Computer scientists create large-scale data set of classical music

Posted on: December 1, 2016

“The composer Johann Sebastian Bach left behind an incomplete fugue upon his death, either as an unfinished work or perhaps as a puzzle for future composers to solve,” reads an unsigned Wednesday (11/30) article at Science Daily (Rockville, MD). MusicNet, “a classical music dataset released by University of Washington researchers … raises the likelihood that a computer could expertly finish the job.… It’s designed to allow machine-learning researchers and algorithms to tackle … open challenges … from note prediction to automated music transcription to listening recommendations based on the structure of a song a person likes…. MusicNet is a collection of 330 freely licensed classical music recordings with annotated labels that indicate the exact start and stop time of each individual note, what instrument plays the note and its position in the composition’s metrical structure. It includes more than 1 million individual labels across 34 hours of chamber music performances that can train computer algorithms to deconstruct, understand, predict and reassemble components of classical music…. It’s similar in design to ImageNet, a public dataset that revolutionized the field of computer vision by labeling basic objects—from penguins to parked cars to people—in millions of photographs.”

Posted December 1, 2016