py-webrtcvad wrapper for trimming speech clips

Requirement

numpy, librosa and py-webrtcvad.

Installation

via pip

$ pip install pyvad

or

from github repository

$ pip install git+https://github.com/F-Tag/python-vad.git

Usage

from pyvad import vad
vact = vad(speech_data, speech_data_fs)

Example

Please see example.ipynb jupyter notebook.

License

MIT License (see LICENSE file).

Announcement

The version 0.1.0 update break backward compatibility.

The changes are as follows:

1. The hoplength argument has been changed to hop_length.
2. The trim returns (start_index, end_index) (return_sec argument is abolished).
3. Slightly changed the method of preprocessing a waveform in vad.
4. End of support for python 2.x.

You can see the new API in the example.ipynb.

The previous version is 0.0.8.

$ pip install pyvad==0.0.8