Decibel pitches itself as a leader in the music metadata space by citing a deeper and more accurate dataset than its competitors. The slick video (embedded below) on its site is quite convincing. Decibel claims to collect music metadata from all sources all over the world, including record labels and users, check it for accuracy, and compile it into the database. It’s ambitions are set high with the goal of creating an industry standard in music metadata, and why not? Someone’s got to do it!
A few metrics on Decibel’s dataset demonstrate that it is a serious player: 12 million tracks, in 1.1 million albums, by 300,000 artists. Each track has 120 potential data fields to describe it. Artificially intelligent robots handle automatic cross translation between languages. Does your database have AI working for it?
There’s an air of secrecy on the site that doesn’t reveal live implementations using the Decibel API, but a few record labels are referred to anonymously. EMI is the only one mentioned by name. Secrecy may be the company theme, but with a data set as tempting as Decibel describes, public API documentation with a test URL would be a big step forward in convincing application developers to give it a shot. Maybe if the documentation was open, open source client API implementations would appear on github.