DyniClick : open-source toolbox for stereo audio click detection, analysis, tracking and exploration

INSTALL

1 - Clone repository

$ git clone https://github.com/dynilib/dyniclick.git

2 - Install Miniconda.

3 - Create environnement:

$ conda env create -f environment.yml

USAGE EXAMPLES

First, activate the environnement:

$ source activate dyniclick

To get help on any script, use '-h' option, e.g.:

python dyniclick/click_detection.py -h

Click detection

python dyniclick/click_detection.py examples/example.wav examples/example.clicks.pk --bandpass_freq 10000 15000 15000 20000

-> the creates a pickle file with the click times and amplitudes

Click analysis

python dyniclick/click_analysis.py examples/example.wav examples/example.clicks.pk examples/example.feat.pk --tdoa_max 0.00047

Note: tdoa_max is the maximum delay between the hydrophones, i.e. the distance between the hydrophones divided by the speed of sound in water (~ 1500 m/s)

Click tracking

python dyniclick/click_tracking.py examples/example.feat.pk examples/example.tracks.pk --click_interval_max 0.3 --diff_max 0.000025 --amp_thres 0.1

Plot click features

python dyniclick/plot_utils.py examples/example.feat.pk --feat_cols 1 4 2 5 --feat_scale 1 1000 1000 0.001 --track_file examples/example.tracks.pk

Concatenate click data files

python dyniclick/concatenate_data.py examples/ examples/dataframe.h5 --track_root examples/

Explore clicks in the dataframe

jupyter notebook examples/dataframe.ipynb