Commit summary from repository starlink
------------------------------------
(Ed Chapin) smurf: add dimmconfig_calcnoise_com.lis and dimmconfig_calcnoise_pca.lis
93.0% applications/smurf/examples/
5.0% applications/smurf/scripts/
1.8% applications/smurf/
(Ed Chapin) smurf: some more comments about common-mode removal
100.0% applications/smurf/examples/
(Ed Chapin) smurf: use noiseclipprecom to flag noisy bolos before common-mode removal
2.0% applications/smurf/defaults/
11.1% applications/smurf/examples/
72.1% applications/smurf/libsmf/
14.6% applications/smurf/libsmurf/
(Ed Chapin) smurf: extra status check
100.0% applications/smurf/libsmurf/
Commits from repository starlink
-----------------------------
commit 73148f788cf8b9e8f2d5c0773b5fb3941321602e
Author: Ed Chapin <[log in to unmask]>
Date: Tue Aug 9 14:55:53 2011 -0700
smurf: add dimmconfig_calcnoise_com.lis and dimmconfig_calcnoise_pca.lis
I re-jigged the calcnoise configuration files so that we now have the
default (dimmconfig_calcnoise.lis) that just uses polynomial baseline
subtraction (robust, and fast, but perhaps noisy and biased upwards),
as well as a variant that uses the new common-mode removal functionality
(dimmconfig_calcnoise_com.lis) and a PCA version which is extremely
slow (dimmconfig_calcnoise_pca.lis). My thought is that the default
should be used in the pipeline, and probably the other two should
only be used for research purposes (PCA is probably best, though
not tested on a variety of data yet).
applications/smurf/Makefile.am | 2 +
.../smurf/examples/dimmconfig_calcnoise.lis | 44 +++++++++---------
.../smurf/examples/dimmconfig_calcnoise_com.lis | 31 ++++++++++++
.../smurf/examples/dimmconfig_calcnoise_pca.lis | 23 +++++++++
applications/smurf/scripts/sc2tsnoise | 4 +-
5 files changed, 81 insertions(+), 23 deletions(-)
commit 7ee2e122b4aab45daa31e52bf4c1ad0f9f38d342
Author: Ed Chapin <[log in to unmask]>
Date: Tue Aug 9 11:47:10 2011 -0700
smurf: some more comments about common-mode removal
.../smurf/examples/dimmconfig_calcnoise.lis | 10 ++++++++--
1 files changed, 8 insertions(+), 2 deletions(-)
commit e225120719377f5782c047bd87954facdbcf37d3
Author: Ed Chapin <[log in to unmask]>
Date: Tue Aug 9 10:43:00 2011 -0700
smurf: use noiseclipprecom to flag noisy bolos before common-mode removal
Common-mode removal depends on the removal of bad bolometers. Normally
we have the COM model iteratively throw-out bolos that don't match
the common-mode template, but this can be problematic for noise
measurements which (a) may have very short time series (so hard to
build up a template with structure that can be fit), and (b) we
want to keep as many bolos as possible for the purposes of fitting.
It seems that a reasonable approach is to optionally flag/removal
the noisiest bolometers immediately before common-mode removal
using this new parameter "noiseclipprecom". In the future, a nicer
approach would be to have a completely dynamic order for the
cleaning options, much like modelorder for the map-maker.
applications/smurf/defaults/smurf_sc2clean.def | 5 +-
applications/smurf/examples/dimmconfig.lis | 13 +++-
applications/smurf/libsmf/smf.h.source | 3 +-
applications/smurf/libsmf/smf_calcmodel_noi.c | 2 +-
applications/smurf/libsmf/smf_checkmem_dimm.c | 2 +-
applications/smurf/libsmf/smf_clean_smfArray.c | 68 ++++++++++++++-----
applications/smurf/libsmf/smf_filter_fromkeymap.c | 2 +-
applications/smurf/libsmf/smf_get_cleanpar.c | 27 +++++++-
applications/smurf/libsmf/smf_get_padding.c | 3 +-
applications/smurf/libsmf/smf_grp_related.c | 2 +-
applications/smurf/libsmf/smf_iteratemap.c | 2 +-
applications/smurf/libsmurf/smurf_calcnoise.c | 2 +-
applications/smurf/libsmurf/smurf_makemap.c | 12 +++-
applications/smurf/libsmurf/smurf_sc2clean.c | 2 +-
14 files changed, 111 insertions(+), 34 deletions(-)
commit 6c8a2167f673a77af0c5e9966d00abbdf400ee7b
Author: Ed Chapin <[log in to unmask]>
Date: Mon Aug 8 15:06:12 2011 -0700
smurf: extra status check
applications/smurf/libsmurf/smurf_calcnoise.c | 4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
|