BurstMon input
Data input
BurstMon takes data from single
interferometer channel. The name of
channel must be specified with option -c,
e.g. -c H1:LSC-AS_Q. Input
data rate must be equal or higher then BurstMon processing rate
specified with option -rate.
If input rate is higher then processing rate the data will be decimated
using biorthogonal wavelets. Only powers of two decimation ratios are
supported. DMT can serve data to BurstMon from next data sources:
- input from single frame file, option -infile, e.g. -infile HL-AS_Q1-730522113-640.gwf
- input from list of frame files, option -inlist, e.g. -inlist list128
- on-line data, neither -infile
nor -inlist options are used
Here is example of list of frame files stored in file 'list128':
../frames/H-AS1-729336752-16.gwf
../frames/H-AS1-729336768-16.gwf
../frames/H-AS1-729336784-16.gwf
../frames/H-AS1-729336800-16.gwf
../frames/H-AS1-729336816-16.gwf
../frames/H-AS1-729336832-16.gwf
../frames/H-AS1-729336848-16.gwf
../frames/H-AS1-729336864-16.gwf
Waveform input
BurstMon reads waveform as time series from
ASCII file supplied by user. The name of file with waveform must be
specified using option -j,
e.g. '-j sinegauss.txt'.
BurstMon can inject several waveforms at the same time, but injection
of too many waveforms may shift h50
estimates significantly. At least one waveform must be specified. As
waveform subjected to wavelet transform it must be long enough to be
decomposed up to level used in whitening. Padding by zeroes must be
done to make waveform time series long enough. For example, 1
second length is enough for whitening with resolution 16 Hz with the
use of symlets of order 32. Lower resolution requires longer waveform.
BurstMon exits with error if waveform doesn't meet requirements. On the
other hand waveform must be short compared to data stride. Otherwise
injection algorithm will not be able to inject waveforms close to data
interval boundaries.
Waveform file may have header at the begining.
Each line in header must start from '#' symbol. Keywords -rate, -size,
-peak and -name may be used to specify
sampling rate (in Hz, default is 16384), number of samples
(default is 1sec×rate), time of waveform peak (middle of the
series by default) and waveform name which will be used as part
of trend names. One keyword per line permitted only. Lines starting
with '#' without keywords will be ignored. Data part of waveform file
consists of lines with one floating point value per line. Waveform with
sampling rate higher than processing rate will be decimated by
BurstMon. All waveforms are normalized before entering injection pipeline.
Example of waveform file:
# Sine-Gaussian waveform:
frequency=555 Hz, duration =0.0036 sec, Q=8.8768801
# -size 8192
# -rate 16384
# -peak 0.25
# -name sg555q9
0
0
...
0.037056524
0.061891414
0.086177271
0.10860555
0.12783429
0.14255668
0.15157442
0.1538721
0.1486887
0.1355819
0.11448104
0.085724742
0.050079724
...
0
0
0
Configuration
file
BurstMon can read parameters from configuration file specified by
command line
option -i, for example, -i test.cf, where test.cf is:
-c H1:LSC-AS_Q
// IFO channel name
-v 10
// verbosity
level, 10 is maximum
-T
60
// input data stride, 60 seconds
-t
15.
// time stride for whitening, 15 seconds
-p
1.
// pixel fraction
-rate 8192.;
// sample rate for data
processing, Hz
-f 64
// frequency
resolution for cluster analysis, Hz
-fw 16
// frequency
resolution for whitening, Hz
-mcs
1
// min cluster size
-gate
0.02
// time gate for coincidence, sec
-nsim
10
// number of simultaneous injections
-ncycles
10
// number of injection cycles
-nmes
5
// number of measurements
-s
720
// dmtviewer history length, 720 mean 12 hours using 1 min stride
-b 2.
// data
padding length is 2 seconds
-w 32
// wavelet
filter length
-j sgauss555q9.txt
// read waveform
#0 from file
-j sgauss235q9.txt
// read waveform #1 from file
Example:
./burstMon
-i test.cf -inlist list128