MusicBrainz and Internet Archive create cover art database
The non-profit music database service, MusicBrainz, has announced the Cover Art Archive, a collaboration between MusicBrainz and the Internet Archive project. The Cover Art Archive is a collection of music cover art images that aims to be freely and easily accessible. The project is hosted on the servers of the Internet Archive and was quietly launched last April behind the scenes at MusicBrainz.
Historically, fetching cover art for albums has been a tricky process. While it adds quite a lot of value to media player software, it traditionally presented several pitfalls both technically and legally. Many media players currently use Amazon's product database to display album covers, but to do so, the developers have to agree to the company's terms of service; in the past this has prevented some projects from adopting this approach. Other projects use Google's image search for the same purpose but that can lead to dissatisfying results as matching cover art to the actual music files can be a hard problem to solve.
Additionally, it is often unclear who holds the copyright to the images. In many instances, the record label publishing an album does not hold the copyright to the cover art, only a licence for it; often the copyright holder for the image is unknown.
The Cover Art Archive promises to solve these problems by tying its image collection to MusicBrainz's extensive database of music metadata. This makes finding definitive cover art for a given piece of music much easier, given the actual music has been identified correctly via MusicBrainz, and aggregating the images in one place should also simplify the copyright situation. The Internet Archive already has significant experience with hosting public domain content as well as material with abandoned or unknown copyright.
Developers can access the cover art database by using a JSON-based API. This standardises the process of accessing album art on an actual stable API instead of having different clients using different methods to aggregate images from disparate sources such as Amazon and Google. The API also introduces a known state to the request as it will not return an unrelated image like a search would, but rather either gives a URL or returns a 404 error. MusicBrainz is providing documentation for the API which currently has bindings for Java, Perl and C.
So far, the archive includes almost 100,000 images connected to approximately 54,000 releases. This represents five per cent of the current MusicBrainz database. Detailed information about the current catalogue of cover art images is available from MusicBrainz's Database Statistics page.