Cette thèse s'inscrit dans le cadre de la synthèse de parole à partir du texte. Elle traite plus précisément du procédé de création de voix en synthèse par sélection d'unités. L'état de l'art repose pour cela sur l'enregistrement d'un locuteur pendant une à deux semaines, suivant un script de lecture de plusieurs dizaines de milliers de mots. Les 5 à 10 heures de parole collectées sont généralement révisées par des opérateurs humains, pour en vérifier la segmentation phonétique et ainsi améliorer la qualité finale de la voix de synthèse.La lourdeur générale de ce procédé freine considérablement la diversification des voix de synthèse ; aussi en proposons-nous ici une rationalisation. Nous introduisons une nouvelle unité, appelée "sandwich vocalique", pour l'optimisation de la couverture des scripts de lecture. Sur le plan phonétique, cette unité offre une meilleure prise en compte des limites segmentales de la synthèse par sélection que les unités traditionnelles (diphones, triphones, syllabes, mots, etc.). Sur le plan linguistique, un nouvel enrichissement contextuel nous permet de mieux focaliser la couverture, sans négliger les aspects prosodiques. Nous proposons des moyens d'accroître le contrôle sur les phrases du script lecture, tant dans leur longueur que dans leur pertinence phonétique et prosodique, afin de mieux anticiper le contenu du corpus de parole final et de rendre automatisable la tâche de segmentation. Nous introduisons également une alternative à la stratégie classique de condensation de corpus en mettant au point un algorithme semi-automatique de création de phrases, grâce auquel nous accroissons de 30 à 40% la densité linguistique du script de lecture.Ces nouveaux outils nous permettent d'établir un procédé très efficace de création de voix de synthèse, procédé que nous validons à travers la création et l'évaluation subjective de nombreuses voix. Des scores perceptifs comparables à l'approche traditionnelle sont ainsi atteints dès 40 minutes de parole (une demi-journée d'enregistrement) et sans post-traitement manuel. Enfin, nous mettons à profit ce résultat pour enrichir nos voix de synthèse de diverses composantes expressives, multi-expressives et paralinguistiques. / This work falls within the scope of text-to-speech (TTS) technology. More precisely, focus is on the voice creation process for unit-selection synthesis. In a standard approach, a textual script of several thousands of words is read by a speaker in order to generate approximately 5 to 10 hours of useable speech. The recording time is spread out over one or two weeks and is followed by the considerable task of manually revising the phonetic segmentation for all of the speech.Such a costly and time-consuming process presents a major obstacle to diversifying synthesized voices. In order to increase efficiency in this process, we introduce a new unit, called a "vocalic sandwich", to optimize coverage of the recording texts. Phonetically, this unit better addresses the segmental limitations of unit-selection TTS than state-of-the-art units (diphones, triphones, syllables, words...). Linguistically, a new set of contextual symbols focuses the coverage, allowing for more control and consideration of prosody. Practically, in order to automate the segmentation process, better anticipation of the phonetic and prosodic content desired in the final database is required. This is achieved here by increasing the readability and consistency of each sentence included in the script. As a side, these properties also help to facilitate the reading stage. Furthermore, as an alternative to the classic corpus condensation, a semi-automatic sentence building algorithm is developed in this work wherein sentences are built rather than selected from a reference corpus. Ultimately, the sentence building provides access to much denser scripts, specifically allowing for increases in density of between 30 and 40%.In incorporating these new approaches and tools, the voice creation process is made very efficient, as is validated in this work through the preparation and evaluation of numerous synthesized voices. Perceptive scores that are comparable to the traditional process are achieved with 40 minutes of speech (half-day recording) and without any manual post-processing. Finally, we take advantage of these results in order to enhance our synthesized voices with various expressive, multi-expressive and paralinguistic features.

La production verbale écrite constitue un vaste champ d’investigation pour les sciences cognitives. A l’heure où il est demandé de plus en plus tôt de savoir produire avec les outils numériques de production verbale écrite (clavier d’ordinateur et tablette tactile), l’utilisation de ces outils nécessite d’être étudiée. Nous proposons d’examiner les effets de trois outils sur la production (stylo numérique, clavier et tablette tactile). Pour ce faire, nous avons réalisé quatre études : la première porte sur la production orthographique en condition de dictée ; la deuxième s’intéresse à la copie de mots ; la troisième concerne la reproduction de phrases ; la quatrième aborde la production de textes.Nos analyses montrent que la variable outil de production a une grande importance lors de la production verbale écrite de mots, de phrases ainsi que sur la production de textes. Nous avons pu mettre en évidence l’effet de ces outils sur les temps de production, le nombre d’erreurs, la récupération d’information, le débit et la qualité des textes produits. Nos recherches ouvrent des perspectives nouvelles tant sur le plan méthodologique que sur le plan théorique. En effet, l’outil de production est une variable qu’il faut prendre en compte dans l’étude de la production verbale écrite. / Written verbal production is a vast field of investigation for cognitive sciences. At a time when it is asked at an increasingly early age to know how to produce with the digital tools of written verbal production (computer keyboard and touch tablet), the use of these tools needs to be studied. We propose to examine the effects of three tools on production (digital pen, keyboard and touch pad). To do so, we conducted four studies: the first deals with orthographic production in dictation condition; the second is about copying words; the third concerns the reproduction of sentences; the fourth deals with the production of texts.Our analyzes show that the variable production tool is of great importance in the verbal production of words, phrases and the production of texts. We have been able to highlight an effect of these tools on production times, the number of errors, the retrieval of information, the throughput and the quality of the texts produced. Our research opens up new perspectives both from a methodological and a theoretical point of view. Indeed, the tool of production is a variable that must be taken into account in the study of written verbal production.

中文的一個獨特之處在於，字和詞這兩個水平的單位都可以傳達意義。在中文里，字和詞分別有其獨自的特點。大多數漢字都由部件組成，這些部件多數情況下與整字的讀音和意義有關。詞通常由兩個或兩個以上的字構成，其中多數是復合詞。根據詞中字之間的關系，復合詞有五種常見結構。由於字和詞都傳達意義，於是就產生了兩個漢語中特有的問題：一是字和詞之中哪一個是閱讀加工的基本單位，二是字詞加工之間的關系如何。之前的一些研究為回答這兩個問題提供了一些依據。一些以成人閱讀者作為被試的研究發現，雖然從形式上來講，字是中文文本的基本單位，但是在閱讀中通常作為基本加工單位的是詞。有些研究也發現詞和字的加工是相互影響的。 / 但是，這些研究很少以兒童作為對象。本文中的三個實驗試圖探索上述兩個問題在兒童閱讀者之中的答案，用三年級和五年級的香港本地兒童作為被試以探討有關加工隨著年齡發展的變化。在實驗 1 中，被試被要求讀出一些單字，以及由這些單字組成的詞。結果顯示多數在單字閱讀中出現的錯誤，在讀詞時得到了更正，說明詞有助於兒童對字的識別。與年長的兒童相比，年幼兒童更容易將單字混淆於其他意義相關的單字，或者混同於那些經常與目標字一起出現組成詞的單字，這表明年幼兒童的字加工更依賴於該字經常出現的語境。實驗 2 使用偏正結構詞，並列結構詞以及非詞等三種結構的刺激檢驗了詞的結構對於詞中字的加工的影響。同預期相反，該實驗的結果表明，在偏正結構的詞中，兒童對后一個字的反應并沒有因為其與整詞意義關聯緊密而更加準確和迅速。更有甚者，三年級和五年級的兒童對於偏正結構詞中前一個字的表現都好於后一個字，並且與並列結構詞和非詞相比，偏正結構中前一個字較之后一個字的優勢似乎更加突出。 / 雖然結果與預期不相符，但是三年級和五年級兒童表現比較一致，說明詞結構很有可能確實對其中的字的加工產生了影響，但是這個影響的機制並未由實驗 2揭示出。實驗 3 比較了字音和字形在詞加工中的作用。實驗結果顯示，與年長兒童相比，年幼兒童對於詞的閱讀更加依賴於讀音。探索字詞加工之間的關系需要更多的研究提供證據，而本研究的三個實驗則為以后的研究提供了一些方向。 / One of the unique features in Chinese is that two levels of unit, character and word, are both related to meaning. Both of them have their own features. Most characters are composed by radicals which may give a cue to the characters’ pronunciation and meaning. Words are mostly composed by no less than two characters, and most of them are compound words which usually have 5 types of morphological structures according to the relationship between the meaning of the whole word and its component characters. Since both characters and words convey meaning, two questions arise: which of them is the basic unit of processing when people read, and what is the relationship between readers’ character and word processing. Previous research has provided some evidence to answer these questions. Although characters are the basic graphic units in Chinese script, several studies have shown that adults’ reading is based on words rather than on characters. Meanwhile, it has been found that the processing of words and characters also affect each other. / However, not many studies used children as participants. In this thesis, three experiments were designed to explore these two questions. Children of grade 3 and 5 in Hong Kong primary school were used as participants to investigate the developmental change. In experiment 1, children were asked to read single characters and words that were composed by these characters. Results showed that words actually helped children to recognize characters, for most errors in single character reading got corrected in word reading. Compared to older children, younger children’s character processing was more likely to rely on the context that the character might appear, for they had more errors than older children of confusing the correct character with the word related or meaning related ones. In experiment 2, the effect of word’s morphological structure on the component characters was examined. Three types of structures, i.e., modifier words, coordinative words and nonwords were used to make comparison. The results were opposite to the expectations, for the closer relationship with the word didn’t get any advantage for the second characters in modifier words in lexical decision task. Moreover, both P3 and P5 children’s performance in modifier words on the first character was better than on the second character, and this difference was even larger than in coordinative words and nonwords. Since children of both grades showed the same pattern, the word’s morphological structure still seemed to have effects on the component characters processing, although the mechanism was not clear. Experiment 3 compared the influence of pronunciation and graphemic similarity on word reading, and compared to older children, younger children were more likely to rely on pronunciation, which was consistent with previous studies. This study provided possible directions for future study, and more evidence on the relation between children’s character and word processing is still needed. / Detailed summary in vernacular field only. / Detailed summary in vernacular field only. / Detailed summary in vernacular field only. / Li, Tong. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2012. / Includes bibliographical references (leaves 55-62). / Abstracts also in Chinese; appendix includes Chinese. / Abstract --- p.i / Abstract (in Chinese) --- p.iii / Acknowledgements --- p.v / Table of contents --- p.vi / Chapter 1 --- Introduction --- p.1 / Chapter 1.1 --- Introduction of the basic characteristics of Chinese characters and words --- p.1 / Chapter 1.2 --- The relationship between processing of word and character --- p.7 / Chapter 1.3 --- The present study --- p.15 / Chapter 2 --- Experiment 1 --- p.18 / Chapter 2.1 --- Method --- p.18 / Chapter 2.1.1 --- Participants --- p.18 / Chapter 2.1.2 --- Material --- p.18 / Chapter 2.2 --- Results --- p.20 / Chapter 2.3 --- Discussion --- p.26 / Chapter 3 --- Experiment 2 --- p.28 / Chapter 3.1 --- Method --- p.29 / Chapter 3.1.1 --- Participants --- p.29 / Chapter 3.1.2 --- Material --- p.29 / Chapter 3.1.3 --- Procedures --- p.33 / Chapter 3.2 --- Results --- p..34 / Chapter 3.3 --- Discussion --- p.35 / Chapter 4 --- Experiment 3 --- p.38 / Chapter 4.1 --- Method --- p.38 / Chapter 4.1.1 --- Participants --- p.38 / Chapter 4.1.2 --- Material --- p.38 / Chapter 4.1.3 --- Procedures --- p.39 / Chapter 4.2 --- Results --- p..40 / Chapter 4.3 --- Discussion --- p.42 / Chapter 5 --- General Discussion --- p.43 / Appendix --- p.48 / Bibliography --- p.54

Submitted by Sabrina de Cassia Martins null (sabrismartins@gmail.com) on 2017-02-08T15:06:43Z No. of bitstreams: 1 Tese-Sabrina Martins_VERSÃO FINAL.pdf: 4892311 bytes, checksum: ded93001eeec82424d2a73ec7626a3d8 (MD5) / Approved for entry into archive by LUIZA DE MENEZES ROMANETTO (luizamenezes@reitoria.unesp.br) on 2017-02-13T18:13:27Z (GMT) No. of bitstreams: 1 martins_sc_dr_sjrp.pdf: 4892311 bytes, checksum: ded93001eeec82424d2a73ec7626a3d8 (MD5) / Made available in DSpace on 2017-02-13T18:13:27Z (GMT). No. of bitstreams: 1 martins_sc_dr_sjrp.pdf: 4892311 bytes, checksum: ded93001eeec82424d2a73ec7626a3d8 (MD5) Previous issue date: 2016-12-12 / Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) / A presente tese tem como objeto de estudo a terminologia da Fauna e da Flora cuja formação envolve o uso do vocabulário das cores. Nosso estudo abrange dois temas gerais: por um lado, a contribuição dos nomes de cores para a variação denominativa em terminologia a partir de uma perspectiva interlinguística; por outro, a descrição de tais variantes em uma obra lexicográfica especializada e multilíngue. A fim de delimitar a nossa pesquisa, concentramo-nos na análise de dois subdomínios da Biologia: a Botânica, especificamente as Angiospermas (monocotiledôneas e eudicotiledôneas), e a Zoologia, exclusivamente os Vertebrados (peixes, anfíbios, répteis, aves e mamíferos). Dessa forma, visamos, em primeiro lugar, ampliar o Dicionário Onomasiológico de Expressões Cromáticas da Fauna e Flora por meio da inserção em sua microestrutura dos correspondentes das palavras-entrada em línguas inglesa e italiana. Seguindo a metodologia proposta em nosso Mestrado (MARTINS, 2013), utilizamos o Corpus Web para a busca e validação dos correspondentes. Em segundo, uma vez verificada a existência de correspondentes para cada entrada, realizamos sua análise estrutural, observando o papel desses itens para a divulgação dos conceitos da Fauna e da Flora em outras culturas. Em terceiro, examinamos as possíveis causas que atuam na formação de tais unidades e que propiciam ou não a escolha do item cor nas línguas em estudo, isto é, português, inglês e italiano. Por último, ainda em relação à obra lexicográfica, propomos formular a metodologia mais adequada para a implementação de uma plataforma on-line que contribuirá para minimizar as divergências terminológicas nessa área, para a normatização de conceitos e para o estabelecimento da comunicação entre culturas. Enfatizamos que o nosso interesse é motivado pelo papel ativo dos nomes de cores na expansão do léxico e a investigação dessa fatia lexical é uma forma de evidenciar o uso dos nomes de cores para a difusão da terminologia da Fauna e da Flora. Finalmente, visamos contribuir qualitativamente, tanto no aspecto teórico (por meio das reflexões apresentadas) quanto no metodológico, para o âmbito das pesquisas que se limitam ao estudo do léxico, bem como para o conjunto das obras lexicográficas produzidas no Brasil. / The present thesis has as its object of study the terminology of Fauna and Flora composed by color names. Hence, it deals with two major themes: on one hand, the contribution of color names to the denominative variation in Terminology from a interlinguistic perspective; on the other, the description of such variants in a multilingual and specialized lexicographical work. In order to delimit our research, we concentrate our analysis on two subdomains of Biology: Botany, specifically the Angiosperms (monocotyledons and dicotyledons), and Zoology, specifically the Vertebrates (fish, amphibians, reptiles, birds and mammals). In this manner, it aims, firstly, to extend the Onomasiological Dictionary of Chromatic Phrases of Fauna and Flora by inserting in the microstructure the equivalents of the entries in English and Italian languages. Following the methodology proposed in a prior research (MARTINS, 2013), we use the Corpus Web to collect and validate the equivalents. Secondly, once checked the existence of equivalents for each entry, it intends to realize their structural analysis, observing the role of these denominative variants to the propagation of Fauna and Flora concepts in other cultures. Thirdly, examining the probable causes that actuate in the formation of such units and that propitiate or not the choice of the color name in the languages in study, i.e., Portuguese, English and Italian. Finally, still regarding the lexicographical work, we propose to formulate an appropriate methodology for the implementation of an on-line platform that will contribute to minimize terminological variances in this area, to concepts normalization, as well as to the communication among cultures. We would like to emphasize that our interest is motivated by the active role of color names in the expansion of the lexicon and the study of this vocabulary demonstrates the importance of color names to the diffusion of Fauna and Flora terminology. To sum up, we expect to contribute qualitatively, as in the theoretical aspect (by means of the reflections presented in the following pages) as in the methodological aspect, to the range of the lexical studies, as well as to the set of lexicographical works produced in Brazil. / FAPESP: 2013/09232-1

Submitted by Filipe dos Santos (fsantos@pucsp.br) on 2018-03-16T12:14:39Z No. of bitstreams: 1 Edgar Godoi Gabriel.pdf: 1767705 bytes, checksum: 65cd62e501122641cd12165d4cfcf6c3 (MD5) / Made available in DSpace on 2018-03-16T12:14:39Z (GMT). No. of bitstreams: 1 Edgar Godoi Gabriel.pdf: 1767705 bytes, checksum: 65cd62e501122641cd12165d4cfcf6c3 (MD5) Previous issue date: 2018-02-23 / Conselho Nacional de Pesquisa e Desenvolvimento Científico e Tecnológico - CNPq / Only few people would dare to challenge the importance of English in the globalized world we live in. To speak English has become a must for those Brazilians who are beginning or developing their career. Thus, language schools strive to persuade society, through their advertisings, of the importance of their services to insert the country in the globalized world. In this scenario, two companies stand out: Wizard and Pearson. Wizard is the world’s largest language school, with more than 1,200 branches where more than 500,000 people study per year. In 2013, Pearson the world business education leader acquired Wizard. In partnership, they released Nação Bilíngue [Bilingual Nation], Wizard’s largest advertising campaign, whose mission is ‘to transform Brazil into a bilingual nation’. We can find lots of studies on Advertising, however, the ones focusing on phrases without text (i.e. slogans, proverbs, sayings…) in language schools’ advertisings are still incipient. Advertising discourse is a profitable field for the so-called phrases without text, especially in the Digital Era, due to its spatiotemporal coercions. From a verbal and visual point of view, this dissertation aims at analyzing the discourses conveyed in the Nação Bilíngue (2014-2017); more specifically, to understand the effects of meaning production in/by the phrases without text in the Nação Bilíngue’s digital banners. Collected from Wizard’s websites, blogs, and social networks, corpus study consists of seven digital banners which belong to Nação Bilíngue three phases: Manifesto (2014-2015), Delivering Results (2016), and Anthem (2017). The analyses are based on the notions of Phrases without text, Hyper-Enunciator, and Discursive Ethos such as advocated and developed by Dominique Maingueneau. The results show that the discourses circulating in Nação Bilíngue produce the effects of meaning that English is both a means of insertion in the globalized world and a means of ascending social strata. They also indicate that, unlike conventional advertising slogans, the ones of Nação Bilíngue reveal their mimotopic character, that is, their capability to adapt to or to metamorphose into other discourses, even pretending to be a non-advertising / Atualmente, poucos ousariam contestar a importância da Língua Inglesa no mundo globalizado em que vivemos. No Brasil, saber se comunicar em inglês passou a ser uma necessidade para quem está entrando no mercado de trabalho ou deseja progredir na carreira profissional. Por meio da publicidade, institutos/escolas de idiomas têm buscado persuadir a sociedade da importância de seus serviços para a inserção do país no mundo globalizado. Nesse contexto, sobressaem-se duas empresas: Wizard e Pearson. A Wizard é a maior rede de escolas de idiomas do mundo, com mais de 1.200 unidades, por onde passam mais de 500 mil alunos por ano; em 2013, foi adquirida pela Pearson, líder mundial em negócios em educação. Juntas, lançaram, em 2014, Nação Bilíngue, a maior campanha publicitária da Wizard, cuja missão é “transformar o Brasil em uma nação bilíngue”. De modo geral, a publicidade tem sido objeto frequente de estudos acadêmicos, no entanto, ainda são incipientes as pesquisas que se debruçam sobre a circulação das chamadas frases sem texto (slogans, provérbios, ditos populares...) na publicidade de institutos/escolas de idiomas. Daí a importância deste trabalho, que tem por objetivo analisar, do ponto de vista verbal e visual, os discursos veiculados na campanha publicitária Nação Bilíngue (2014-2017), mais especificamente, depreender os efeitos de sentido produzidos nas/pelas frases sem texto que circulam nos cartazes/banners digitais da referida campanha. Coletados das páginas digitais (websites, blogs e redes sociais) da escola de idiomas Wizard, o corpus é constituído por sete cartazes/banners, referentes às três fases iniciais da campanha: Manifesto (2014-2015), Entrega Resultados (2016) e Hino (2017). As análises fundamentam-se nas noções de Frases sem texto, Hiperenunciador e Ethos Discursivo, como preconizadas e desenvolvidas por Dominique Maingueneau. Os resultados mostram que os discursos que circulam na campanha publicitária Nação Bilíngue produzem o efeito de sentido segundo o qual a língua inglesa é tanto uma forma de inserção no mundo globalizado quanto um meio de ascensão social. Indicam ainda que, diferentemente dos slogans publicitários convencionais, os de Nação Bilíngue revelam seu caráter mimotópico, isto é, sua capacidade de adaptar-se a ou metamorfosear-se em outros discursos, simulando até mesmo ser uma não publicidade

Dans cette thèse nous avons exploré la capacité des modèles magnétiques de la physique statistique à extraire l'information essentielle contenue dans les textes. Les documents ont été représentés comme des ensembles d'unités en interaction magnétique, l'intensité de telles interactions a été mesurée et utilisée pour calculer de quantités qui sont des indices de l'importance de l'information portée. Nous proposons deux nouvelles méthodes. Premièrement, nous avons étudié un modèle de spins qui nous a permis d'introduire l'énergie textuelle d'un document. Cette quantité a été utilisée comme indicatrice de pertinence et appliquée à une vaste palette de tâches telles que le résumé automatique, la recherche d'information, la classification de documents et la segmentation thématique. Par ailleurs, et de façon encore exploratoire, nous proposons un deuxième algorithme qui définie un couplage grammatical pour conserver les termes importants et produire des contractions. De cette façon, la compression d'une phrase est l'état fondamental de la chaîne de termes. Comme cette compression n'est pas forcement bonne, il a été intéressant de produire des variantes en permettant des fluctuations thermiques. Nous avons fait des simulations Métropolis Monte-Carlo avec le but de trouver l'état fondamental de ce système qui est analogue au verre de spin. Les deux systèmes, utilisant des méthodes numériques, restent indépendants de la langue.

Cette thèse a pour objectif de répondre à deux questions fondamentales sur la compréhension du langage :<br />Quelle stratégie adopte le lecteur lorsqu'il rencontre une ambiguïté ?<br />Quelles sont les informations utilisées pour résoudre une ambiguïté syntaxique ?<br />Les modèles actuels de la compréhension de phrases ne permettent pas de rendre compte de l'ensemble des données inter-langues observées dans le cas de l'attachement d'une Proposition Relative (PR) ambiguë. Une série d'expériences a donc été réalisée afin de déterminer les facteurs responsables de l'attachement d'une PR. Nous avons utilisé la technique d'enregistrement des mouvements oculaires afin d'appréhender le décours temporel des opérations qui ont lieu pendant la lecture. Nous avons observé que la longueur des constituants influence la préférence d'attachement d'une PR (Expériences 1-4). Pour rendre compte de ces résultats, nous faisons l'hypothèse que certains attachements peuvent être laissés en suspens (sous-spécifiés) en attendant que l'information désambiguïsatrice soit disponible. À partir de cette hypothèse, nous avons implémenté un modèle qui nous permet de simuler nos résultats et de faire de nouvelles hypothèses quant aux facteurs susceptibles d'influencer la préférence d'attachement d'une PR. Ces prédictions ont été vérifiées dans l'Expérience 5. L'idée qu'en cas d'ambiguïté syntaxique, la décision syntaxique peut être reportée est développée dans la théorie Construal (Frazier et Clifton, 1996). Cependant, contrairement à notre modèle qui prédit que tous les constituants peuvent recevoir une description sous-spécifiée, la Construal propose que seules les relations non primaires reçoivent une analyse sous-spécifiée, les relations primaires recevant immédiatement une analyse syntaxique unique et spécifiée. Les Expériences 6-11 révèlent que la distinction entre relations primaires et non primaires proposée par la Construal n'est pas pertinente et nous permettent de conclure en faveur de notre modèle.

University of Macau / Faculty of Social Sciences and Humanities / Department of Chinese

University of Macau / Faculty of Social Sciences and Humanities / Department of Chinese

published_or_final_version / abstract / English / Master / Master of Arts