Dynamic Bayesian networks for symbolic polyphonic pitch modeling

Stanislaw Raczynski 1 Emmanuel Vincent 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 : The performance of many MIR analysis algorithms, most importantly polyphonic pitch transcription, can be improved by introducing musicological knowledge to the estimation process. We have developed a probabilistically rigorous musicological model that takes into account dependencies between consequent musical notes and consequent chords, as well as the dependencies between chords, notes and the observed note saliences. We investigate its modeling potential by measuring and comparing the cross-entropy with symbolic (MIDI) data.
Complete list of metadatas

Cited literature [17 references]  Display  Hide  Download

https://hal.inria.fr/hal-00663954
Contributor : Emmanuel Vincent <>
Submitted on : Friday, January 27, 2012 - 6:12:12 PM
Last modification on : Thursday, March 21, 2019 - 2:20:12 PM
Long-term archiving on : Saturday, April 28, 2012 - 2:43:52 AM

File

raczynski_SIGMUS11.pdf
Files produced by the author(s)

Identifiers

  • HAL Id : hal-00663954, version 1

Citation

Stanislaw Raczynski, Emmanuel Vincent, Shigeki Sagayama. Dynamic Bayesian networks for symbolic polyphonic pitch modeling. 91st IPSJ Special Interest Group on MUSic and computer (SIGMUS) Meeting, Jul 2011, Ibaraki, Japan. pp.no. 8. ⟨hal-00663954⟩

Share

Metrics

Record views

352

Files downloads

192