SecondHandSongs has teamed up with Columbia University to create the SecondHandSongs Dataset.
A team at Columbia compared the data in our database to their Million Song Dataset and found matches for more than 18000 performances and almost 6000 works. These are now released as a separate dataset publicly available for research purposes.
This is likely to be the largest cover dataset ever released for research and will be of great value for researchers all over the world. For example, our data will be used to research cover song recognition based on pitches and beat information.
SecondHandSongs is proud to participate in this project, which confirms our authority on cover song information once more.
Last edit: 2011-05-22 00:56:15 UTC by Denis