Sonic Print: Timbre Classification with Live Training for Musical Applications

Jean-François Charles; Gil Dori; Joseph Norman

doi:10.5281/zenodo.4285338

Back

Sonic Print: Timbre Classification with Live Training for Musical Applications

Conference proceeding

Open access

Sonic Print: Timbre Classification with Live Training for Musical Applications

Jean-François Charles, Gil Dori and Joseph Norman

Proceedings of the 1st Joint Conference on AI Music Creativity, AIMC, Stockholm, Sweden

10/19/2020

DOI: 10.5281/zenodo.4285338

Files and links (1)

url

https://doi.org/10.5281/zenodo.4285338View

Published (Version of record) Open Access

Abstract

Sonic Print provides performers, composers and improvisers the ability to use automatic timbre recognition in live musical applications. Sonic Print can be trained quickly, during a performance, to discriminate between four classes of sounds. The musician can use the result of the live timbre classification to process sound with different audio effects, or to trigger musical events. Sonic Print is implemented as native Max code (widely used by artists), and as a Max for Live device (for integration in Ableton Live, a popular Digital Audio Workstation). We detail how Sonic Print contributed to two distinct creations in Israel and the United States.

Music Composition

Interactive machine learning

Live timbre classification

Multivariate linear regression

Music performance

Details

Title: Subtitle: Sonic Print: Timbre Classification with Live Training for Musical Applications
Creators: Jean-François Charles - University of Iowa, School of Music
Gil Dori
Joseph Norman - Grinnell College
Resource Type: Conference proceeding
Publication Details: Proceedings of the 1st Joint Conference on AI Music Creativity, AIMC, Stockholm, Sweden
DOI: 10.5281/zenodo.4285338
Language: English
Date published: 10/19/2020
Academic Unit: School of Music; University College Courses
Record Identifier: 9984182076602771

Metrics

79 Record Views