Python Libraries

Libraries for analyzing music files

Librosa

Librosa load audio files including midi, wav files through its load() in librosa.core. This package provides a way of analyzing audio in its raw format. It does not have track information for different tracks of midi file.

The function returns

• y:np.ndarray [shape=(n,) or (2, n)]audio time series

• sr:number > 0 [scalar]sampling rate of y

Audio time series is a list of amplitude of the audio, normalized to (-1, 1). For instance if the value is 8192 and the file is a 16-bit wav, then this value is normalized to +0.25. Sample rate is something like 44100 for 44100 Hz audio.