librosa

To make visuals follow the beat or analyze a song's rhythm and pitch, you first turn sound into numbers. librosa is the standard Python library for audio analysis: load audio, then detect beats and extract spectral and timbre features. It is widely used for music retrieval and speech front ends, and also gives automatic editing and animation their beat points. ISC licensed.

Features

Beats and tempo: find beat positions and overall tempo for music sync.

Spectral features: mel spectrograms, MFCC, chroma and more in one call.

Audio processing: resampling, time stretch without pitch change, harmonic percussive separation.

Visualization: plot waveforms and spectrograms with matplotlib.