Description:
If supplied, this will
switch on an experimental method for handling missing or flagged data when filtering the time
streams, based on replacing missing data with zero, and then normalising the smoothed
data using a mask of good samples smoothed in the same way. The flt.filt_wlim value
specifies the minimum fraction of good values that must contribute to a filtered value.
For instance, if wlim is 0.9 then a filtered data value is flagged as bad unless at least 0.9
of the input values that contribute to it are good (i.e. have not been flagged as unusable
for any reason). Thus a high filt_wlim value (i.e. close to 1.0) will cause more data to be
rejected, and a low value (i.e. close to 0.0) will cause less data to be rejected. A value of
undef
causes the old filtering algorithm to be used that is based on filling gaps with artificial data. If the experimental
algorithm is used the following additional settings can be made: "
apod=0, fillgaps=0, noi.fillgaps=0"
.
[undef
]
Type:
real SMURF Usage
MAKEMAP, CALCQU
Copyright © 2012 University of British Columbia & the Science & Technology Facilities Council