Audio & Video Processing in Python

For audio, soundfile reads and writes WAV, FLAC, OGG and other formats into NumPy arrays through libsndfile. librosa analyzes music and audio: spectrograms, tempo and beat tracking, pitch and features such as MFCCs used in machine learning. pydub offers a simple API for slicing, concatenating, fading and exporting audio, and relies on FFmpeg for formats other than WAV.

For video, PyAV binds FFmpeg's libraries directly and gives frame-level access to decoding and encoding, which suits processing pipelines. MoviePy is higher level and edits video in scripts: cutting, concatenating, compositing, titles and effects. For metadata, mutagen reads and writes tags across many audio formats, tinytag reads them with no dependencies, and beets is a music library manager that corrects tags using MusicBrainz.