musique contemporaine

Ircam - articles scientifiques notice originale

From Boulez to Ballads: Training Ircam's Score Follower

Type

text
 

Genre(s)

article
 

Forme(s)

document numérique
 

Cette ressource est disponible chez l'organisme suivant : Ircam - Centre Pompidou

Identification

Titre

From Boulez to Ballads: Training Ircam's Score Follower
 

Nom(s)

Schwarz, Diemo (auteur)
 
Cont, Arshia (auteur)
 
Schnell, Norbert (auteur)
 

Publication

Barcelona, Spain , 2005
 

Description

Sujet(s)

score following   automatic accompaniment   performing   training   learning   Hidden Markov Models   Gaussian Mixture Models   probabilistic modeling
 

Résumé

This paper describes our attempt to make the Hidden Markov Model (HMM) score following system developed at Ircam sensible to past experiences in order to obtain better audio to score real-time alignment for musical applications. A new observation modeling based on Gaussian Mixture Models is developed which is trainable using a learning algorithm we call automatic discriminative training. The novelty of this system lies in the fact that this method, unlike classical methods for HMM training, is not concerned with modeling the music signal but with correctly choosing the sequence of music events that was performed. Besides obtaining better alignment, new system's parameters are controllable in a physical manner and the training algorithm learns different styles of music performance as discussed. Experience with the piece "...explosante-fixe..." by Boulez, and with an advanced karaoke system that allows to sing ballads in free tempo with automatic accompaniment are given.
 

Note(s)

Contribution au colloque ou congrès : International Computer Music Conference (ICMC)
 

Localisation

Envoyer la notice

Bookmark and Share 
 

Identifiant OAI

 

Date de la notice

2010-02-25 01:00:00
 

Identifiant portail

 

Contact