acm-header
Sign In

Communications of the ACM

ACM TechNews

Computer Scientist Publishes Manifesto For Expressive Algorithmic Music


robot band, illustration

Credit: Science Node

Gerhard Widmer at Austria's Johannes Kepler University has published a manifesto for music information research (MIR) in the October issue of ACM Transactions on Intelligent Systems and Technology. He argues the MIR field, encompassing such technologies as music recommender systems and automated music recognition, should refocus so it can facilitate a "qualitative leap in musically intelligent systems." Widmer cites problems beyond computers' abilities — including differentiating between songs that a listener might find dull or interesting, or playing along with musicians in a musically sympathetic manner — as a starting point.

The manifesto says computers must be imbued with capabilities that transcend processing music as data and patterns, with an emphasis on perception and appreciation. "This is now the time for the MIR community to embark on massive feature/representation learning endeavors — much like the current trend in image analysis, which starts to produce quite spectacular results," Widmer says. "Given the computational and data-related demands, the MIR community should join forces and pool its resources, efforts, and learned models (in cases where the training data cannot be shared) — and indeed, it has already begun to do so."

Widmer details a project, Con Expressione, designed to characterize and recognize music's expressive aspects via performance to develop new models for perception and generating expressive music.

From Motherboard 
View Full Article

 

Abstracts Copyright © 2016 Information Inc., Bethesda, Maryland, USA


 

No entries found

Sign In for Full Access
» Forgot Password? » Create an ACM Web Account