A query by humming system that learns from experience

David Little*, David Raffensperger, Bryan Pardo

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contribution

10 Scopus citations

Abstract

Query-by-Humming (QBH) systems transcribe a sung or hummed query and search for related musical themes in a database, returning the most similar themes. Since it is not possible to predict all individual singer profiles before system deployment, a robust QBH system should be able to adapt to different singers after deployment. Currently deployed systems do not have this capability. We describe a new QBH system that learns from user provided feedback on the search results, letting the system improve while deployed, after only a few queries. This is made possible by a trainable note segmentation system, an easily parameterized singer error model and a straight-forward genetic algorithm. Results show significant improvement in performance given only ten example queries from a particular user.

Original languageEnglish (US)
Title of host publicationProceedings of the 8th International Conference on Music Information Retrieval, ISMIR 2007
Pages335-338
Number of pages4
StatePublished - 2007
Event8th International Conference on Music Information Retrieval, ISMIR 2007 - Vienna, Austria
Duration: Sep 23 2007Sep 27 2007

Publication series

NameProceedings of the 8th International Conference on Music Information Retrieval, ISMIR 2007

Other

Other8th International Conference on Music Information Retrieval, ISMIR 2007
Country/TerritoryAustria
CityVienna
Period9/23/079/27/07

ASJC Scopus subject areas

  • Music
  • Information Systems

Fingerprint

Dive into the research topics of 'A query by humming system that learns from experience'. Together they form a unique fingerprint.

Cite this