LSC Detector Characterization Working Group
Subgroup on Reduced Data
Sets
Leader: Jim Brau
There are two obvious ways to reduce data sets to more manageable volumes:
eliminating extraneous channels (extraneous to a given purpose) and decimating
the retained channels. Complicating this task are differing needs of various
experimenters. For example, a person wishing to study the correlation of
a seismometer's reading with the interferometer's differential mode readout
channel may want a customized data set for some period of time that contains
only those two channels, with the interferometer channel decimated to match
the seismometer's sampling rate. Another experimenter may wish to look
only at correlations between wave-front-sensor signals and light intensity
transmitted through each arm's end mirror. Again, only a short period of
data may be needed for this study.
To accommodate these various needs, a convenient software structure
must be set up that allows customized specification of reduced data sets.
The specifications would include (as a minimum) the channels to be recorded,
the decimation rate for each channel, and the time period for recording.
Other extensions to the specifications might include the type of decimation
filter to use on a given channel and interferometer state conditions (e.g.,
full lock) that should be satisfied before recording the data set.
Tasks:
-
Define parameters to be used in creating reduced data sets
-
Define associated software interface
-
Write, test and document the software (the hard part)
-
Create and make available sample reduced data sets from Hanford and/or
Livingston
In addition to setting up this structure which is geared toward creating
special-purpose data sets, this subgroup should also contribute strongly
to defining channel selection and decimation to be used in LIGO's standard
"Lightweight Data Set" which will be recorded on-site and distributed more
widely than the full data set to be archived at CACR.