Session JEP orale - O1
Diversité des langues
Lundi 9 Juin - 13h30 15h30
L'espace vocalique perceptif dépend de la densité des systèmes vocaliques: Étude translinguistique en arabe marocain, en arabe jordanien et en français.
- Jalaleddin Al-Tamimi ( Université de Newcastle Upon Tyne)
- Résumé : The aim of this study was to evaluate the effect of vowel density on the size and on the internal organisation of vowel spaces in perception of speech from two Arabic dialects, Moroccan and Jordanian, and from French. 30 listeners (10 per system) identified the vocalic prototypes of their systems from a MOA design (Method Of Adjustment, Johnson et al., ). Results from this work corroborate the hypothesis of an effect of the inventory size on the size and internal organisation of vowel spaces: the larger the vowel inventory, the bigger the perceptual vowel space.
Modélisation acoustique multilingue pour la reconnaissance automatique de la parole non native
- Tien-Ping Tan ( Laboratoire LIG/GETALP)
- Laurent Besacier ( Laboratoire LIG/GETALP)
- Résumé : Automatic speech recognition system performance on non-native speakers is still poor. Non-native speech has a different characteristic compare to native speech as it is influenced by their mother tongue. Getting sufficient non-native speech to create a non-native acoustic model is not always possible. In this paper, we investigate the usage of multilingual acoustic models to adapt the target language acoustic model for non-native speakers. A hybrid of acoustic interpolation and merging is proposed for adapting the target acoustic model by using multilingual acoustic models without requiring the raw corpus. Three types of multilingual resources were tested for adapting non-native speakers: the native language of the speakers, non-native speech from a language different from the target language but spoken by the same native origin, and language close to the target language.
Aspects prosodiques du français parlé en Alsace, Belgique et Suisse
- Cécile Woehrling ( LIMSI-CNRS)
- Philippe Boula-de-Mareüil ( LIMSI-CNRS)
- Martine Adda-Decker ( LIMSI-CNRS)
- Résumé : The object of this article is a prosodic study of the French language spoken in Alsace, Belgium and Switzerland, also compared with standard French through large corpora of scripted and spontaneous speech. The data were segmented into phonemes by automatic alignment. pitch values were extracted and averaged over phonemes. Two features are addressed: initial stress (through pitch and duration correlates) and penultimate lengthening. Different patterns enable us to contrast the three varieties under investigation. Swiss speakers exhibit pitch rise and polysyllabic word onset lengthening in clitic/nonclitic sequences, while Alsatians tend to lengthen the initial vowel of nonclitic words. Belgians show prepausal penultimate lengthening.
Reconnaissance automatique de la parole en langue khmère : quelles unités pour la modélisation du langage et la modélisation acoustique?
- Sopheap Seng ( Laboratoire Informatique de Grenoble)
- Sethserey Sam ( Laboratoire Informatique de Grenoble)
- Viet-Bac Le ( Laboratoire Informatique de Grenoble)
- Brigitte Bigi ( Laboratoire Informatique de Grenoble)
- Laurent Besacier ( Laboratoire Informatique de Grenoble)
- Résumé : In this paper we present an overview on the development of a large vocabulary continuous speech recognition system for Khmer language. Methods and tools used for language resources collection for quick development of an ASR system for a new under-resourced language are presented. Face with the problem of lack of text data and the word error segmentation in language modeling, we investigate how different views of the text data (word and sub-word units) can be exploited for Khmer language modeling. We propose to work both at the model level (by making hybrid vocabularies with both word and sub-word units) as well as at the ASR output level (systems combination). For acoustic modeling, we use basic linguistic rules to automatically generate pronunciation dictionaries based on grapheme or phoneme. An experimental framework is setup to evaluate the performance of each modeling units.
Conséquences d'un cas de réparation phonologique en espagnol : special est un mot, scuro est un non-mot
- Pierre Hallé ( Laboratoire de Phonétique et Phonologie, CNRS-Paris 3)
- Fernando Cuetos ( Universidad de Oviedo)
- Juan Segui ( Laboratoire de Psychologie et des Neurosciences Cognitives, CNRS-Paris 5)
- Alberto Dominguez ( Universidad de La Laguna)
- Résumé : Cross-linguistic studies of speech perception have shown that speech perception is strongly dependent on the listener's native phonological system. In particular, nonnative illegal sound sequences are reinterpreted as permissible sequences in the native language. The present study addresses the case of /e/ prosthesis in Spanish, in which word-initial /s/+consonant clusters are banned. To this date, there is only scarce experimental evidence that such clusters be repaired in perception by /e/ prosthesis. In this study, we use auditory lexical decision to indirectly show that Spanish listeners hear an initial /e/ in either special (from especial) or scuro (from oscuro). A parallel is drawn with cases of phonological repair in Japanese and French.
Réalisations phonétiques des suites /tr/ en arabe marocain : données par EMA et par endoscopie
- Chakir Zeroual ( Faculté Pluridisciplinaire de Taza & Laboratoire de Phonétique et Phonologie, Paris. 3)
- Phil Hoole ( Institut de Phonétique de Munich.)
- John.-H. Esling ( University of Victoria, Victoria.)
- Lise Crevier-Buchman ( Hôpital Européen Georges Pompidou & Laboratoire de Phonétique et Phonologie, Paris. 3)
- Résumé : This study confirmed that the variations of the duration of the VOT between three voiceless coronal consonants in Moroccan Arabic (/ts/, /T/, [t]) are controlled by supralaryngeal and laryngeal adjustments. The EMA analysis showed that affricated /ts/ had a laminal contact and an important jaw elevation during its release. On the other hand, the emphatic /T/ and the non-affricated [t] in [tr] are apical consonants produced with a lower jaw position during the release. Several arguments have been presented to show that the supralaryngeal differences between /ts/ and [t] result from coarticulation with /r/. Endoscopic analysis showed a more important opening of the glottis during /ts/ compared to its allophone [t] produced before /r/.