171 |
Se représenter les émotions du personnage du récit : contributions méthodologiques chez l’enfant âgé de huit à dix ans et perspectives interculturelles / The protagonist's emotion in children's representation : methodological contributions and cross-cultural perspectivesQuénette, Guy 19 December 2017 (has links)
La présente thèse examine la capacité des enfants âgés de huit à dix ans à produire des inférences sur l’état émotionnel du protagoniste au moment même où ils découvrent l’histoire (i.e., évaluation en temps réel). Quatre objectifs sont visés. Premièrement, la question de la spécificité des inférences émotionnelles produites est posée et appréhendée à partir de deux types de tâches déjà utilisées et validées chez l’adulte, à savoir une tâche de complètement et une tâche de choix multiples. Pour la tâche de complètement, les enfants sont invités à désigner par écrit l’émotion que pourrait ressentir le personnage à ce moment précis de l’histoire alors que pour la tâche de choix multiples ils doivent sélectionner parmi une liste d’émotions celle qui correspond le mieux à la situation décrite. Deuxièmement, outre les apports méthodologiques retirés de la comparaison de ces deux tâches, cette thèse interroge le développement des connaissances émotionnelles en explorant tant leur maîtrise des émotions de base que des émotions sociales. Troisièmement, s’agissant d’étudier les inférences émotionnelles élaborées en situation d’écoute d’histoires, l’expressivité du narrateur est considérée avec l’idée qu’une prosodie marquée pourrait favoriser l’identification des émotions du personnage. Enfin, dans l’optique d’ouvrir ce travail à une comparaison interculturelle, la capacité des enfants à inférer l’émotion du personnage à partir de textes de littérature jeunesse a été envisagée tant en France qu’à l’Ile Maurice. Au regard de ces différents objectifs, les retombées des études menées sont de deux ordres : méthodologique et théorique. La présente thèse permet ainsi d’alimenter la réflexion tant du point de vue des connaissances scientifiques actuellement disponibles sur ce sujet dans le domaine de la psychologie que du point de vue des pratiques pédagogiques à privilégier dans le domaine de l’éducation. / This thesis examines the ability of children aged eight to ten years to produce inferences about the emotional state of the protagonist while they discover the story (i.e., online evaluation). Four main objectives have to be distinguished. First, the specificity of the emotional inferences children produced is examined through two types of tasks already used and validated in adults, namely a completion task and a multiple-choice task. Regarding the completion task, children are invited to write down the emotion the character is supposed to feel at this moment of the story. For the multiple-choice task, they have to select from a list of emotional labels the one that best fits the situation described. Secondly, this thesis provides a relevant framework to assess the development of children’s emotional knowledge by considering their ability to infer both basic emotions and social emotions. Thirdly, our purpose is to highlight that the expressivity of the narrator can be of help in inferring the character’s emotions. Finally, in order to open this work to an intercultural comparison, children’s ability to infer the emotion of the character has been considered both in France and in Mauritius. In view of these different objectives, the series of studies we carried address both methodological and theoretical issues. Overall, the present thesis provides additional data to the current state of knowledge in psychology about children’s ability to represent the character’s emotional state and underlines related pedagogical contributions to reinforce practices in the field of education.
|
172 |
Individual Differences in the Perceptual Learning of Degraded Speech: Implications for Cochlear Implant Aural RehabilitationJanuary 2015 (has links)
abstract: In the noise and commotion of daily life, people achieve effective communication partly because spoken messages are replete with redundant information. Listeners exploit available contextual, linguistic, phonemic, and prosodic cues to decipher degraded speech. When other cues are absent or ambiguous, phonemic and prosodic cues are particularly important because they help identify word boundaries, a process known as lexical segmentation. Individuals vary in the degree to which they rely on phonemic or prosodic cues for lexical segmentation in degraded conditions.
Deafened individuals who use a cochlear implant have diminished access to fine frequency information in the speech signal, and show resulting difficulty perceiving phonemic and prosodic cues. Auditory training on phonemic elements improves word recognition for some listeners. Little is known, however, about the potential benefits of prosodic training, or the degree to which individual differences in cue use affect outcomes.
The present study used simulated cochlear implant stimulation to examine the effects of phonemic and prosodic training on lexical segmentation. Participants completed targeted training with either phonemic or prosodic cues, and received passive exposure to the non-targeted cue. Results show that acuity to the targeted cue improved after training. In addition, both targeted attention and passive exposure to prosodic features led to increased use of these cues for lexical segmentation. Individual differences in degree and source of benefit point to the importance of personalizing clinical intervention to increase flexible use of a range of perceptual strategies for understanding speech. / Dissertation/Thesis / Doctoral Dissertation Speech and Hearing Science 2015
|
173 |
Prosodie et structure du discours en français : analyse multiparamétrique et multidimensionnelle de narrations / Prosody and discourse structure in French : a multilayer multi-parametric analysis of narrationsFezza, Nassima 29 June 2016 (has links)
Cette thèse s’inscrit à l’interface d’une part de la phonétique et de la phonologie prosodiques et d’autre part de l’analyse du discours. Plus spécifiquement, ce travail a eu pour objet de décrire et d’expliquer les phénomènes relatifs au marquage prosodique de la structure du discours narratif en français. Notre méthodologie s’appuie sur l’analyse d’un corpus de 58mn de parole naturelle issue des narrations de 20 locuteurs français natifs, qui devaient raconter le contenu de BD. D’une part, le corpus a été aligné aux niveaux phonématique, syllabique, lexical et en unités inter-pausales ; des données acoustico-phonétiques ont ensuite été extraites à tous ces niveaux de granularité. D’autre part, nous avons appliqué une segmentation en parties narratives selon les critères du modèle de Labov (1972) qui se base sur des critères sémantiques et informationnels. La mise en correspondance de ces deux analyses a été faite sur deux niveaux structurels : global et local. Les résultats ont démontré l’existence d’un lien entre le découpage séquentiel narratif et les variations à long terme de la mélodie, du débit de parole et de l’intensité. Toutefois ces résultats n’ont pas mis en évidence un effet des variations prosodiques à court terme sur le découpage structurel. Par ailleurs, nous avons pu démontrer, pour chaque domaine prosodique, la réalisation d’une réinitialisation au niveau de l’enchainement narratif. Sans prôner une relation biunivoque entre organisation prosodique et structuration discursive, nos résultats démontrent ainsi que la prosodie, au même titre que la syntaxe ou la sémantique, participe aux marquages de la continuité et de la discontinuité informationnelles. / This thesis is at the interface of, on one hand, prosodic phonetics and phonology and, on the other hand, of discourse analysis. More specifically, this work aims to describe, explain and compare the phenomena related to prosodic marking of the narrative structure of discourse in French.Our methodology is based on the analysis of a 58-minute-long corpus of 20 natural speech narratives by French native speakers, who were asked to tell the content of comics without any warning. On one hand, the corpus was aligned with phonemic, syllabic, lexical and inter-pausales units levels; acoustic-phonetic data was then extracted automatically with all levels of granularity. On the second hand, we applied a narrative segmentation according to the criteria of Labov’s evaluative model (1972) which is based on semantic and informational criteria.The mapping of these two analyzes was made on two structural levels: global and local. The results showed the existence of a link between the narrative sequential segmentation and long-term variations of melody, speech rate and intensity. However, these results did not show an effect in the short term prosodic variations related to the structural segmentation. Furthermore, we demonstrated that for each prosodic domain, the realization of a reset at the narrative junction.Without advocating one relationship between prosodic and discursive organization structure, our results demonstrate, that prosody, as well as syntax or semantics, takes part in marking informational continuity and discontinuity.
|
174 |
Para um estudo da estruturação ritmica na fala disartrica / Towards a study of the rhythmic structural in dysarthric speechVieira, Jussara Melo 23 February 2007 (has links)
Orientador: Plinio Almeida Barbosa / Tese (doutorado) - Universidade Estadual de Campinas, Instituto de Estudos da Linguagem / Made available in DSpace on 2018-08-10T11:46:05Z (GMT). No. of bitstreams: 1
Vieira_JussaraMelo_D.pdf: 3447938 bytes, checksum: baa53bcff3e3572313562da26d96e687 (MD5)
Previous issue date: 2007 / Resumo: A disartria é uma desordem da fala decorrente de comprometimentos neuromusculares. Tais comprometimentos podem causar alterações na respiração, na fonação, na ressonância e na articulação da fala. Estes comprometimentos podem ser decorrentes de traumatismo craniano (TC). O TC é uma agressão cerebral resultante de uma pancada na cabeça e/ou uma queda com fratura craniana. Na fala disártrica pode haver, também, falha no mecanismo velofaríngeo devido à paralisia/paresia do palato mole. Neste caso, o resultado acústico para a fala é a hipernasalidade e a emissão de ar nasal, que podem ser tratadas através do uso de uma prótese de palato elevadora (PPE). Na disartria, o ritmo da fala também pode estar alterado, implicando alterações nos locais de acento frasal (AF), na inserção e na duração das pausas silenciosas (PS). Sendo assim, interessou-nos investigar a estruturação rítmica da fala disártrica decorrente de TC. Para tanto, obtivemos a colaboração de uma falante disártrica, usuária de PPE, falante nativa do português brasileiro (PB) acometida por TC. O estudo desta fala isártrica teve como objetivos: comparar as condições de uso e não uso da PPE; comparar a estruturação rítmica da fala disártrica brasileira (FDB) com uma fala sem comprometimentos (FSC); comparar a estruturação rítmica da FDB com uma fala isártrica francesa (FDF); estudar a relação sintaxeprosódia na estruturação rítmica da fala tanto na comparação com a FSC quanto na comparação com a FDF; adquirir melhor entendimento da fala disártrica pós TC, considerando o modelo de ritmo da fala para o PB de Barbosa (2006). Para cumprir estes objetivos, acompanhamos o tratamento com PPE da falante disártrica de 1995 a 2005. Neste período coletamos e analisamos oito leituras de um mesmo texto realizadas por ela usando a PPE e duas leituras sem esta prótese: uma em 1995 e outra em 2005. A FSC correspondeu à leitura do mesmo texto da FDB. A FDF correspondeu à leitura de outro texto francês. Nestas leituras realizamos análises acústicas e de sua organização rítmica em unidades vogal-vogal (VV), grupos acentuais (GA), AF e PS. Nas análises das oito leituras da FDB houve redução no número de PS, redução da participação das PS na duração do texto lido e aumento da taxa de elocução (TE). A estruturação rítmica da FDB com e sem a prótese foi a mesma. Houve diferença significante entre a FDB e a FSC para a TE e para a duração das PS, mas não para a duração dos GA nem para o número de unidades VV. Na comparação entre a FDB e a FDF não houve diferenças significativas quanto à TE, taxa de articulação e número de unidades VV por GA. Na relação sintaxe-prosódia, as marcas sintáticas IDF (independência forte) e COORD (independência de uma conjunção coordenada) foram significativas tanto para a FDB quanto para a FSC. Mas não para a FDF. Nosso trabalho contribuiu para destacar a pertinência de estudos que considerem a relação dinâmica entre fatores biomecânicos e lingüísticos da fala, notadamente para a fala comprometida e, em especial, para a fala disártrica pós TC / Abstract: Dysarthria is a speech disorder caused by neuromotors problems. These neuromotors problems may cause breath, phonation, resonance and speech articulation disturbance. The cause of these neuromotors problems can be the traumatic brain injury (TBI). TBI is a lesion that result of cerebral aggression, a collision in the head/or a fall happening cranial breaking. Dysarthric speech can have too velopharyngeal dysfunction by total/partial paralysis soft palate. In case of soft palate palsy the hypernasality may be present, with emission of nasal air during oral sounds. In order to deal with hypernasality and nasal air emission a palatal lift (PL) is indicated. The use of PL stimulates the correct movement of the soft palate, which produces the appropriate closure of the velopharyngeal port. Rhythm speech can be altered in dysarthric speech. We realize then a study of the rhythmic structure of the dysarthric speech by TBI. For doing so, eight dysarthric speaker?s readings during prosthesis treatment with the palatal lift, are compared with readings without prosthesis, with a reference speaker and a French dysarthric speaker. These comparisons are done by analyzing vowel-to-vowel units (VV), stress groups, phrase stress and silent pauses under the theoretical framework of Barbosa (2006)?s speech rhythm model applied to Brazilian Portuguese. In the eight readings of the Brazilian dysarthric speaker, there were reductions on the number of silent pauses and pause occurrence. An increase of speech rate was also observed. The results showed that there is no distinction in rhythmic structure with and without prosthesis. The comparative analysis between the Brazilian dysarthric speaker and the reference speaker showed significant differences in speech rate and silent pause duration, but no significant difference in the duration of stress groups or in the number of VV units. The comparative analysis between the Brazilian dysarthric speaker and the French dysarthric speaker showed no significant differences in speech rate, articulation rate and number of VV units inside stress groups. As regards the syntax-prosody interface, the syntactic markers IDF (strong independency) and COORD (coordinated conjunction independency) were significant for both the Brazilian dysarthric and the reference speaker. However, these markers were not significant for the French dysarthric speaker. This work reinforces the importance of considering both linguistics and biomechanics aspects on the analysis of dysarthric speech / Doutorado / Doutor em Linguística
|
175 |
Variações entoacionais na língua portuguesa falada por mulheres guatós / Intonational variations in the Portuguese spoken by Guató womenNatalina Sierra Assêncio Costa 08 February 2011 (has links)
Esta pesquisa descreve a entoação da língua portuguesa falada por mulheres guatós, fazendo comparação com mulheres não-índias. Foram analisadas as falas de quinze sujeitos. Descreve também a imanência da prosódia da língua guató, adquirida na infância, das guatós, mesmo depois de muito convívio com a população de Corumbá. Os dados foram obtidos por meio de pesquisa de campo e pesquisa bibliográfica. A análise e tabulação dos dados realizaram-se por meio do aplicativo ExProsodia. A análise centrou-se especificamente nas finalizações de frases em contexto diverso daquele dos falantes da língua portuguesa. Os resultados obtidos apontaram para a diferenciação na finalização das frases da categoria das meninas corumbaenses em relação às senhoras corumbaenses e às senhoras guatós, que não fizeram finalização descendente de frases assertivas. A diferenciação entre as senhoras guatós e as senhoras corumbaenses manifestou-se no tom médio das frases assertivas. / This research aims at describing the intonation of the Portuguese spoken by Guató women and comparing their intonational patterns to the ones produced by non Guató women. We analyzed the speech of fifteen subjects and observed that the Guató prosody, which is the language acquired in their childhood, remains although our subjects have been living with people from Corumbá for a long time. The data were collected through field research and a literature review. Analysis and tabulation of data were conducted using the software ExProsodia. The analysis focused specifically the end of sentences in different contexts from those in the speech of non Guató women. The results pointed out to differences at the end of sentences produced by young Corumbá women in relation to the production of old Guató and Corumbá women, who did not produce the falling contour of assertive sentences. The difference between the production of Guató and Corumbá women was found in the middle tone of assertive sentences.
|
176 |
Percepção do pareamento entre prosódia e sintaxe por falantes do português brasileiroCosta, Gabriela Fernandes 03 December 2015 (has links)
Submitted by Renata Lopes (renatasil82@gmail.com) on 2016-04-28T14:32:53Z
No. of bitstreams: 1
gabrielafernandescosta.pdf: 1513724 bytes, checksum: d25cbe1111ae096e0f3028ea5955cd21 (MD5) / Approved for entry into archive by Adriana Oliveira (adriana.oliveira@ufjf.edu.br) on 2016-05-02T01:10:23Z (GMT) No. of bitstreams: 1
gabrielafernandescosta.pdf: 1513724 bytes, checksum: d25cbe1111ae096e0f3028ea5955cd21 (MD5) / Made available in DSpace on 2016-05-02T01:10:23Z (GMT). No. of bitstreams: 1
gabrielafernandescosta.pdf: 1513724 bytes, checksum: d25cbe1111ae096e0f3028ea5955cd21 (MD5)
Previous issue date: 2015-12-03 / CAPES - Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / O trabalho investiga como as pistas prosódicas são usadas por bebês, durante a
aquisição de língua, e por adultos e crianças, no processamento linguístico.
Particularmente, focaliza-se o pareamento entre fronteiras prosódicas e sintáticas, e em
que medida as primeiras são percebidas como pista para identificação das segundas. São
tomadas como base teórica a Fonologia Prosódica (NESPOR & VOGEL, 1986) e a
hipótese do Bootstrapping Prosódico (MORGAN e DEMUTH, 1996; CHRISTOPHE et
al., 1997; 2008). De modo a investigar a relação sintaxe-prosódia e o continuum
perceptual defendido pela hipótese do Bootstrapping Prosódico, foram realizados dois
experimentos: um com bebês e um com crianças de oito anos e adultos. Verificamos se
o contorno prosódico bem formado, aliado a uma unidade sintática, facilita o
reconhecimento de elementos na sentença. Os resultados do experimento com bebês
apontam para uma preferência pelos estímulos bem formados prosódica e
sintaticamente. Os resultados do experimento com adultos revelaram um efeito
significativo de congruência prosódia/sintática. Por outro lado, as crianças de 8 anos
reagiram aleatoriamente aos estímulos, o que sugere que elas não atentaram para o
contorno prosódico bem feito associado ao sintagma sintático, como fizeram os adultos. / This study aims how the prosodic cues are used by babies during the language
acquisition, and by adults and children during the language processes. Specially, the
pairing between prosodic and syntactic boundaries are focused, and to what extent
prosodic boundaries is perceived as cues for recognition of syntactic boundaries. The
Prosodic Phonologic (NESPOR & VOGEL, 1986) and the Prosodic Bootstrapping
(MORGAN e DEMUTH, 1996; CHRISTOPHE et al., 1997; 2008) are used as
theoretical basis. To investigate the prosodic-syntax relationship and the perceptual
continuum holded by Prosodic Bootstrapping hypothesis, it was made two experiments:
with babies, 8-years-children and adults. It was verified if the well-formed prosodic
surround, combined with a syntactic unit, makes clause’s elements recognition easier.
The babies’ findings shows a preference for syntactic and prosodic well-formed stimuli.
The adults’ data, on one hand, reveals a relevant effect of syntactic/prosodic
congruence. The 8-years-children, on the other hand, had reacted to stimuli in a
randomly way, which suggests that those children did not pay attention to the wellformed
prosodic surround combined with the syntactic phrase, as was done by adults.
|
177 |
Repetições hesitativas em fala afásica e não-afásica / Hesitational repetitions in aphasic and nonaphasic speechViscardi, Janaisa M., 1980- 19 August 2018 (has links)
Orientador: Edwiges Maria Morato / Tese (doutorado) - Universidade Estadual de Campinas, Instituto de Estudos da Linguagem / Made available in DSpace on 2018-08-19T14:03:59Z (GMT). No. of bitstreams: 1
Viscardi_JanaisaM._D.pdf: 3660729 bytes, checksum: 06863fa1e7ad76498d3cb62c2e793fa3 (MD5)
Previous issue date: 2012 / Resumo: Esta pesquisa tem como objetivo central analisar os parâmetros prosódicos que permeiam a atividade da repetição hesitativa na produção de frases nominais e preposicionais, em contexto de fala afásica e não-afásica. Entre os aspectos de maior relevância para as análises realizadas neste trabalho, estão 1) a relação entre os movimentos de retração e projeção, como descritos em Auer (2009), no decorrer da produção da repetição hesitativa; 2) a relação entre os elementos prosódicos produzidos ao longo da repetição hesitativa, a saber, a frequência fundamental, a duração e a intensidade e, por fim, 3) a co-ocorrência entre os aspectos prosódicos e os movimentos de retração e projeção durante a produção da repetição hesitativa. Nesta pesquisa, os dados analisados foram extraídos do corpus Aphasiacervus, um acervo que reúne dados linguístico-interacionais registrados audiovisualmente dos encontros semanais ocorridos no Centro de Convivência de Afásicos, na UNICAMP. O interesse pelo estudo comparativo dos casos de repetição hesitativa se configura pela presença de especificidades linguísticas na produção das repetições hesitativas na fala afásica e não-afásica. Da comparação entre casos afásicos e não-afásicos, foi possível identificar as semelhanças e diferenças na realização dessas repetições. Entre as semelhanças, destacam-se 1) o alongamento vocálico associado ao movimento de curva entoacional descendente na caracterização da repetição hesitativa, 2) a presença dos mesmos movimentos de retração simples e múltipla em ambos os grupos e 3) uma maior similaridade das características prosódicas dos dois grupos nos casos de retração múltipla. Entre as diferenças mais importantes, destacam-se 1) a distribuição da ocorrência das retrações simples e múltiplas - afásicos produzem um número significativamente maior de retrações múltiplas -, 2) pausas significativamente mais longas nas sequências de retração simples em fala afásica, além de maior ocorrência da sequência M (P) R1 P na fala afásica; 3) a extensão das características hesitativas da produção ao longo de toda a estrutura M R1 Rx na fala afásica, enquanto na fala não-afásica essas características se encerram na penúltima repetição. A escolha pelos parâmetros apontados anteriormente se deu com o intuito de compreender os processos envolvidos na ocorrência dessas formas de repetição, o que poderá levar a uma melhor compreensão dos mecanismos empregados pelos sujeitos afásicos para se manterem atuantes na interação, além de permitir identificar a maior ou menor relevância de determinados recursos linguísticos em ambientes específicos de produção, como no caso das repetições hesitativas / Abstract: This thesis analyzes the prosodic parameters that underlie the activity of hesitational repetitions in the production of nominal and prepositional phrases by aphasic and nonaphasic speakers. The most relevant aspects of the analysis are 1) the relationship between the movements of retraction and projection, as described in Auer (2009a), during the production of hesitational repetitions, 2) the relationship between the prosodic parameters, namely, fundamental frequency, duration and intensity, and finally, 3) the cooccurrence between the prosodic parameters and the movements of retraction and projection during the production of hesitational repetitions. Among the similarities between the two groups analyzed are a) vowel lengthening associated with the downward intonation curve in the characterization of the hesitational repetitions, 2) the occurrence of the same retraction movements in both aphasics and nonaphasics and 3) a greater similarity of prosodic features in the two groups in cases of multiple retraction. Among the most important differences are 1) the distribution of single and multiple retractions - aphasics produce a significantly greater number of multiple retractions - and 2) the extent of the features of hesitation throughout the structure M R1 Rx in aphasic speech. This study promotes a better understanding of the use of prosodic features by aphasic subjects in order to keep their participation in the interaction, besides assigning the relevance of certain linguistic parameters in specific production environments, as in the case of hesitational repetitions / Doutorado / Linguistica / Doutor em Linguística
|
178 |
Oh foxy lady, where art thou? : A corpus based analysis of the word foxy, from a gender stereotype perspectivePellén, Angelica January 2009 (has links)
Abstract The aim of this essay is to establish whether or not the word foxy can serve to illustrate gender differences and gender stereotypes in English. The analysis is conducted by using one American English corpus and one British English corpus in order to make a comparison of the two English varieties. Apart from the comparative study, foxy is examined and categorized according to gender and a number of features to help answering the research questions which are: • What difference in meaning, if any, does the word foxy carry when used for males, females and inanimate things? • Can the word foxy serve to illustrate gender stereotypes in English? • Are there any differences regarding how foxy is used in American English compared to British English? Throughout the essay previous studies are presented, terms and tools that have been used are defined and argued for. One of the conclusions drawn in this study is that there is a significant difference in meaning when foxy is used in American English compared to British English. There are, however, also differences concerning the use of foxy when referring to males, females and inanimate things. Keywords: Collocation, corpus studies, foxy, gender, language, linguistics, semantic prosody, stereotypes.
|
179 |
Synthèse incrémentale de la parole à partir du texte / Incremental text-to-speech synthesisPouget, Maël 23 June 2017 (has links)
Ce travail de thèse porte sur un nouveau paradigme pour la synthèse de la parole à partir du texte, à savoir la synthèse incrémentale. L'objectif est de délivrer la parole de synthèse au fur et à mesure de la saisie du texte par l'utilisateur, contrairement aux systèmes classiques pour lesquels la synthèse est déclenchée après la saisie d'une ou plusieurs phrases. L'application principale visée est l'aide aux personnes présentant un trouble sévère de la communication orale, et communiquant principalement à l'aide d'un synthétiseur vocal. Un synthétiseur vocal incrémental permettrait de fluidifier une conversation en limitant le temps que passe l'interlocuteur à attendre la fin de la saisie de la phrase à synthétiser. Un des défi que pose ce paradigme est la synthèse d'un mot ou d'un groupe de mot avec une qualité segmentale et prosodique acceptable alors que la phrase qui le contient n'est que partiellement connue au moment de la synthèse. Pour ce faire, nous proposons différentes adaptations des deux principaux modules d'un système de synthèse de parole à partir du texte : le module de traitement automatique de la langue naturelle (TAL) et le module de synthèse sonore. Pour le TAL en synthèse incrémentale, nous nous sommes intéressé à l'analyse morpho-syntaxique, qui est une étape décisive pour la phonétisation et la détermination de la prosodie cible. Nous décrivons un algorithme d'analyse morpho-syntaxique dit "à latence adaptative". Ce dernier estime en ligne si une classe lexicale (estimée à l'aide d'un analyseur morpho-syntaxique standard basé sur l'approche n-gram), est susceptible de changer après l'ajout par l'utilisateur d'un ou plusieurs mots. Si la classe est jugée instable, alors la synthèse sonore est retardée, dans le cas contraire, elle peut s'effectuer sans risque a priori de dégrader de la qualité segmentale et suprasegmentale. Cet algorithme exploite une ensemble d'arbre de décisions binaires dont les paramètres sont estimés par apprentissage automatique sur un large corpus de texte. Cette méthode nous permet de réaliser un étiquetage morpho-syntaxique en contexte incrémental avec une précision de 92,5% pour une latence moyenne de 1,4 mots. Pour la synthèse sonore, nous nous plaçons dans le cadre de la synthèse paramétrique statistique, basée sur les modèles de Markov cachés (Hidden Markov Models, HMM). Nous proposons une méthode de construction de la voix de synthèse (estimation des paramètres de modèles HMM) prenant en compte une éventuelle incertitude sur la valeur de certains descripteurs contextuels qui ne peuvent pas être calculés en synthèse incrémentale (c'est-à-dire ceux qui portent sur les mots qui ne sont pas encore saisis au moment de la synthèse).Nous comparons la méthode proposée à deux autres stratégies décrites dans la littérature. Les résultats des évaluations objectives et perceptives montrent l’intérêt de la méthode proposée pour la langue française. Enfin, nous décrivons un prototype complet qui combine les deux méthodes proposées pour le TAL et la synthèse par HMM incrémentale. Une évaluation perceptive de la pertinence et de la qualité des groupes de mots synthétisés au fur et à mesure de la saisie montre que notre système réalise un compromis acceptable entre réactivité (minimisation du temps entre la saisie d'un mot et sa synthèse) et qualité (segmentale et prosodique) de la parole de synthèse. / In this thesis, we investigate a new paradigm for text-to-speech synthesis (TTS) allowing to deliver synthetic speech while the text is being inputted : incremental text-to-speech synthesis. Contrary to conventional TTS systems, that trigger the synthesis after a whole sentence has been typed down, incremental TTS devices deliver speech in a ``piece-meal'' fashion (i.e. word after word) while aiming at preserving the speech quality achievable by conventional TTS systems.By reducing the waiting time between two speech outputs while maintaining a good speech quality, such a system should improve the quality of the interaction for speech-impaired people using TTS devices to express themselves.The main challenge brought by incremental TTS is the synthesis of a word, or of a group of words, with the same segmental and supra-segmental quality as conventional TTS, but without knowing the end of the sentence to be synthesized. In this thesis, we propose to adapt the two main modules (natural language processing and speech synthesis) of a TTS system to the incremental paradigm.For the natural language processing module, we focused on part-of-speech tagging, which is a key step for phonetization and prosody generation. We propose an ``adaptive latency algorithm'' for part-of-speech tagging, that estimates if the inferred part-of-speech for a given word (based on the n-gram approach) is likely to change when adding one or several words. If the Part-of-speech is considered as likely to change, the synthesis of the word is delayed. In the other case, the word may be synthesized without risking to alter the segmental or supra-segmental quality of the synthetic speech. The proposed method is based on a set of binary decision trees trained over a large corpus of text. We achieve 92.5% precision for the incremental part-of-speech tagging task and a mean delay of 1.4 words.For the speech synthesis module, in the context of HMM-based speech synthesis, we propose a training method that takes into account the uncertainty about contextual features that cannot be computed at synthesis time (namely, contextual features related to the following words). We compare the proposed method to other strategies (baselines) described in the literature. Objective and subjective evaluation show that the proposed method outperforms the baselines for French.Finally, we describe a prototype developed during this thesis implementing the proposed solution for incremental part-of-speech tagging and speech synthesis. A perceptive evaluation of the word grouping derived from the proposed adaptive latency algorithm as well as the segmental quality of the synthetic speech tends to show that our system reaches a good trade-off between reactivity (minimizing the waiting time between the input and the synthesis of a word) and speech quality (both at segmental and supra-segmental levels).
|
180 |
Prosodic Features in Child-directed Speech during the Child's First YearRask, Linnea January 2015 (has links)
This study investigates prosodic features of child-directed speech during the child’s first year, using the automated prosodic annotation software Prosogram. From previous studies on first language acquisition and child-directed speech we know that speech directed to infants and small children is characterised by exaggerated use of several prosodic features, including a higher pitch, livelier pitch movement and slower speech rate. Annotation of these phenomena has previously been done manually, which is time consuming and includes a risk of circularity. If we can use semi-automated systems to carry out this task, it would be a huge methodological gain. This study analysed recordings of 10 parent-child pairs at four occasions (3, 6, 9 and 12 months of age) for a total of 40 recordings. The audio files were analysed in Prosogram in order to detect possible differences depending on the child’s age. The results showed a noticeable change in child-directed speech over the first year of the child’s life. A change in several characteristic prosodic features was noted to occur between the ages of 6 and 9 months. Pitch levels decreased, and articulation rate increased. Additionally, parents seemed to use pitch values much higher than their mean pitch speaking to children aged 3 and 6 months than to children aged 9 and 12 months. Despite using a relatively small sample, the results show several interesting trends in the usage of child-directed speech. Furthermore, this study shows that Prosogram is a useful tool for automatic analysis of child-directed speech.
|
Page generated in 0.1367 seconds