[daip] 31DEC06

[Eric Greisen]

Lawrence Rudnick writes:
 > Eric - we've gone to 31DEC06.  thanks for the MWFLT mods.  I don't know 
 > if you want me to write a few sentence description fo the "help" or 
 > "explain" files? to give the prescription for how to do filtering  
 > (basically, it's just, set cparm(1) and cparm(2) to at least 3x beam 
 > extent in pixels, set dparm 0,1. Run 'min' on map, run 'max' on output 
 > of 'min' to get all emission larger than 3x beam.  Can be subtracted 
 > from original map to get only more compact emission.

     I just incorporated the above remarks into the explain file with
a bit of grammatical improvemeents.

 >            Shea is making some progress on the WENSS editing --  FLAGR 
 > is so powerful, takes quite a while to figure out how to tune --
he's 

      The default is pretty good - set the SOLINT to 3.5x the
integration time and almost everything else default.  I am working
right now on the full polarization version.

 > mostly working with SPFLG, UVFLG right now on clearly bad channels.  At 
 > some point, it definitely would be worthwhile to be able to do clipping 
 > based on cross-hands.  However, it looks like you can't do the 
 > equivalent of a UVFLG clip on SPLIT data by amplitude (getting rid of 
 > individual high channels) in FLAGR, just get rid of a whole spectrum(IF, 
 > correlator)-baseline pair,  but isn't a channel-based clip what you'd 
 > want to do for RFI that's concentrated?

FLAGR is really designed for calibration sources and pseudo-continuum
data.  Line data is frequently handled in FLGIT which is used for
4-band, P-band, etc data.  Note that real line observers are nervous
about channel-dependent flagging since that can introduce apparent
spectral features due to changes in uv distribution.  Tasks CLIP and
CLIPM (multi-source) are available for plain clipping.

Eric Greisen