Skip to content

Instantly share code, notes, and snippets.

@noetits
noetits / speech_noise.py
Created December 28, 2021 14:25 — forked from johnmeade/speech_noise.py
Use VAD to separate out regions without speech content, and estimate mean power of background noise
"""
Estimate background noise power level of a speech waveform.
Requires some non-speech regions in the wave.
Requirements:
pip install numpy librosa soundfile webrtcvad
MIT License John Meade 2021
"""