open source audio identification
About the AcoustID project
AcoustID is a project providing a complete audio identification service, based entirely on open source software.
It consists of a client library for generating compact fingerprints from audio files, a large crowd-sourced database of audio fingerprints, many of which are linked to the MusicBrainz metadata database using their unique identifiers, and a web service that enables applications to quickly search in the fingerprint database.
Today the database holds over 94 million fingerprints, grouped into more than 76 million AcoustIDs, submitted by over 1 million contributors. Over 29 thousand applications are registered to use the web service, which answers around 28 million lookups a day. See the statistics for the exact numbers.
History
The project was started in 2010 by Lukáš Lalinský as a completely open source alternative to the proprietary MusicDNS service, which MusicBrainz Picard then used for identifying music files. It was the first open source audio identification technology and the first large database of audio fingerprints to be freely available, and today it is used by many other applications too, such as beets, MusicBee, FileBot and VLC. Since 2016, it has been funded by AcoustID OÜ, which provides the service to commercial users.
- Chromaprint is introduced, a library for extracting audio fingerprints.
- AcoustID is announced. Lookups run on PostgreSQL, using GIN indexes.
- The first fingerprint is added to the database, on October 8.
- 1.1 million submissions so far. The service moves to its own server, and the data becomes available for download.
- The server is rewritten from Java to Python, with a new web service that returns MusicBrainz metadata directly.
- Searches move to a dedicated fingerprint index server, acoustid-index, written in C++.
- 10 million fingerprints. The service handles over half a million lookups a day.
- MusicBrainz shows AcoustIDs on its recording pages.
- 25 million fingerprints.
- Five years of AcoustID: 8.3 million MusicBrainz recordings can be identified, and over 100 applications use the service every month.
- AcoustID OÜ is founded. It offers the service to commercial users, and the income pays for running the service and developing the open source project.
- 50 million fingerprints.
- The service, until then run with supervisord, is containerized with Docker and moves to Docker Swarm.
- AcoustID moves from Docker Swarm to a self-hosted Kubernetes cluster.
- 75 million fingerprints.
- The worst outage in the project's history: all services are down for three days.
- A record 34.9 million lookups in a single day.
- Submissions without metadata are no longer accepted.
- 90 million fingerprints.
- The service moves to a new self-hosted Kubernetes cluster, and searches switch to a new, much faster generation of acoustid-index, written from scratch in Zig.
Get involved
You can help by submitting fingerprints from your music collection, contributing to the code, or just spreading the word. Get in touch through the contact page.