value per specified pol, antenna, and spw.

  • 'amp'= amplitude correction
  • 'ph' = phase correction
  • 'sbd'= single-band delay (phase-frequency slope for each spw)
  • 'mbd'= multi-band delay (phase-frequency slope over all spw)
  • 'opac' = Tropospheric opacity (1 real parameter [in nepers] per antenna, spw)
  • 'antpos' = ITRF antenna position corrections (3 real parameters [in m] for each antenna; see below)
  • 'antposvla' = VLA-centric antenna position corrections (3 real parameters [in m] for each antenna; see below)

Specialized caltypes

The following caltype options automatically generate caltables from ancilliary information found in the MS or elsewhere. The pol, antenna, spw, and parameter options are ignored for these.

  • 'tsys' = Tsys from the MS.SYSCAL table (ALMA)
  • 'swpow' = VLA switched-power gains from MS.SYSPOWER, CALDEVICE
  • 'rq' = VLA requantizer gains _only_
  • 'swp/rq' = VLA switched-power gains divided by requantizer gain
  • 'gc' = Gain curve (zenith-angle-dependent gain) (VLA only; auto-lookup)
  • 'eff' = Antenna efficiency (sqrt(K/Jy)) (VLA only)
  • 'gceff' = Gain curve and efficiency (VLA only)
  • 'tecim' = Time-dependent TEC image specified in infile subparameter
  • 'antpos' = For VLA datasets, automatic lookup of antenna position corrections if antenna=''


Specifying calibration values: pol, antenna, spw, parameter

Generic calibration values for the "manual caltypes" listed above should be specified in the parameter argument as a list. The length of the list must correspond to the net length of the specific polarizations, antennas, and spws specified in the pol, antenna, and spw selection arguments.  The values specified in parameter will be duplicated over all members of any selection axis that is not explicitly specified (pol='', spw='' and/or antenna='') E.g., if pol=antenna=spw='', it only makes sense to specify a single parameter value (or three, for antpos and antposvla), and this will be duplicated for all pols, antennas, and spws. If multiple parameter values are specified, at least one of pol, spw, or antenna must be non-trivial, and the number of values in parameter must be consistent with the range of specified pol, spw, and/or antenna. E.g., if only a non-trivial spw selection is specified, then the parameter value list should match the number of spws specified, and these values will be duplicated for all polarizations and antennas. If more than one of pol, spw, and antenna is non-trivially specified, the number of parameter values specified should match the product of the number specified selection elements. The parameter values should be sorted by pol (fastest), antenna, and spw (slowest). Un-specified elements on non-trivially specified axes will be filled with nominal values (i.e., it is not necessary to exhaustively specify all elements on any axis or use nominal parameter values explicitly). Please consult the examples for additional guidance. There is currently no support for time-dependent calibration specfication; in all cases, the specified parameter values will be assumed constant in time (though their impact on the data may be time-dependent, depending on the caltype).

The same caltable can be specified for multiple runs of gencal, in which case the specified parameter values will be incorporated cumulatively. E.g., amplitude-like values (caltype='amp') multiply and phase-like values ('ph', 'sbd','mbd','antpos') add. Also, 'amp' and 'ph' calibrations can be incorporated into the same caltable (in separate cumulative runs), but each of the other types require their own unique caltable. A mechanism for specifying manual corrections via a text file will be provided in the future.

The calibration tables generated by gencal can be applied to the data in all other tasks that accept specified calibration for (pre-)application, e.g., applycal. gaincal, bandpass, etc.

Consult the Examples for more information on the many caltype options in gencal.

Notes on specific caltypes

  • 'antpos'  For antenna position corrections (caltype='antpos'), the antenna position offsets are specified in the ITRF frame. For the Karl G. Jansky VLA, automated lookup of the antenna position corrections is enabled when antenna is unspecified (antenna='') for this caltype. Note that this requires internet connection to access the VLA antenna position correction site.
  • 'antposvla'  For (old) pre-upgrade VLA position corrections, specify the values in the VLA-centric frame and gencal will rotate them to ITRF before storing them in the output caltable.
  • VLA switched power calibration is supported in three modes: 'swpow' (formerly 'evlagain', a syntax which has been deprecated) yields the formal VLA switched power calibration which describes voltage gain as sqrt(Pdif/Tcal) (used to correct the visibility data) and Tsys as Psum*Tcal/Pdif/2 (used to correct the weights). 'swpow' implicitly includes any requantizer gain scale and adjustments. 'rq' yields only the requantizer voltage gains (Tsys is set to 1.0 to avoid weight adjustments). 'swp/rq' yields the ordinary switched power voltage gains divided by the requantizer voltage gain (Tsys is calculated normally). The 'rq' and 'swp/rq' modes are are mainly intended for testing and evaluating the VLA switched power systems.
  •  For caltype='opac', only constant (in time) opacities are supported via gencal.  
  •  For gaincurve and efficiency (caltype='gc', 'gceff', or 'eff'), observatory-provided factors are determined per spw according to the observing frequencies. The parameter argument is ignored. These caltypes are currently only supported for VLA (including pre-upgrade VLA) processing. (Appropriate factors for ALMA are TBD.)