ABSTRACT Nous montrons comment la fréquence variable de traits prosodiques, syntaxiques et intono... more ABSTRACT Nous montrons comment la fréquence variable de traits prosodiques, syntaxiques et intonosyntaxiques constitue un indice pour caractériser des types de textes à l’oral et les genres qui les engendrent. Parmi l’ensemble des traits formels explorés, on s’interroge sur la manière dont ils se répartissent en termes de redondance de l’information, de complémentarité ou de pertinence pour l’étude des variables situationnelles considérées. Lors de notre étude quantitative, nous confrontons deux méthodes de classification supervisée : arbres de décision et machine à vecteurs supports, et nous abordons le problème de la représentation graphique des traits grâce à l’analyse en composantes principales. L’ensemble donne un retour précis et inédit sur le rôle de l’interface intonosyntaxique dans l’identification des genres en français parlé.
In the area of large French speech corpora, there is a demonstrated need for a common prosodic no... more In the area of large French speech corpora, there is a demonstrated need for a common prosodic notation system allowing for easy data exchange, comparison, and automatic annotation. The major questions are: (1) how to develop a single simple scheme of prosodic transcription which could form the basis of guidelines for non-expert manual annotation (NEMA), used for linguistic teaching and
Prosodic transcription of spoken corpora relies mainly on the identification of perceived promine... more Prosodic transcription of spoken corpora relies mainly on the identification of perceived prominence. However, the manual annotation of prominent phenomena is extremely timeconsuming, and varies greatly from one expert to another. Automating this procedure would be of great importance. In this study, we present the first results of a methodology aiming at an automatic detection of prominence syllables. It is based on 1. a spontaneous French corpus that has been manually annotated according to a strict methodology and 2. some acoustic prosodic parameters, shown to be corpusindependent, that are used to detect prominent syllables. Some automatic tools, used to handle large corpora, are also described.
The goal of this paper is to shed new light on accentuation in French, more precisely to discuss ... more The goal of this paper is to shed new light on accentuation in French, more precisely to discuss the role of grammatical constraints and of phonetic factors implicated in the perception of French final and non-final accent. The study is based on the analysis of a 70-minute long corpus, including various speaking styles. The corpus has been annotated manually and automatically for prominence detection and tagged semi-automatically for grammatical categories. We first describe the rate of accentuation for each grammatical category (discussing the notion of "clitic" in French) and then discuss the divergences between manual and automatic prominence detection, in relation with the phonological structure.
In the area of prosody/discourse interface analysis, there is a demonstrated need for sharable pr... more In the area of prosody/discourse interface analysis, there is a demonstrated need for sharable prosodic notation systems for the segmentation of continuous speech into discourse units (Mettouchi & al. 2007, Simon & Degand 2011); that can be used independently to a particular discourse genre (Cresti & Moneglia2005). The fundamental issue for the transcription of prosody in discourse is to define stable prosodic units (phrasing and boundaries, accents, others?) which can be first: identified in common by experts on ...
This paper assesses the ability of a HMM-based speech synthesis systems to model the speech chara... more This paper assesses the ability of a HMM-based speech synthesis systems to model the speech characteristics of various speaking styles 1. A discrete/continuous HMM is presented to model the symbolic and acoustic speech characteristics of a speaking style. The proposed model is used to model the average characteristics of a speaking style that is shared among various speakers, depending on specific situations of speech communication. The evaluation consists of an identification experiment of 4 speaking ...
Actes des 27èmes journées d’étude sur la parole, 2008
This paper presents the inner details of three different algorithms for prominence detection. On ... more This paper presents the inner details of three different algorithms for prominence detection. On the basis of a 50- minute corpus made of five speaking styles and manually annotated for prominence, a quantitative evaluation compares the three approaches.
Proceedings of the Fourth Linguistic …, Jul 15, 2010
In the area of large French speech corpora, there is a demonstrated need for a common prosodic no... more In the area of large French speech corpora, there is a demonstrated need for a common prosodic notation system allowing for easy data exchange, comparison, and automatic annotation. The major questions are:(1) how to develop a single simple scheme of prosodic transcription which could form the basis of guidelines for non-expert manual annotation (NEMA), used for linguistic teaching and research;(2) based on this NEMA, how to establish reference prosodic corpora (RPC) for different discourse genres (Cresti and ...
This paper presents a work-in-progress on the automatic analysis of discourse genre in non-elicit... more This paper presents a work-in-progress on the automatic analysis of discourse genre in non-elicited speech. The study is focused on the development of bottom-up methods for automatic validation of discourse typologies found in linguistic descriptions (prosodic, syntactic, pragmatic and/or contextual and situational cues). The linguistic classification examined here opposes five discourse genres+/-controlled. To test this a priori classification under prosodic criteria, we propose a method that provides an automatic ...
ABSTRACT Nous montrons comment la fréquence variable de traits prosodiques, syntaxiques et intono... more ABSTRACT Nous montrons comment la fréquence variable de traits prosodiques, syntaxiques et intonosyntaxiques constitue un indice pour caractériser des types de textes à l’oral et les genres qui les engendrent. Parmi l’ensemble des traits formels explorés, on s’interroge sur la manière dont ils se répartissent en termes de redondance de l’information, de complémentarité ou de pertinence pour l’étude des variables situationnelles considérées. Lors de notre étude quantitative, nous confrontons deux méthodes de classification supervisée : arbres de décision et machine à vecteurs supports, et nous abordons le problème de la représentation graphique des traits grâce à l’analyse en composantes principales. L’ensemble donne un retour précis et inédit sur le rôle de l’interface intonosyntaxique dans l’identification des genres en français parlé.
In the area of large French speech corpora, there is a demonstrated need for a common prosodic no... more In the area of large French speech corpora, there is a demonstrated need for a common prosodic notation system allowing for easy data exchange, comparison, and automatic annotation. The major questions are: (1) how to develop a single simple scheme of prosodic transcription which could form the basis of guidelines for non-expert manual annotation (NEMA), used for linguistic teaching and
Prosodic transcription of spoken corpora relies mainly on the identification of perceived promine... more Prosodic transcription of spoken corpora relies mainly on the identification of perceived prominence. However, the manual annotation of prominent phenomena is extremely timeconsuming, and varies greatly from one expert to another. Automating this procedure would be of great importance. In this study, we present the first results of a methodology aiming at an automatic detection of prominence syllables. It is based on 1. a spontaneous French corpus that has been manually annotated according to a strict methodology and 2. some acoustic prosodic parameters, shown to be corpusindependent, that are used to detect prominent syllables. Some automatic tools, used to handle large corpora, are also described.
The goal of this paper is to shed new light on accentuation in French, more precisely to discuss ... more The goal of this paper is to shed new light on accentuation in French, more precisely to discuss the role of grammatical constraints and of phonetic factors implicated in the perception of French final and non-final accent. The study is based on the analysis of a 70-minute long corpus, including various speaking styles. The corpus has been annotated manually and automatically for prominence detection and tagged semi-automatically for grammatical categories. We first describe the rate of accentuation for each grammatical category (discussing the notion of "clitic" in French) and then discuss the divergences between manual and automatic prominence detection, in relation with the phonological structure.
In the area of prosody/discourse interface analysis, there is a demonstrated need for sharable pr... more In the area of prosody/discourse interface analysis, there is a demonstrated need for sharable prosodic notation systems for the segmentation of continuous speech into discourse units (Mettouchi & al. 2007, Simon & Degand 2011); that can be used independently to a particular discourse genre (Cresti & Moneglia2005). The fundamental issue for the transcription of prosody in discourse is to define stable prosodic units (phrasing and boundaries, accents, others?) which can be first: identified in common by experts on ...
This paper assesses the ability of a HMM-based speech synthesis systems to model the speech chara... more This paper assesses the ability of a HMM-based speech synthesis systems to model the speech characteristics of various speaking styles 1. A discrete/continuous HMM is presented to model the symbolic and acoustic speech characteristics of a speaking style. The proposed model is used to model the average characteristics of a speaking style that is shared among various speakers, depending on specific situations of speech communication. The evaluation consists of an identification experiment of 4 speaking ...
Actes des 27èmes journées d’étude sur la parole, 2008
This paper presents the inner details of three different algorithms for prominence detection. On ... more This paper presents the inner details of three different algorithms for prominence detection. On the basis of a 50- minute corpus made of five speaking styles and manually annotated for prominence, a quantitative evaluation compares the three approaches.
Proceedings of the Fourth Linguistic …, Jul 15, 2010
In the area of large French speech corpora, there is a demonstrated need for a common prosodic no... more In the area of large French speech corpora, there is a demonstrated need for a common prosodic notation system allowing for easy data exchange, comparison, and automatic annotation. The major questions are:(1) how to develop a single simple scheme of prosodic transcription which could form the basis of guidelines for non-expert manual annotation (NEMA), used for linguistic teaching and research;(2) based on this NEMA, how to establish reference prosodic corpora (RPC) for different discourse genres (Cresti and ...
This paper presents a work-in-progress on the automatic analysis of discourse genre in non-elicit... more This paper presents a work-in-progress on the automatic analysis of discourse genre in non-elicited speech. The study is focused on the development of bottom-up methods for automatic validation of discourse typologies found in linguistic descriptions (prosodic, syntactic, pragmatic and/or contextual and situational cues). The linguistic classification examined here opposes five discourse genres+/-controlled. To test this a priori classification under prosodic criteria, we propose a method that provides an automatic ...
Uploads
Papers by Anne Lacheret