From 7886b54d147e7b6b97953d5d3d97395904a5c44f Mon Sep 17 00:00:00 2001
From: Patrick Godwin <patrick.godwin@ligo.org>
Date: Fri, 4 Aug 2017 19:36:01 -0700
Subject: [PATCH] idq_multirate_datasource.py: fixed normalization in
downsampler, fixed dc offset that appears in certain channels
---
gstlal-ugly/bin/gstlal_idq_trigger_gen | 3 +-
.../python/idq_multirate_datasource.py | 31 ++++++++-----------
gstlal/python/pipeparts/__init__.py | 4 +--
3 files changed, 17 insertions(+), 21 deletions(-)
diff --git a/gstlal-ugly/bin/gstlal_idq_trigger_gen b/gstlal-ugly/bin/gstlal_idq_trigger_gen
index c5d483034a..7ce440de29 100755
--- a/gstlal-ugly/bin/gstlal_idq_trigger_gen
+++ b/gstlal-ugly/bin/gstlal_idq_trigger_gen
@@ -651,7 +651,7 @@ for channel in channels:
n_rates = int(numpy.log2(max_samp_rate/min_samp_rate) + 1)
if data_source_info.latency_output:
head[channel] = pipeparts.mklatency(pipeline, head[channel], name = 'stage2_beforeWhitening_%s' % channel)
- for rate, thishead in idq_multirate_datasource.mkwhitened_multirate_src(pipeline, head[channel], [min_samp_rate*2**i for i in range(n_rates)], int(samp_rate), instrument, channel_name = channel, width=32, quality=1, cutoff=None).items():
+ for rate, thishead in idq_multirate_datasource.mkwhitened_multirate_src(pipeline, head[channel], [min_samp_rate*2**i for i in range(n_rates)], int(samp_rate), instrument, channel_name = channel, width=32).items():
if data_source_info.latency_output:
thishead = pipeparts.mklatency(pipeline, thishead, name = 'stage3_afterWhitening_%s_%s' % (str(rate).zfill(5), channel))
if data_source_info.extension == 'ini':
@@ -683,6 +683,7 @@ for channel in channels:
thishead = pipeparts.mkcapsfilter(pipeline, thishead, caps = "audio/x-raw, format=Z64LE, rate=%i" % rate)
thishead = pipeparts.mktaginject(pipeline, thishead, "instrument=%s,channel-name=%s" %( instrument, channel))
thishead = pipeparts.mktrigger(pipeline, thishead, rate, max_snr = True)
+ #thishead = pipeparts.mktrigger(pipeline, thishead, rate, snr_thresh = 4)
if data_source_info.latency_output:
thishead = pipeparts.mklatency(pipeline, thishead, name = 'stage5_afterTrigger_%s_%s' % (str(rate).zfill(5), channel))
src[(channel, rate)] = thishead
diff --git a/gstlal-ugly/python/idq_multirate_datasource.py b/gstlal-ugly/python/idq_multirate_datasource.py
index 13f8b8d828..749d647155 100644
--- a/gstlal-ugly/python/idq_multirate_datasource.py
+++ b/gstlal-ugly/python/idq_multirate_datasource.py
@@ -105,7 +105,7 @@ from gstlal import datasource
#
# }
# @enddot
-def mkwhitened_multirate_src(pipeline, src, rates, native_rate, instrument, psd = None, psd_fft_length = 32, veto_segments = None, track_psd = False, block_duration = int(1 * Gst.SECOND), zero_pad = 0, width = 64, cutoff = 12, quality = 9, unit_normalize = True, channel_name = "hoft"):
+def mkwhitened_multirate_src(pipeline, src, rates, native_rate, instrument, psd = None, psd_fft_length = 32, veto_segments = None, track_psd = False, block_duration = int(1 * Gst.SECOND), zero_pad = 0, width = 64, channel_name = "hoft"):
"""!
Build pipeline stage to whiten and downsample auxiliary channels.
@@ -118,19 +118,19 @@ def mkwhitened_multirate_src(pipeline, src, rates, native_rate, instrument, psd
- psd_fft_length: length of fft used for whitening
- veto_segments: segments to mark as gaps after whitening
- track_psd: decide whether to dynamically track the spectrum or use the fixed spectrum provided
- - width: type convert to either 32 or 64 bit floati
- - cutoff: high pass frequency cutoff
- - quality: quality factor of whitening, determines power loss at highest frequencies
+ - width: type convert to either 32 or 64 bit float
- channel_name: channel to whiten and downsample
"""
#head = pipeparts.mkchecktimestamps(pipeline, src, "%s_%s_%d_timestamps" % (instrument, channel_name, native_rate))
max_rate = max(rates)
- head = pipeparts.mkaudioamplify(pipeline, src, 1. / math.sqrt(pipeparts.audioresample_variance_gain(quality, native_rate, max_rate)))
- head = pipeparts.mkaudiocheblimit(pipeline, head, cutoff = 0.9 * (max_rate/2))
- head = pipeparts.mkaudioundersample(pipeline, head)
- head = pipeparts.mkcapsfilter(pipeline, head, caps = "audio/x-raw, rate=%d" % max_rate)
+ if native_rate > max_rate:
+ head = pipeparts.mkaudiocheblimit(pipeline, src, cutoff = 0.9 * (max_rate/2), type = 2, ripple = 0.1)
+ head = pipeparts.mkaudioundersample(pipeline, head)
+ head = pipeparts.mkcapsfilter(pipeline, head, caps = "audio/x-raw, rate=%d" % max_rate)
+ else:
+ head = src
#
# add a reblock element. the whitener's gap support isn't 100% yet
@@ -181,6 +181,8 @@ def mkwhitened_multirate_src(pipeline, src, rates, native_rate, instrument, psd
psd.data.data = psd_data
kernel, latency, sample_rate = psd_fir_kernel.psd_to_linear_phase_whitening_fir_kernel(psd)
kernel, theta = psd_fir_kernel.linear_phase_fir_kernel_to_minimum_phase_whitening_fir_kernel(kernel, sample_rate)
+ # subtract DC offset from signal
+ kernel -= numpy.mean(kernel)
firbank.set_property("fir-matrix", numpy.array(kernel, ndmin = 2))
whiten.connect_after("notify::mean-psd", set_fir_psd, head, psd_fir_kernel)
@@ -249,7 +251,7 @@ def mkwhitened_multirate_src(pipeline, src, rates, native_rate, instrument, psd
#head = {max_rate: pipeparts.mktee(pipeline, head)}
tee = pipeparts.mktee(pipeline, head)
head = {rate: None for rate in rates}
- head[max_rate] = pipeparts.mkqueue(pipeline, tee, max_size_buffers = 0, max_size_bytes = 0, max_size_time = 0)
+ head[max_rate] = pipeparts.mkqueue(pipeline, tee, max_size_buffers = 0, max_size_bytes = 0, max_size_time = Gst.SECOND * 8)
#
# down-sample whitened time series to remaining target sample rates
# while applying an amplitude correction to adjust for low-pass
@@ -276,17 +278,11 @@ def mkwhitened_multirate_src(pipeline, src, rates, native_rate, instrument, psd
#
for rate in sorted(set(rates))[:-1]:
- head[rate] = pipeparts.mkqueue(pipeline, tee, max_size_buffers = 0, max_size_bytes = 0, max_size_time = 0)
- # downsample. make sure each output stream is unit
- # normalized, otherwise the audio resampler removes power
- # according to the rate difference and filter rolloff
- if unit_normalize:
- head[rate] = pipeparts.mkaudioamplify(pipeline, head[rate], 1. / math.sqrt(pipeparts.audioresample_variance_gain(quality, max_rate, rate)))
-
+ head[rate] = pipeparts.mkqueue(pipeline, tee, max_size_buffers = 0, max_size_bytes = 0, max_size_time = Gst.SECOND * 8)
# low pass filter + audio undersampler to downsample
# NOTE: as long as this fudge factor (0.9) for the high frequency cutoff is
# higher than the cutoff for the FIR bank basis, this should be fine
- head[rate] = pipeparts.mkaudiocheblimit(pipeline, head[rate], cutoff = 0.9 * (rate/2))
+ head[rate] = pipeparts.mkaudiocheblimit(pipeline, head[rate], cutoff = 0.9 * (rate/2), type = 2, ripple = 0.1)
head[rate] = pipeparts.mkaudioundersample(pipeline, head[rate])
head[rate] = pipeparts.mkcapsfilter(pipeline, head[rate], caps = "audio/x-raw, rate=%d" % rate)
@@ -298,4 +294,3 @@ def mkwhitened_multirate_src(pipeline, src, rates, native_rate, instrument, psd
#
return head
-
diff --git a/gstlal/python/pipeparts/__init__.py b/gstlal/python/pipeparts/__init__.py
index 7bb72a4d8f..e22564516e 100644
--- a/gstlal/python/pipeparts/__init__.py
+++ b/gstlal/python/pipeparts/__init__.py
@@ -468,8 +468,8 @@ def mkaudiochebband(pipeline, src, lower_frequency, upper_frequency, poles = 8):
## Adds a <a href="@gstpluginsgooddoc/gst-plugins-good-plugins-audiocheblimit.html">audiocheblimit</a> element to a pipeline with useful default properties
-def mkaudiocheblimit(pipeline, src, cutoff, mode = 0, poles = 8):
- return mkgeneric(pipeline, src, "audiocheblimit", cutoff = cutoff, mode = mode, poles = poles)
+def mkaudiocheblimit(pipeline, src, cutoff, mode = 0, poles = 8, type = 1, ripple = 0.25):
+ return mkgeneric(pipeline, src, "audiocheblimit", cutoff = cutoff, mode = mode, poles = poles, type = type, ripple = ripple)
## Adds a <a href="@gstpluginsgooddoc/gst-plugins-good-plugins-audioamplify.html">audioamplify</a> element to a pipeline with useful default properties
--
GitLab