Modeling Long and Short-term prosody for language identification

Jean-Luc Rouas (1, 2, 3, 4, 5)

1 L2F - Laboratorio de Sistemas de Lingua Falada
2 LaBRI - Laboratoire Bordelais de Recherche en Informatique
3 LIMSI - Laboratoire d'Informatique pour la Mécanique et les Sciences de l'Ingénieur
4 INRETS/LEOST - Laboratoire Electronique, Ondes et Signaux pour les Transports
5 IRIT-SAMoVA - Équipe Structuration, Analyse et MOdélisation de documents Vidéo et Audio

This paper addresses the problem of modeling prosody for language identification. The main goal is to validate (or invalidate) some languages characteristics proposed by the linguists by the mean of an automatic language identification (ALI) system. In previous papers, we defined a prosodic unit, the pseudo-syllable. Static modeling has proven the relevance of the pseudo-syllable unit for ALI. In this paper, we try to model the prosody dynamics. This is achieved by the separation of long-term and short-term components of prosody and the proposing of suitable models. Experiments are made on seven languages and the efficiency of the modeling is discussed.

Domaines: Traitement du signal et de l'image [eess.SP]; Traitement du signal et de l'image [eess.SP]

Jean-Luc Rouas. Modeling Long and Short-term prosody for language identification. 9th European Conf. on Speech Communication and Technology (INTERSPEECH'2005 - EUROSPEECH), Lisboa, Portugal, 04/09/05-08/09/05, 2005, Lisboa, Portugal. ⟨hal-00664989⟩