brew install libsndfile
pip install scikits.audiolab
pip install pyaudio
This will record wave files:
python recordwav.py <output-filename>
In order to use the above command with the beamforming code, you must have a set of wave files of the same prefix followed by the recording number used. For example, my filenames could be called
recording1.wav
recording2.wav
recording3.wav
recording4.wav
The prefix in this case would be recording. In the end, each of these four recordings denote a speaker at a specific location.
Note that these are not the signals heard at the microphone. Those will be the sum of these recordings, each of which are delayed by a slight amount.
beamform.py <input-prefix> <speaker-id>
The speaker-id is an integer specifying which speaker you would like to hear. For example, in the above set of recordings, where the prefix is recording, if I'd like to hear a speaker at recording 3, the would then be 3.
For example, here is an example of what will work out of the box:
beamform.py wavs/labmates/recording- 16