#+
#  Name:
#     dimmconfig_blank_field.lis

#  Purpose:
#     A MAKEMAP configuration suitable for deep scans of blank fields.

#  Description:
#     This file provides values for parameters used by the SMURF:MAKEMAP
#     command to control the details of the map-making algorithm. To
#     use it, assign it to the CONFIG parameter on the command line when
#     running MAKEMAP. For instance:
#
#     % makemap config=^/star/share/smurf/dimmconfig_blank_field.lis
#
#     (substitute the path to your Starlink installation in place of
#     "/star").
#
#     This config is aimed primarily at reducing blank fields -- extremely
#     deep observations for detecting the individual point sources that
#     produce the cosmic infrared background.
#
#     Since there are no bright objects, a single high-pass filter
#     is applied at the start, and no other iterative filtering is
#     performed. A harsher high-pass filter is applied than in the
#     default configuration since no appreciable large-scale structure
#     is expected.

#  Notes:
#     - Any parameter not set in this configuration defaults to the value
#     specified in file $STARLINK_DIR/smurf_makemap.def.
#     - For a full list of all available parameters, and their purposes,
#     see the file $SMURF_DIR/smurf_makemap.def. All available parameters
#     are also documented in SUN/258, appendix "Configuration Parameters".
#     - A single parameter can be given different values to use when
#     processing 450 or 850 um data. This is done by including the
#     parameter twice, prefixing the parameter name with "450." and "850."

#  Authors:
#     HP: Harriet Parsons (JAC, Hawaii)
#     DSB: David Berry (JAC, Hawaii)

#  History:
#     12-FEB-2013 (HP):
#        General tidy.
#     14-FEB-2013 (DSB):
#        Use a standard prologue format. Re-instated comments describing
#        rationale for each value.
#     17-MAR-2015 (DSB):
#        Remove dependence on dimmconfig.lis.
#-

   numiter = 4

#  No FLT model is needed since we are filtering out low frequencies as
#  part of the initial cleaning process.
   modelorder = (com,ext,ast,noi)

#  Use time-domain de-spiker first because we are only doing a single
#  FFT-based high-pass filter before the iterations start.
   spikethresh = 10

#  Heavier high-pass filtering. Note that the default padding/apodization
#  that will be used is the number of samples that corresponds to the
#  period of the knee frequency in the high-pass filter (i.e. 200*(1/freq) )
   filt_edge_largescale = 200

#  Large-scale structure is not an issue so treat each subarray
#  independently for common-mode removal
   com.perarray = 1