L’objectif de cette thèse est l’étude et la conception d’une plateforme de synthèse de parole expressive.Le système de synthèse — LIPS3, développé dans le cadre de ce travail, incorpore deux éléments : un module linguistique et un module de synthèse paramétrique par apprentissage statistique (construit à l’aide de HTS et de STRAIGHT). Le système s’appuie sur un corpus monolocuteur conçu, enregistréet étiqueté à cette occasion.Une première étude, sur l’influence de la qualité de l’étiquetage du corpus d’apprentissage, indique que la synthèse paramétrique statistique est robuste aux erreurs de labels et d’alignement. Cela répond au problème de la variation des réalisations phonétiques en parole expressive.Une seconde étude, sur l’analyse acoustico-phonétique du corpus permet la caractérisation de l’espace expressif utilisé par la locutrice pour réaliser les consignes expressives qui lui ont été fournies. Les paramètres de source et les paramètres articulatoires sont analysés suivant les classes phonétiques, ce qui permet une caractérisation fine des phonostyles.Une troisième étude porte sur l’intonation et le rythme. Calliphony 2.0 est une interface de contrôlechironomique temps-réel permettant la modification de paramètres prosodiques (f0 et tempo) des signaux de synthèse sans perte de qualité, via une manipulation directe de ces paramètres. Une étude sur la stylisation de l’intonation et du rythme par contrôle gestuel montre que cette interface permet l’amélioration, non-seulement de la qualité expressive de la parole synthétisée, mais aussi de la qualité globale perçue en comparaison avec la modélisation statistique de la prosodie.Ces études montrent que la synthèse paramétrique, combinée à une interface chironomique, offre une solution performante pour la synthèse de la parole expressive, ainsi qu’un outil d’expérimentation puissant pour l’étude de la prosodie. / The subject of this thesis was the study and conception of a platform for expressive speech synthesis.The LIPS3 Text-to-Speech system — developed in the context of this thesis — includes a linguistic module and a parametric statistical module (built upon HTS and STRAIGHT). The system was based on a new single-speaker corpus, designed, recorded and annotated.The first study analyzed the influence of the precision of the training corpus phonetic labeling on the synthesis quality. It showed that statistical parametric synthesis is robust to labeling and alignment errors. This addresses the issue of variation in phonetic realizations for expressive speech.The second study presents an acoustico-phonetic analysis of the corpus, characterizing the expressive space used by the speaker to instantiate the instructions that described the different expressive conditions. Voice source parameters and articulatory settings were analyzed according to their phonetic classes, which allowed for a fine phonostylistic characterization.The third study focused on intonation and rhythm. Calliphony 2.0 is a real-time chironomic interface that controls the f0 and rhythmic parameters of prosody, using drawing/writing hand gestures with a stylus and a graphic tablet. These hand-controlled modulations are used to enhance the TTS output, producing speech that is more realistic, without degradation as it is directly applied to the vocoder parameters. Intonation and rhythm stylization using this interface brings significant improvement to the prototypicality of expressivity, as well as to the general quality of synthetic speech.These studies show that parametric statistical synthesis, combined with a chironomic interface, offers an efficient solution for expressive speech synthesis, as well as a powerful tool for the study of prosody.

No description available.

No description available.

Die vorliegende Arbeit befasst sich mit der Fragestellung, wie die Wortbetonung in der kanonischen Lallenphase organisiert ist. Diese Entwicklungsphase tritt für gewöhnlich in der zweiten Hälfte des ersten Lebensjahres auf, wenn Kleinkinder erstmalig phonetisch wohlgeformte Silben äußern. Dabei gibt es jedoch noch keine klare Beziehung zwischen Wortbedeutung und bestimmten Sprachlauten. Die Einzelsilben werden allerdings schon zu größeren wortartigen Einheiten zusammengefügt und erscheinen in betonten wie unbetonten Positionen. Es wurden Sprachdaten von sieben deutschlernenden Kindern im Alter zwischen 0;5 und 1;0 gesammelt. Die phonetisch wohlgeformten Lalläußerungen wurden einer Reihe von perzeptiven und akustischen Analysen unterzogen, um der Frage nach einer bereits vorhandenen prosodischen Organisation der Wortbetonung nachzugehen. Die Äußerungen der Probanden werden anschließend mit einem theoretischen Parametermodell verglichen. Hierbei wird untersucht, ob Lalläußerungen eher sprachliche Universalien vorweisen oder doch eher durch ihre sprachliche Umgebung beinflusst sind. Dazu wird eigens eine Reanalyse des Betonungssystems des Deutschen vorgenommen. Die Ergebnisse zeigen, dass eine leichte Tendenz hin zum trochäischen Betonungsmuster in Zweisilbern zu erkennen ist. Diese Tendenz lässt sich allerdings nur mit zwei der fünf Analysemethoden der vorliegenden Arbeit nachweisen. Bei Mehrsilbern scheint die Lage der hauptbetonten Silbe durch keine Analysemethode vorhersagbar zu sein. Es kann dennoch festgestellt werden, dass Silben vom Typus CVV in signifikanter Weise Betonung auf sich ziehen. Diese Ergebnisse lassen den Schluss zu, dass die Parameter der Betonungszuweisung während des kanonischen Lallens noch nicht zur Gänze gesetzt sind. Nur diejenigen für Fußform und Quantitätssensitivität erscheinen so konfiguriert, dass ein Einfluß der Umgebungssprache, des Deutschen, erkennbar wird. / The current work addresses the question, how word stress is organised during the developmental stage of Canonical Babbling, which usually emerges in the second half of the first year of life. In this period infants begin to utter phonetically well-formed syllables, although there is no clear relationship between meaning and single speech sounds. However, syllables are combined to form larger word-like units containing prominent and less prominent syllables. Data was gathered from seven German-learning infants aged between 0;5 and 1;0. The phonetically well-formed babbling vocalizations were subject to perceptual and acoustical analyses, in order to find out whether there is already a prosodic organization of word stress. The subjects'' utterances are then compared to a theoretical parametric framework. It will be examined, whether babbling vocalizations are more universal or rather influenced by the ambient language. Therefore a reanalysis of the German system of word stress will be undertaken. Results indicate that there is a weak tendency towards the trochaic stress pattern in disyllabic babbles. This tendency is, however, only predictable in two of the five analysis parameters used in this study. In polysyllabic vocalizations, the location of main stress appears not to be predictable by any analysis. It can nevertheless be stated that CVV syllables tend to attract stress significantly. These results point to the fact, that the parameters of word stress assignment are not yet completely set during Canonical Babbling. Instead, only those for foot shape and quantity-sensitivity appear to be set in a fashion reflecting an influence of the subjects'' ambient language, German.

Die Alternativensemantik nach (Rooth, 1985, 1992) geht davon aus, dass sprachlicher Fokus eine Menge von Alternativen evoziert und Fokuspartikeln auf diese Menge Bezug nehmen. Die vorliegende Dissertation untersucht den Ein uss von Fokuspartikeln und prosodischem Fokus auf die mentale Repräsentation von Alternativenmengen. Im Speziellen wurden die sprachlichen und kognitiven Mechanismen, die am Aufbau einer Alternativenmenge beteiligt sind, mittels psycholinguistischer Experimente getestet. Experiment 1a und 1b untersuchen späte Repräsentation der gesamten Alternativenmenge. Die Experimente zeigen, dass Fokuspartikeln zu einer reicheren Enkodierung der Alternativenmenge führen, was sich in einem verbesserten Gedächtnisabruf äußert. Experiment 2 und 3 untersuchen die Prozesse, die am Aufbau von Alternativenmengen beteiligt sind. In Experiment 2 (Probe Recognition) führten Fokuspartikeln zu Interferenzeffekten beim Erkennen von erwähnten Alternativen sowie bei der Ablehnung nicht-erwähnter Alternativen. Eine weitere Analyse und Experiment 4 untersuchen die Mitglieder der Alternativenmenge. Die Daten deuten darauf hin, dass die Alternativenmenge aus sämtlichen möglichen Ersetzungen des fokussierten Elements inklusive unrelatierter Elemente besteht. Im Vergleich von intonatorischem Fokus und Fokuspartikeln (Experiment 5 und 6) zeigte sich, dass Fokus den Abruf von kontextuellen Alternativen erleichtert während Fokuspartikeln zu zusätzlichen Kompetitionseffekten führen. Im Allgemeinen liefert deutet die vorliegende Dissertation darauf hin, dass intonatorischer Fokus eine Alternativenmenge im Kopf des Hrers evoziert und dabei hilft relevante Alternativen zu identifizieren. Fokuspartikeln haben eine zusätzliche Funktion und führen whrend der Sprachverarbeitung zu einem strkeren Wettberwerb zwischen Mitgliedern der Alternativenmenge. / According to Rooth'' alternative semantics (Rooth, 1985, 1992), focus evokes a set of alternatives and that focus particles make reference to this set. The present thesis examines the impact of focus particles and intonational focus on the mental representation of alternatives. In particular, it explores the linguistic and cognitive mechanisms that underlie the establishment of alternative sets in a series of psycholinguistic experiments. Two delayed recall experiments (Exp. 1a and 1b) revealed that participants were better at recalling mentioned alternatives to a focused element when the discourse contained the particles only or even. Experiments 2 and 3 found interference effects of focus particles in the recognition of mentioned alternatives and the rejection of unmentioned alternatives. An additional analysis and Experiment 4 indicate that the set of alternatives consists of various possible replacements of the focused element. Experiment 5 showed that contrastive accents facilitated the retrieval of alternatives while focus particles caused interference effects relative to the condition with an L+H* accent. Overall, this thesis indicates that intonational focus activates alternatives in the listeners mind and helps identifying relevant alternatives. Focus particles, on the other hand, cause stronger competition between the focused element and its alternatives.

Made available in DSpace on 2016-04-28T18:23:14Z (GMT). No. of bitstreams: 1 IzabelCristinaViola.pdf: 5167096 bytes, checksum: 5ae574c694e79ab10fe91f2c4a099ac9 (MD5) Previous issue date: 2006-05-12 / This thesis elucidates the elements that build the relationship between sound and meaning in speech and contributes theoretically and methodologically to the analysis of the oral expression. The symbolism of sounds, the style and the expression of emotions and attitudes are discussed from the point of view of speech. The models of communication The Live Voice and Covariance and Configuration help interpret the expressive potential of the speech. The corpus is constituted by the poem I-Juca Pirama performed by a professional actor. The data analysis dwells upon its auditive-perceptive and acoustic-phonetic analysis together with the parallelism between phonic materialization and the meaning. The qualitative analysis deals with the study the qualities of voice, the allophones of /r/ and the loudness. The measurement of the duration of speech segments, in ms, was carried out in order to analyse speech and articulation rates, rhythm and continuity. Furthermore, fundamental frequency values, in Hz, of vowel segments in words and phrases were obtained to investigate intonation aspects. The confrontation of duration and f0 contours are used to interpret the data of the prosodic analysis. We favor the conception that the relationship between sound and meaning in speech is either motivated or arbitrary, used consciously or unconsciously, historically established or continuously changed within culture, by the demotivation or remotivation processes of the sign. The individual promotes in practical activities the indexes of the nature of gestures, known as vocal gestures, which are composed by the interaction (and impossible dissociation) of prosodic elements (quality and dynamics of voice) with the phonetic elements (vowels and consonants) and non-verbal sounds produced in the communication (breathing, mouth and tongue sounds). As body movements of one or more structures of the vocal tract that move through time and space, in synchronic or non-synchronic ways, the vocal gestures are used to reduce tension being or not the voluntary reproduction an emotion or mark the presence of an emotion. The vocal gestures, as symbolic elements, belong to the system of expressive signs that respond to a universal sound symbolism. They can represent other animate or inanimate objects that are either associated by similarity or by functional analogy, may have a specific configuration and/or are subject to the activation of the organism when express linguistic information or an emotion or an attitude. The vocal gestures, as stylistical facts, represent the dual character of the language. They are practical trademarks of the job of the locutor who shows his/her singularity and subjectivity in a recurrent and salient way in the speech, building and being built by meaning, one time playing the main role, other times a secondary role. The vocal gestures, as physiologic and dynamic linguistic elements carry out in the individual expression the subjective and contextual demands that reflect in the variability of language and, therefore, integrate the voice into the universe of the language / Esta tese tem como objetivo enfocar os elementos que constroem as relações entre som e sentido na fala, contribuindo teórica e metodologicamente para a análise da expressividade oral. O simbolismo sonoro, o estilo e a expressão de emoções e atitudes são discutidos do ponto de vista da fala. Os modelos de comunicação A Viva Voz e da Covariância e da Configuração apóiam a interpretação do potencial expressivo da fala. O corpus é constituído pelo poema de I-Juca Pirama , interpretado por um ator profissional, e a análise dos dados é realizada por uma combinação das análises perceptivo-auditiva e fonético-acústica e pelo pareamento entre materialidade fônica e o sentido. As análises qualitativas abarcam o estudo das qualidades de voz, dos alofones do /r/, da loudness e de seus usos expressivos. As análises quantitativas abrangem as medidas de duração em ms em segmentos, sílabas, GIPC, silêncios, ruídos inspiratórios e enunciados e medidas de freqüência fundamental (f0) das vogais em palavras e enunciados. São analisados aspectos rítmicos, entoacionais, de continuidade e de taxa de elocução e articulação. A confrontação de contornos de duração e f0 é utilizada para a interpretação dos dados da análise prosódica. Pautamo-nos pela concepção de que o vínculo entre som e sentido na fala é de natureza motivada ou arbitrária, usado de forma consciente ou inconsciente, estabelecido historicamente e modificado constantemente na cultura, pelos processos de desmotivação e remotivação do signo. O indivíduo suscita na atividade prática os índices de caráter gestual, denominados de gestos vocais, que são compostos pela interação (e impossível dissociação) dos elementos prosódicos (qualidade e dinâmica da voz) com os segmentos fonéticos (vogais e consoantes) e sons não verbais produzidos na comunicação (ruídos respiratórios, sons bucais e linguais). Enquanto movimentos corporais, de uma ou mais estruturas do trato vocal que se deslocam no tempo e no espaço, de forma sincrônica ou não, os gestos vocais são utilizados para reduzir imediatamente a tensão e podem ser a reprodução voluntária ou assinalar a presença de uma emoção. Enquanto elementos simbólicos, os gestos vocais pertencem ao sistema total de signos expressivos que respondem a um simbolismo sonoro universal, podendo representar outros objetos animados ou inanimados que lhes são associados pela semelhança ou por uma analogia funcional, e podem assumir uma determinada configuração e/ou estão sujeitos à ativação do organismo, ao veicular uma informação lingüística ou uma emoção ou atitude. Os gestos vocais, enquanto fatos estilísticos, são representantes da dialogia da língua e são uma marca prática do trabalho do locutor, que exibe sua singularidade e sua subjetividade, de forma recorrente e saliente no discurso, construindo e sendo constituídos pelo sentido, ora no papel principal ora como coadjuvante. Os gestos vocais são os elementos fisiológicos e lingüísticos dinâmicos que efetivam na expressão do indivíduo as demandas contextuais e subjetivas que refletem a variabilidade da língua e, por isso, integram a voz no universo da linguagem

Nesse trabalho procuro examinar as relações de João Gilberto com o samba e com os modos tradicionais do canto brasileiro, particularmente o estilo do samba sincopado. Representante fundamental de um estilo que levou a música brasileira ao centro da indústria cultural, estarei em busca das continuidades e desenvolvimentos que esse estilo promove em relação ao samba feito nos subúrbios das cidades brasileiras em formação. Invertendo a orientação frequentemente biográfica da bibliografia tradicional, estarei em busca das novidades trazidas por João Gilberto no que diz respeito aos parâmetros musicais dos timbres, durações, alturas e intensidades. Longe de desprezar os estudos biográficos já feitos sobre o autor, a estratégia da inversão pretende fomentar um novo ambiente de debate com bases tão sólidas quanto aquelas, no qual se criem condições de dialogar com a bibliografia tradicional por um novo viés: o da musicologia. / In this academic work I examine the relation of João Gilberto with samba and traditional Brazilian singing, particularly the syncopated samba style. Through this representative of a style that brought Brazilian music to the center of the cultural industry, I will be searching for continuities and developments of the samba made in the outskirts of emerging Brazilian cities, promoted by that style. Inverting the orientation, frequently biographic of his traditional bibliography, I will be looking for new developments brought by João Gilberto on tones, durations, intensities and timbres. The strategy of this inversion does not intend to ignore the biographical studies already done, but to foster a new environment for discussion with a solid bases as well, in which conditions are created to dialogue with the traditional bibliography by a new bias: one of musicology.

Made available in DSpace on 2016-04-28T20:56:58Z (GMT). No. of bitstreams: 1 Sandra Puliezi.pdf: 2112077 bytes, checksum: 65039a7eb052fad7e1682ee760969c50 (MD5) Previous issue date: 2015-12-10 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / Current studies about reading show that there is a strong relationship between comprehension and reading fluency, to the extent that one influences another. In order to contribute to the growth of this knowledge, we conducted a study to investigate the relationship between reading comprehension and reading fluency of Brazilian children who attend the 4th year of elementary school. Objectives of this research were: verify the effects of an intervention in comprehension; verify the effects of an intervention in reading fluency; and create a series of activities for the improvement and stimulation of reading fluency and comprehension. To accomplish these objectives, we created two intervention groups (one in fluency and the other in comprehension) and a control group. At the beginning of the school year all the groups completed two tests: an assessment of fluency (accuracy, speed and prosody) and an assessment of comprehension. The intervention lasted three months. At the end of that time, every group completed the same assessments of fluency and comprehension, plus a test to evaluate inferential comprehension. We then analyzed the data to compare student performance between the three groups. The results showed that interventions helped to improve the performance of students in the skills that were being targeted, but not the other skill. Pearson's correlations allowed us to see that high performance in comprehension is accompanied by high performance in fluency, highlighting the importance of the two skills for success in reading. Our study indicated the need to direct teaching in order to fully develop the three dimensions of fluency and the different levels of comprehension, leading us to conclude that the teaching of one of these skills alone does not ensure success in the other / Estudos atuais sobre leitura mostram que existe uma relação muito forte entre compreensão e fluência na leitura, a ponto de haver influência de uma sobre a outra. Com o intuito de contribuir para o crescimento desse conhecimento, conduzimos um estudo para investigar relações entre a compreensão de textos e a fluência na leitura de crianças brasileiras que cursam o 4º ano do ensino fundamental. Os objetivos foram: verificar os efeitos de uma intervenção em compreensão; verificar os efeitos de uma intervenção em fluência e criar uma série de atividades para a melhoria e estimulação da fluência de leitura e da compreensão. Para tanto, foram constituídos dois grupos de intervenção (um em fluência e outro em compreensão) e um grupo controle. No início do ano letivo todos os grupos passaram pelo pré-teste que continha duas provas: uma para avaliar a fluência (precisão, velocidade e prosódia) e uma para avaliar a compreensão. A intervenção durou três meses. Ao fim desse tempo, todos os grupos passaram pelo pós-teste, constituído pelas mesmas provas do pré-teste, acrescido de uma prova que avaliou a compreensão inferencial. Na análise dos dados foram comparados o desempenho dos alunos dos três grupos. Os resultados mostraram que as intervenções ajudaram a melhorar o desempenho dos alunos nas habilidades que estavam sendo treinados, mas não na outra. As correlações de Pearson nos permitiram ver que o bom desempenho em compreensão é acompanhado do bom desempenho em fluência, evidenciando a importância das duas habilidades para o sucesso na leitura. Nosso estudo indicou que é preciso haver um ensino direcionado para podermos desenvolver ao máximo as três dimensões da fluência e os diferentes níveis de compreensão, fazendo-nos concluir que apenas o ensino de uma dessas habilidades é insuficiente para o sucesso da outra

Made available in DSpace on 2014-06-11T19:22:20Z (GMT). No. of bitstreams: 0 Previous issue date: 2007-03-08Bitstream added on 2014-06-13T20:09:21Z : No. of bitstreams: 1 serra_mp_me_sjrp.pdf: 745870 bytes, checksum: b5514841b084b9e9d740d730a13fc3c2 (MD5) / Propusemos, neste trabalho, a observação de operações de reelaboração da segmentação em textos infantis. Interessaram-nos, particularmente, as marcas de reelaboração textual (mais especificamente, de reelaboração da segmentação). O propósito foi observar que fatores estariam envolvidos em momentos nos quais crianças, em momentos iniciais do processo de aquisição da escrita, distribuíam o material gráfico que produziam em porções menores e realizavam o que denominamos reelaboração da segmentação textual. O material que usamos foi formado de produções textuais de alunos de uma escola da rede pública de ensino do município de São José do Rio Preto (SP), que foram coletadas quinzenalmente e acompanhadas de modo longitudinal por um período de quatro anos. A partir desse conjunto, elegemos 62 textos, de 20 sujeitos que, durante o ano de 2002, cursavam a segunda série do ensino fundamental I. Cada produção foi selecionada por apresentar pelo menos duas ocorrências, num mesmo texto, em que o escrevente flutuava quanto à segmentação escrita. Fizemos um levantamento quantitativo das ocorrências de reelaboração em relação a constituintes prosódicos mobilizados, a fim de possibilitar a formulação de mais argumentos para o fato de que os constituintes prosódicos mobilizados nessas ocorrências teriam algum tipo de realidade para a criança/aprendiz da escrita. Estivemos atentos também para o que, nestas ocorrências, nos possibilitou conjeturar sobre a relação entre o que pôde ser considerado como fatos episódicos de escrita infantil e para o que pôde ser considerado como genérico. Com base, de um lado, em pressupostos teórico-metodológicos vinculados ao Paradigma Indiciário proposto por Ginzburg (1986) e, de outro, na idéia de um modo heterogêneo de constituição da escrita, tal como formulada por Corrêa (1997 e 2004).... / In this work, we propose the observation of operations involved in the redevelopment of segmentation of infant writing. We concentrate, particulary, on textual redevelopment marks (more specifically on segment redevelopment). Our aim was to observe which factors are involved in the initial process of written acquisition and how, at this crucial time, children distribute and produce graphic material along the texts and realize sucessfully what we refer to as textual section development. Resource material came from written work from public school students in São José do Rio Preto (SP), which were collected on a fortnightly basis during a four year period. From this material, we selected 62 texts written by twenty people who attended the second grade of fundamental education in 2002. Each text was chosen due to the appearance of floating writing in the written segments on at least two occasions. A quantitative analysis between the incidence of redevelopment and the relation of prosodic constituents that are involved was carried out with the aim to enabling the formulation of arguments to the fact that prosodic constituents on these occasions had some contribution to children skill aquision. We also took care to note that these events enable us to deliberate about the possibility of what might be consider sporadic facts in infant writting and that could be considered as generic. Based upon presumed theorectical-methods tied to Paradigma Indiciário on one side, and the idea of a heterogenic method of written formulation, such as that proposed by Côrrea (1997 and 2004), on the other, we arrive at a proposition that, on these occasions, the interlacing of reconstructed aspects of the origin of written skills (in the case of children), and the institucionalized written code stirred up not only tension concerning the delimitation of works...(Complete abstract, click electronic address below).

Notre recherche a pour objectif de tester l’incidence de la pratique linguistique à l’oral sur les compétences en lecture d’apprenants de Français Langue Etrangère (FLE). A partir d’une hypothèse en psycholinguistique sur le rôle de la prosodie implicite dans le décodage de l’écrit en Langue Maternelle (Fodor, 2002), nous postulons qu’un travail poussé de correction phonétique centré sur la sensibilisation à la structure prosodique du français (via la Méthode Verbo Tonale, MVT) va non seulement améliorer la fluence des apprenants en parole spontanée et en parole lue, mais aussi faciliter le décodage de l’écrit. Nous pensons que cet apprentissage sera plus efficace s’il est proposé dès le niveau débutant. Pour tester ces hypothèses, nous avons mené deux études longitudinales avec des apprenants anglophones de FLE. Nous avons mesuré la fluence en lecture oralisée des apprenants avant et après la formation. Les résultats des analyses acoustiques et perceptives de la première étude (étude pilote) nous ont permis de confirmer nos hypothèses et d’appuyer nos choix méthodologiques. Dans la deuxième étude, nous avons souhaité approfondir les résultats obtenus dans l’étude pilote et valider expérimentalement l’efficacité de la MVT. Nous avons comparé l’impact de la MVT et de la méthode Articulatoire dans l’évolution de la fluence des apprenants. Nous avons également testé un des présupposés de la méthodologie SGAV (dont est issue la MVT) selon lequel l’écrit ne devrait pas être introduit avant que les paramètres prosodiques ne soient maitrisés, sans quoi le passage à l’écrit entrainerait une chute de la fluence des apprenants. Les résultats de cette seconde étude vont dans le sens de nos hypothèses : seuls les apprenants qui ont suivi les cours de correction phonétique via la MVT ont amélioré leur fluence en parole lue. De plus, l’introduction de l’écrit fait chuter la fluence des productions chez tous les apprenants. En conclusion, la maitrise de la prosodie constitue un prérequis nécessaire au passage à l’écrit en langue étrangère. / This study investigates the influence of oral training on reading out loud in French as a Foreign Language (FFL). Based on the implicit prosody hypothesis by Fodor (2002), we venture that working on the phonetic correction orally will improve speaking fluency, as well as reading fluency, and facilitate the decoding of the written test. We think that the impact of this training will be stronger with beginner than advanced learners. In order to test these hypotheses, we have conducted two longitudinal studies with English FFL students. We have measured the students’ fluency before and after training. The acoustics and perceptive results of the first study have confirmed our hypothesis and supported our methodology. In the second study, we follow up on these results on a larger-scale longitudinal study, designed to systematically compare the impact of two phonetics teaching methods on reading skills for FFL learners of French: the Articulatory Method (AM) and the Verbo-Tonal Method (VTM). We have also tested one of the presuppositions of the Structuro-Global Audiovisual Methodology (SGAV) – from which comes the VTM – according to which readings and writings activities should be delayed until students’ phonetic and prosodic skills have reached a basic and steady level. Our data support our hypotheses: only the students who have received VTM classes have improved their reading fluency. Furthermore, the introduction of reading activities during the training has caused a decline in all the students' fluency. Thus mastering second language prosody is essential before introducing reading or writing activities.