Submitted by Aline Batista (alinehb.ufpel@gmail.com) on 2016-06-29T19:34:38Z No. of bitstreams: 2 license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) A dinâmica dos gestos articulatórios da líquida lateral palatal.pdf: 6446781 bytes, checksum: 47d086436a73ebb11f6b2217fdbf858d (MD5) / Approved for entry into archive by Aline Batista (alinehb.ufpel@gmail.com) on 2016-06-30T20:24:48Z (GMT) No. of bitstreams: 2 A dinâmica dos gestos articulatórios da líquida lateral palatal.pdf: 6446781 bytes, checksum: 47d086436a73ebb11f6b2217fdbf858d (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Made available in DSpace on 2016-06-30T20:25:05Z (GMT). No. of bitstreams: 2 A dinâmica dos gestos articulatórios da líquida lateral palatal.pdf: 6446781 bytes, checksum: 47d086436a73ebb11f6b2217fdbf858d (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Previous issue date: 2016-06-06 / Sem bolsa / Devido as suas características linguísticas, articulatórias e acústicas, as laterais são concebidas como sons complexos. Promovem curiosidades, dúvidas e controvérsias (LADEFOGED, 1986; SILVA, 1996 e ALBANO, 2001) no meio acadêmico, mas são pouco investigadas. De fato, descrições da consoante líquida lateral palatal, por exemplo, ainda são escassas no Português Brasileiro (PB), sobretudo em populações atípicas. A presente pesquisa tem por objetivo a investigação da produção dessa consoante por usuários de implante coclear e por adultos falantes do português brasileiro como língua materna. As análises foram guiadas pela Fonologia Gestual, que concebe a fonologia das línguas como um sistema dinâmico. Foram analisadas as produções orais de 10 sujeitos adultos com idades entre 20 e 25 anos – 5 do gênero masculino e 5 do gênero feminino – e de uma usuária de implante coclear, de 13 anos. Os dados foram gravados por meio de um gravador digital, modelo Zoom H4N sincronizado a um ultrassom, modelo Mindray DP-6600. As coletas ocorreram em duas etapas. Na primeira, imagens foram apresentadas na tela de um computador, as quais deveriam ser nomeadas e inseridas na frase veículo Digo (palavra) bem bonito. Na segunda, o mesmo procedimento foi realizado com logatomas. A totalidade do corpus é de 3.420 itens. Os dados acústicos foram analisados por meio do software Praat, versão 5.3.77 (BOERSMA & WEENINK, 2007), e os dados articulatórios por meio do Software AAA (WRENCH, 2012). Os dados foram submetidos, ainda, à análise estatística através do software SPSS Statistics. Os resultados foram ao encontro dos obtidos por Silva (1996): Identificou-se as três fases acústico-articulatórias da consoante [?] quanto à trajetória dos formantes, bem como sua palatalização. Tais fatos acústicos demonstraram possuir correlatos com as imagens ultrassonográficas obtidas. A informante usuária de implante coclear realizou a consoante mais lentamente do que os informantes ouvintes, mas os aspectos acústicos de suas produções são semelhantes aos desses informantes. Entretanto, ao analisar os dados articulatórios, detectou-se uma variabilidade na palatalização da consoante, fato não detectado para os ouvintes. / Due to their linguistic, articulatory and acoustic characteristics, the lateral consonants are conceived as complex sounds. They promote curiosities, doubts and controversies (Ladefoged, 1986; SILVA, 1996 and ALBANO, 2001) in the academia, but it is scarcely investigated. In fact, descriptions of the palatal lateral approximant consonant, for example, are still rare in Brazilian Portuguese (BP), especially in atypical populations. This research aims to investigate the production of this consonant by people who received cochlear implant and adult speakers of Brazilian Portuguese as their mother tongue. The analysis was guided by the Gestual Phonology that conceives the phonology of the languages as a dynamic system. Ten adult subjects with ages between 20 and 25 years - 5 males and 5 females - and one female user of cochlear implant with the age 13 years had their oral production analyzed. The data were recorded by a digital recorder Zoom H4N model synchronized to an ultrasound, Mindray DP-6600 model. The samples were collected in two stages. First, images were displayed on a computer screen, whose names should be inserted in the gap of the vehicle phrase Digo (word) bem bonito – I say (…) pretty well. In the second, the same procedure was performed with logatomas. The totality of the corpus is of 3,420 items. The acoustic data were analyzed by Praat software version 3.5.77 (BOERSMA & Weenink, 2007), and for the articulatory data was used AAA Software (WRENCH, 2012). The data were also submitted to statistical analysis using SPSS software. The results agree with those obtained by Silva (1996), it means: the three acoustic-articulatory phases of the consonant [x] were identified regarding the trajectory of the formants, as well as palatalization. Such acoustic facts demonstrate correlations with the obtained ultrasound images. The informant with cochlear implant articulated the consonant more slowly than listeners, but the acoustic aspects of her productions are similar to those informants. However, when we analyze the articulatory data it was detected a variability in palatalization of the consonant, a fact that was not detected for the listeners.

L'objectif général de cette thèse est d'explorer l'impact des lésions du cervelet sur les séquelles motrices et cognitives des enfants traités pour tumeur du cervelet maligne ou bénigne. En nous basant sur trois études cliniques, nos objectifs généraux ont été (1) d'examiner si ces enfant présentaient des déficits dans l'établissement d'automatismes moteurs et cognitifs, (2) de préciser les facteurs associés aux difficultés d'automatisation et (3) d'examiner l'impact de ces difficultés sur la scolarité et le devenir à long terme de ces enfants. La première étude avait pour objectif spécifique d'examiner le devenir à long terme de 46 adultes et de 18 adolescents traités dans l'enfance chirurgicalement pour un astrocytome pilocytique du cervelet. Elle a mis en évidence un devenir à long terme satisfaisant dans l'ensemble, bien que des difficultés cognitives et motrices aient été rapportées, en particulier par les sujets qui ont le plus réussi leurs parcours scolaires. La perte d'autonomie était associée à des complications postopératoires telles que le mutisme cérébelleux et l'invasion du tronc cérébral. La deuxième étude a porté sur une cohorte de 17 enfants et adolescents traités pour un astrocytome pilocytique cérébelleux. Elle a exploré l'automatisation de la lecture et discuté la théorie cérébelleuse de la dyslexie. Les résultats ont mis en évidence une difficulté de suppression des mouvements articulatoires chez la plupart des sujets qui était associée à un faible indice de Mémoire de Travail Verbale. La troisième étude a porté sur 16 enfants traités pour un astrocytome pilocytique du cervelet et 16 enfants traités pour un médulloblastome. Elle avait pour objectifs (1) d'étudier l'apprentissage moteur et différents automatismes cognitifs intervenant notamment dans la lecture et le calcul mental et (2) de préciser les relations entre les différentes mesures de la difficulté d'automatisation motrice et cognitive. Les résultats ont confirmé la difficulté de suppression des mouvements articulatoires et ont montré que les enfants traités pour tumeur du cervelet se différenciaient des enfants sains de âge par un apprentissage moteur moindre, en particulier avec la main non dominante et par une lenteur en lecture, en calcul mental, en dénomination rapide et en double tâche. Par ailleurs, l'atteinte des noyaux dentelés était liée à une baisse de l'efficience intellectuelle, en particulier chez les enfants traités pour un médulloblastome, à un apprentissage moteur moindre avec la main dominante, à une difficulté de suppression des mouvements articulatoires, et à une lenteur de calcul mental. Ce travail de thèse offre des résultats pionniers dans la compréhension de l'impact des lésions cérébelleuses sur les apprentissages chez l'enfant. / The general aim this doctoral dissertation is to explore the impact of cerebellar lesions on motor and cognitive sequelae in children treated for malignant or benign tumor of the cerebellum. In three clinical studies, we in (1) examine whether these children exhibited deficits in motor and cognitive automation, (2) identify factors associated with difficulties in automation and (3) examine the impact of these difficulties on schooling and long-term outcome. The first study examines the long-term outcome of 46 adults and 18 adolescents treated surgically in childhood for cerebellar pilocytic astrocytoma. Most subjects showed a positive long-term outcome, despite report of cognitive and motor difficulties, especially by subjects with successful school careers. The loss of autonomy was associated with postoperative complications, post cerebellar mutism, or invasion of the brain stem. The second study in 17 children and adolescents treated for pilocytic cerebellar astrocytoma aimed to examine the automation of reading and to discuss the cerebellar theory of dyslexia. The results highlighted a difficulty of suppressing articulatory movements in most subjects, associated with low index of Verbal Working Memory. The third study concerned 16 children treated for pilocytic astrocytoma of the cerebellum and 16 children treated for medulloblastoma. Its objectives were to (1) investigate motor learning and different cognitive automations involved in particularly in reading and mental calculation (2) clarify the relationship between different measures of motor and cognitive automation. The results confirmed the difficulty to suppress articulatory movements and showed lower motor learning effect, especially with the non-dominant hand, slowness in reading, mental calculation, rapid naming and dual task. Furthermore, dentate nuclei damage was linked to lower intellectual efficiency (IQ), particularly in children treated for medulloblastoma; to a lesser motor learning for the dominant hand, a difficulty to suppress articulatory movements, and slowness in mental calculation. This work offers pioneer results in understanding the impact of cerebellar lesions in children learning.

Articulatory copy synthesis (ACS), a subarea of speech inversion, refers to the reproduction of natural utterances and involves both the physiological articulatory processes and their corresponding acoustic results. This thesis proposes two novel methods for the ACS of human speech using the articulatory speech synthesizer VocalTractLab (VTL) to address or mitigate the existing problems of speech inversion, such as non-unique mapping, acoustic variation among different speakers, and the time-consuming nature of the process. The first method involved finding appropriate VTL gestural scores for given natural utterances using a genetic algorithm. It consisted of two steps: gestural score initialization and optimization. In the first step, gestural scores were initialized using the given acoustic signals with speech recognition, grapheme-to-phoneme (G2P), and a VTL rule-based method for converting phoneme sequences to gestural scores. In the second step, the initial gestural scores were optimized by a genetic algorithm via an analysis-by-synthesis (ABS) procedure that sought to minimize the cosine distance between the acoustic features of the synthetic and natural utterances. The articulatory parameters were also regularized during the optimization process to restrict them to reasonable values. The second method was based on long short-term memory (LSTM) and convolutional neural networks, which were responsible for capturing the temporal dependence and the spatial structure of the acoustic features, respectively. The neural network regression models were trained, which used acoustic features as inputs and produced articulatory trajectories as outputs. In addition, to cover as much of the articulatory and acoustic space as possible, the training samples were augmented by manipulating the phonation type, speaking effort, and the vocal tract length of the synthetic utterances. Furthermore, two regularization methods were proposed: one based on the smoothness loss of articulatory trajectories and another based on the acoustic loss between original and predicted acoustic features. The best-performing genetic algorithms and convolutional LSTM systems (evaluated in terms of the difference between the estimated and reference VTL articulatory parameters) obtained average correlation coefficients of 0.985 and 0.983 for speaker-dependent utterances, respectively, and their reproduced speech achieved recognition accuracies of 86.25% and 64.69% for speaker-independent utterances of German words, respectively. When applied to German sentence utterances, as well as English and Mandarin Chinese word utterances, the neural network based ACS systems achieved recognition accuracies of 73.88%, 52.92%, and 52.41%, respectively. The results showed that both of these methods not only reproduced the articulatory processes but also reproduced the acoustic signals of reference utterances. Moreover, the regularization methods led to more physiologically plausible articulatory processes and made the estimated articulatory trajectories be more articulatorily preferred by VTL, thus reproducing more natural and intelligible speech. This study also found that the convolutional layers, when used in conjunction with batch normalization layers, automatically learned more distinctive features from log power spectrograms. Furthermore, the neural network based ACS systems trained using German data could be generalized to the utterances of other languages.

No description available.

La présente étude s’intéresse aux mécanismes neuronaux qui sous-tendent la rétention en mémoire à court terme auditive (MCTA) en utilisant la technique des potentiels reliés aux événements (PRE). Dans l’Expérience 1, nous avons isolé une composante de PRE, nommée SAN pour « sustained anterior negativity ». La SAN augmentait en amplitude négative plus le nombre de sons à maintenir en MCTA augmentait. Cet effet de charge était présent, bien que la durée totale des stimuli restait la même entre les conditions. L’effet de charge observé par la SAN dans l’Expérience 1 disparaissait dans l’Expérience 2, où les mêmes sons étaient utilisés, mais où la mémorisation de ceux-ci n’était plus requise. Finalement, dans l’Expérience 3, la tâche de MCTA a été effectuée avec et sans suppression articulatoire durant l'intervalle de rétention. L’effet de charge trouvé dans l’Expérience 1 était de nouveau observé, lorsque les participants faisaient la tâche de suppression articulatoire ou non. Ces résultats suggèrent que la SAN reflète l'activité nécessaire pour le maintien des objets acoustiques dans un système de MCTA qui serait distinct de la répétition phonologique. / We studied the neuronal mechanisms that implement acoustic short-term memory (ASTM) for pitch using event-related potentials (ERP). Experiment 1 isolated an ERP component, the sustained anterior negativity (SAN), that increased in amplitude with increasing memory load in ASTM using stimuli with equal duration at all memory loads. The SAN load effect found in Experment 1, when pitch had to be remembered to perform the task, was absent in Experiment 2 using the same sounds when memory was not required. In Experiment 3 the memory task was performed without or with concurrent articulatory suppression during the retention interval, to suppress rehearsal via an articulatory loop. Load-related effects found in Experiment 1 were found again, whether participants engaged in concurrent suppression or not. The results suggest that the SAN reflects activity required to maintain pitch objects in an ASTM system that is distinct from articulatory rehearsal.

Les résultats obtenus en psycholinguistique via des études perceptives sur le traitement de la consonne de liaison montrent que le traitement de ce phénomène complexe n’est qu’imparfaitement prédit par les différentes théories phonologiques dans lesquelles il a été intégré de manière explicite. C’est pourquoi nous proposons dans cette thèse, qui s’inscrit dans une perspective résolument nouvelle, d’envisager la liaison du point de vue de la production et de la formaliser dans le cadre de la Phonologie articulatoire de Browman et Goldstein (1986). Ce modèle présente l’avantage d’une organisation directe entre phonologie et phonétique en utilisant une seule unité qui sert à la fois de primitive phonologique et d’unité d’action motrice dans la production de la parole : le geste articulatoire. Nous supposons que le statut phonologique de « segment flottant » proposé pour la consonne de liaison dans le cadre de la phonologie autosegmentale (Clements & Keyser, 1985 ; Encrevé 1988) puisse trouver un corrélat au niveau articulatoire et puisse être observé et mesuré expérimentalement. L’analyse qui vient à l’appui de cette hypothèse compare la consonne de liaison avec la consonne initiale de mot dans des contextes vocaliques identiques au moyen de mesures électro-palatographiques. Nos résultats suggèrent que la consonne de liaison est soumise à une modification quantitative des caractéristiques dynamiques qui lui sont associés. Les mouvements des articulateurs ainsi que les relations entre les gestes consonantique et vocalique sont ré-échelonnés en fonction de la position qu’occupe la consonne de manière graduelle. / Perceptual studies in the field of psycholinguistics on the processing of the liaison consonant have shown that this complex phenomenon is not perfectly predicted by the different phonological theories in which it has been explicitly integrated. In this thesis, we propose to investigate liaison from a totally new perspective, that is from a production point of view, and to formalise this phenomenon in the framework of Browman and Goldstein (1986)’s Articulatory Phonology. This model is advantageous as it provides a direct connection between phonology and phonetics where only one unit is used both as a phonological primitive and as a unit of motor action in speech production, that is articulatory gesture. In the Framework of autosegmental phonology (Clements & Keyser, 1985 ; Encrevé 1988), the phonological status of the liaison consonant is considered to be a floating segment. In this thesis, we hypothesise that there is a correlate at the articulatory level that can be observed and experimentally measured. This hypothesis is tested by comparing the liaison consonant with the word-initial consonant in identical vowel contexts by using electro-palatographical measurements.Results suggest that the liaison consonant is subjected to a quantitative modification of the dynamic caracteristics which are associated with it. The articulator movements, as well as the relations between consonant and vowel gestures, are gradually re-scaled according to the consonant position.

La présente étude s’intéresse aux mécanismes neuronaux qui sous-tendent la rétention en mémoire à court terme auditive (MCTA) en utilisant la technique des potentiels reliés aux événements (PRE). Dans l’Expérience 1, nous avons isolé une composante de PRE, nommée SAN pour « sustained anterior negativity ». La SAN augmentait en amplitude négative plus le nombre de sons à maintenir en MCTA augmentait. Cet effet de charge était présent, bien que la durée totale des stimuli restait la même entre les conditions. L’effet de charge observé par la SAN dans l’Expérience 1 disparaissait dans l’Expérience 2, où les mêmes sons étaient utilisés, mais où la mémorisation de ceux-ci n’était plus requise. Finalement, dans l’Expérience 3, la tâche de MCTA a été effectuée avec et sans suppression articulatoire durant l'intervalle de rétention. L’effet de charge trouvé dans l’Expérience 1 était de nouveau observé, lorsque les participants faisaient la tâche de suppression articulatoire ou non. Ces résultats suggèrent que la SAN reflète l'activité nécessaire pour le maintien des objets acoustiques dans un système de MCTA qui serait distinct de la répétition phonologique. / We studied the neuronal mechanisms that implement acoustic short-term memory (ASTM) for pitch using event-related potentials (ERP). Experiment 1 isolated an ERP component, the sustained anterior negativity (SAN), that increased in amplitude with increasing memory load in ASTM using stimuli with equal duration at all memory loads. The SAN load effect found in Experment 1, when pitch had to be remembered to perform the task, was absent in Experiment 2 using the same sounds when memory was not required. In Experiment 3 the memory task was performed without or with concurrent articulatory suppression during the retention interval, to suppress rehearsal via an articulatory loop. Load-related effects found in Experiment 1 were found again, whether participants engaged in concurrent suppression or not. The results suggest that the SAN reflects activity required to maintain pitch objects in an ASTM system that is distinct from articulatory rehearsal.

Le présent travail de thèse s’inscrit dans le domaine de recherche sur la perception audio-visuelle (AV) de la parole. Son objectif est de répondre aux questions suivantes : (i) Quelle est la nature du traitement de l’input visuel (holistique vs analytique) dans la perception AV de la parole ? (ii) Quelle est l’implication des régions faciales extra-buccales dans la perception AV de la parole ? (iii) Quel est le comportement oculaire des sujets lors de la perception AV de la parole ? (iv) Quelle est l’évolution de la perception AV de la parole dans les aspects (i), (ii) et (iii) au cours du développement. Le paradigme de dégradation de l’information auditive par le bruit a été utilisé dans deux expériences qui ont été menées avec des participants de quatre groupes d’âge, enfants, préadolescents, adolescents, adultes (16 participants par groupe). La tâche des participants consistait à répéter les syllabes de type consonne-voyelle /a/, faiblement et fortement dégradées par le bruit rose, présentées dans quatre conditions différentes. Ces conditions étaient les suivantes : une auditive (AU) et trois audio-visuelles (AV) (AV visage (AVV)), AV « bouche extraction » (AVB-E) (format bouche sans contrastes lumineux), AV « bouche-masquage » (AVB-M) (format bouche avec contrastes lumineux) pour l’expérience 1, et AVV, AV « bouche active » (AVV-BA) (format « visage » avec un contexte facial statique), AV « régions extra-buccales actives » (AVV-EBA) (format « visage » sans bouche) pour l’expérience 2. Le nombre total des répétitions correctes par condition (performance totale), la différence dans ce score entre chaque condition AV et la condition auditive (gain AV) et la durée totale des fixations oculaires dans la région buccale et les autres régions faciales (pour les formats AVV) ont été analysés. Les principaux résultats montrent que les mécanismes de traitement AV de la parole atteignent leur maturité avant l’enfance tardive. La vision du visage entier de l’orateur n’est pas avantageuse pour ce type de traitement. Elle semble même désavantageuse pour les adultes possiblement car elle déclenche le traitement d’autres aspects du visage (identité, expressions faciales) qui pourrait interférer avec le traitement des indices acoustiques relatifs à la parole. Pour les quatre groupes d’âge, la contribution des mouvements articulatoires dans les régions extra-buccales à la perception AV de la parole s’est avérée faible et limitée aux conditions de haute incertitude quant à l’information auditive. Pour les stimuli respectant les caractéristiques écologiques de l’information faciale, les patterns du comportement oculaire dans la perception bimodale de la parole varient en fonction du degré de dégradation de l’information auditive, mais semblent relativement stables durant la période allant de l’enfance à l’âge adulte. Finalement, les modalités de présentation de l’information visuelle localisée à la bouche ont affecté le comportement oculaire chez les adultes, les pré-adolescents et les enfants. Ceci suggère que le traitement visuo-attentionnel dans le cadre de la perception AV de la parle est sensible aux caractéristiques de bas niveau des stimuli visuels chez ces populations. Les variations au niveau du traitement visuo-attentionnel s’accompagnent, dans une certaine mesure, de variations dans la perception AV de la parole. / The present thesis work fits into the domain/is incorporated within the framework of research on audio-visual (AV) speech perception. Its objective is to answer the following questions: (i) What is the nature of visual input processing (holistic vs analytic) in AV speech perception? (ii) What is the implication of extra-oral facial movement in AV speech perception? (iii) What are the oculomotor patterns in AV speech perception? (iv) What are the developmental changes in the above-mentioned aspects (i), (ii) and (iii)? The classic noise degradation paradigm was applied in two experiments conducted in the framework of the present thesis. Each experiment were conducted on participants of 4 age groups, adults, adolescents, pre-adolescents and children. Each group consisted of 16 participants. Participants’ task was to repeat consonant-vowel (/a/) syllables. The syllables were both mildly and strongly degraded by pink noise and were presented in four audio(-visual) conditions, one purely auditory (AO) and three audio-visual conditions. The AV conditions were the following: (i) AV face (AVF), (ii) AV « mouth extraction » (AVM-E ; mouth format without visual contrasts), (iii) AV « mouth window » (AVM-W ; mouth format with high visual contrasts) in experiment 1, and (i) AVF, (ii) AVF « mouth active (and facial frame static) » (AVF-MA), (iii) AVF « extra-oral regions active (and mouth absent) » (AVF-EOA) in experiment 2. The data relative to (i) the total number of correct repetitions (total performance), (ii) the difference in the correct repetitions score between each AV and the AO condition (AV gain), and (iii) the total fixations duration in the oral area and other facial areas (for the AV formats) were analyzed. The main results showed that the mechanisms involved in AV speech perception reach their maturity before late childhood. The vision of the talker’s full face does not seem to be advantageous in this context. It seems that the vision of the talker’s full face might perturb AV speech processing in adults, possibly because it triggers processing of other types of information (identity, facial expressions) which could in terms interfere with the processing of acoustic aspects of speech. The contribution of the extra-oral articulatory movement to AV speech perception was poor and limited to the condition of highly degraded auditory information. For ecologically presented facial information, the oculomotor patterns in AV speech perception varied as a function of the level of auditory information degradation, but appeared rather stable across the 4 groups. Finally, the modalities of the featural (mouth) facial information presentation affected the oculomotor behavior patterns in adults, pre-adolescents and children, thus suggesting a certain sensitivity of visuo-attentional processing to low-level visual stimuli characteristics in AV speech perception. The variations in visuo-attentional processing seemed to be associated to a certain extent with variations in AV speech perception.

La capacité de production de la parole est apprise et maintenue au moyen d'une boucle de perception-action qui permet aux locuteurs de corriger leur propre production en fonction du retour perceptif reçu. Ce retour est auditif et proprioceptif, mais pas visuel. Ainsi, les sons de parole peuvent être complétés par l'affichage des articulateurs sur l'écran de l'ordinateur, y compris ceux qui sont habituellement cachés tels que la langue ou le voile du palais, ce qui constitue de la parole augmentée. Ce type de système a des applications dans des domaines tels que l'orthophonie, la correction phonétique et l'acquisition du langage. Ce travail a été mené dans le cadre du développement d'un système de retour articulatoire visuel, basé sur la morphologie et les stratégies articulatoires d'un locuteur de référence, qui anime automatiquement une tête parlante 3D à partir du son de la parole. La motivation de cette recherche était d'adapter ce système à plusieurs locuteurs. Ainsi, le double objectif de cette thèse était d'acquérir des connaissances sur la variabilité inter-locuteur, et de proposer des modèles pour adapter un clone de référence, composé de modèles des articulateurs de la parole (lèvres, langue, voile du palais, etc.), à d'autres locuteurs qui peuvent avoir des morphologies et des stratégies articulatoires différentes. Afin de construire des modèles articulatoires pour différents contours du conduit vocal, nous avons d'abord acquis des données qui couvrent l'espace articulatoire dans la langue française. Des Images médio-sagittales obtenues par Résonance Magnétique (IRM) pour onze locuteurs francophones prononçant 63 articulations ont été recueillis. L'un des principaux apports de cette étude est une base de données plus détaillée et plus grande que celles disponibles dans la littérature. Cette base contient, pour plusieurs locuteurs, les tracés de tous les articulateurs du conduit vocal, pour les voyelles et les consonnes, alors que les études précédentes dans la littérature sont principalement basées sur les voyelles. Les contours du conduit vocal visibles dans l'IRM ont été tracés à la main en suivant le même protocole pour tous les locuteurs. Afin d'acquérir de la connaissance sur la variabilité inter-locuteur, nous avons caractérisé nos locuteurs en termes des stratégies articulatoires des différents articulateurs tels que la langue, les lèvres et le voile du palais. Nous avons constaté que chaque locuteur a sa propre stratégie pour produire des sons qui sont considérées comme équivalents du point de vue de la communication parlée. La variabilité de la langue, des lèvres et du voile du palais a été décomposé en une série de mouvements principaux par moyen d'une analyse en composantes principales (ACP). Nous avons remarqué que ces mouvements sont effectués dans des proportions différentes en fonction du locuteur. Par exemple, pour un déplacement donné de la mâchoire, la langue peut globalement se déplacer dans une proportion qui dépend du locuteur. Nous avons également remarqué que la protrusion, l'ouverture des lèvres, l'influence du mouvement de la mâchoire sur les lèvres, et la stratégie articulatoire du voile du palais peuvent également varier en fonction du locuteur. Par exemple, certains locuteurs replient le voile du palais contre la langue pour produire la consonne /ʁ/. Ces résultats constituent également une contribution importante à la connaissance de la variabilité inter-locuteur dans la production de la parole. Afin d'extraire un ensemble de patrons articulatoires communs à différents locuteurs dans la production de la parole (normalisation), nous avons basé notre approche sur des modèles linéaires construits à partir de données articulatoires. Des méthodes de décomposition linéaire multiple ont été appliquées aux contours de la langue, des lèvres et du voile du palais ... / The capacity of producing speech is learned and maintained by means of a perception-action loop that allows speakers to correct their own production as a function of the perceptive feedback received. This auto feedback is auditory and proprioceptive, but not visual. Thus, speech sounds may be complemented by augmented speech systems, i.e. speech accompanied by the virtual display of speech articulators shapes on a computer screen, including those that are typically hidden such as tongue or velum. This kind of system has applications in domains such as speech therapy, phonetic correction or language acquisition in the framework of Computer Aided Pronunciation Training (CAPT). This work has been conducted in the frame of development of a visual articulatory feedback system, based on the morphology and articulatory strategies of a reference speaker, which automatically animates a 3D talking head from the speech sound. The motivation of this research was to make this system suitable for several speakers. Thus, the twofold objective of this thesis work was to acquire knowledge about inter-speaker variability, and to propose vocal tract models to adapt a reference clone, composed of models of speech articulator's contours (lips, tongue, velum, etc), to other speakers that may have different morphologies and different articulatory strategies. In order to build articulatory models of various vocal tract contours, we have first acquired data that cover the whole articulatory space in the French language. Midsagittal Magnetic Resonance Images (MRI) of eleven French speakers, pronouncing 63 articulations, have been collected. One of the main contributions of this study is a more detailed and larger database compared to the studies in the literature, containing information of several vocal tract contours, speakers and consonants, whereas previous studies in the literature are mostly based on vowels. The vocal tract contours visible in the MRI were outlined by hand following the same protocol for all speakers. In order to acquire knowledge about inter-speaker variability, we have characterised our speakers in terms of the articulatory strategies of various vocal tract contours like: tongue, lips and velum. We observed that each speaker has his/her own strategy to achieve sounds that are considered equivalent, among different speakers, for speech communication purposes. By means of principal component analysis (PCA), the variability of the tongue, lips and velum contours was decomposed in a set of principal movements. We noticed that these movements are performed in different proportions depending on the speaker. For instance, for a given displacement of the jaw, the tongue may globally move in a proportion that depends on the speaker. We also noticed that lip protrusion, lip opening, the influence of the jaw movement on the lips, and the velum's articulatory strategy can also vary according to the speaker. For example, some speakers roll up their uvulas against the tongue to produce the consonant /ʁ/ in vocalic contexts. These findings also constitute an important contribution to the knowledge of inter-speaker variability in speech production. In order to extract a set of common articulatory patterns that different speakers employ when producing speech sounds (normalisation), we have based our approach on linear models built from articulatory data. Multilinear decomposition methods have been applied to the contours of the tongue, lips and velum. The evaluation of our models was based in two criteria: the variance explanation and the Root Mean Square Error (RMSE) between the original and recovered articulatory coordinates. Models were also assessed using a leave-one-out cross validation procedure ...

This dissertation focuses on the contributions of Phonetics in the teaching and learning processes of languages pronunciation in Singing. For that, it investigates a teaching proposal based on this discipline of Linguistics, the Phonetics, with emphasis on articulation and representation of speech sounds - the Articulatory Phonetics and the International Phonetic Alphabet (IPA) in practice, the phonetic transcription. It is based on the literature on Phonetics and Phonology, in the Linguistics area, on the Teaching of Languages, in the Applied Linguistics, and on the Diction for Singers, in the Arts - specifically, the field of Music, singing. Based on the diagnosis made on the subject of Diction in technical and undergraduate courses of Singing in institutions in the triangle region of Minas Gerais, it reflects and suggests changes on didactic and pedagogical aspects, related to the training of singers and singing teachers. It intends to contribute with theoretical reflections, teaching materials and methodological approaches for the teaching of pronunciation of the major languages in the classical repertoire performed in Brazil. / Esta dissertação centra-se nas contribuições da Fonética no processo ensino e aprendizagem da pronúncia de línguas no Canto. Para isto, investiga uma proposta de ensino elaborada com base nesta disciplina da Linguística, a Fonética, com ênfase na articulação e representação dos sons da fala a Fonética Articulatória e o Alfabeto Fonético Internacional (AFI) em sua prática, a transcrição fonética. Fundamenta-se em literaturas sobre Fonética e Fonologia, da área de Linguística, sobre o Ensino de Línguas, da Linguística Aplicada e, sobre Dicção para Cantores, das Artes especificamente, da subárea Música, o instrumento Canto. Com base no diagnóstico da disciplina Dicção em cursos técnicos e graduação em Canto de algumas instituições na região do Triângulo Mineiro, reflete e sugere mudanças de aspectos didáticos e pedagógicos, em função da formação de cantores professores de Canto. Pretende contribuir com reflexões teóricas, material didático e abordagens para o ensino da pronúncia das principais línguas do repertório erudito praticado no Brasil. / Mestre em Artes