Multiple pitch transcription using DBN-based musicological models

Stanislaw Raczynski 1 Emmanuel Vincent 2 Frédéric Bimbot 2 Shigeki Sagayama 1
2 METISS - Speech and sound data modeling and processing
IRISA - Institut de Recherche en Informatique et Systèmes Aléatoires, Inria Rennes – Bretagne Atlantique
Abstract : We propose a novel approach to solve the problem of estimating pitches of notes present in an audio signal. We have developed a probabilistically rigorous model that takes into account temporal dependencies between musical notes and between the underlying chords, as well as the instantaneous dependencies between chords, notes and the observed note saliences. We investigated its modeling ability by measuring the cross-entropy with symbolic (MIDI) data and then proceed to observe the model's performance in multiple pitch estimation of audio data.
Liste complète des métadonnées

Cited literature [10 references]  Display  Hide  Download

https://hal.inria.fr/inria-00544064
Contributor : Emmanuel Vincent <>
Submitted on : Tuesday, December 7, 2010 - 11:11:47 AM
Last modification on : Thursday, March 21, 2019 - 2:20:42 PM
Document(s) archivé(s) le : Tuesday, March 8, 2011 - 3:18:48 AM

File

raczynski_ISMIR10.pdf
Publisher files allowed on an open archive

Identifiers

  • HAL Id : inria-00544064, version 1

Citation

Stanislaw Raczynski, Emmanuel Vincent, Frédéric Bimbot, Shigeki Sagayama. Multiple pitch transcription using DBN-based musicological models. 2010 Int. Society for Music Information Retrieval Conf. (ISMIR), Aug 2010, Utrecht, Netherlands. pp.363--368. ⟨inria-00544064⟩

Share

Metrics

Record views

435

Files downloads

162