Automatic analysis and visualization of urban soundscapes
This package contains:
- python functions that compute descriptors of environmental sound detection in a corpus of sound.
- HTML and javascript code to visualise a field recording corpus
To run the package, you need to install python 2.7, and the following python modules: numpy, scipy and scikits.
- Installing dependencies on Mac OS X:
install Xcode install homebrew -> http://brew.sh/ cmd: brew install gcc libsndfile cmd: brew install python --framework cmd: pip install numpy scipy install scikit-audioloab from source http://scikits.appspot.com/audiolab
- Installing dependencies on Linux:
sudo apt-get install python-numpy python-scipy install scikit-audioloab from source http://scikits.appspot.com/audiolab
This package contains functions to extract descriptors on the following sound events: motor, footstep, horn. To run the extraction of all the descriptors, run 'python main.py -MFH'
To visualize the TM-charts of the audio files, launch 'plot_™-CIESS.html', and browse files from the annotation folder.