Spectral Decomposition

2020-01-21

Using the python library librosa we can run a PCA on any audio signal to get the main n frequencies of the signal. We can then re-arrange those n frequency components to re-construct the sound.

import librosa

librosa.load('my_file.wav', sr=44100)
Components of spectral decomposition

Components of the spectral decomposition (left) and their rearangement in time.