published_or_final_version / abstract / toc / Linguistics / Master / Master of Arts

This thesis studies disfluency in spontaneous Swedish speech, i.e., the occurrence of hesitation phenomena like eh, öh, truncated words, repetitions and repairs, mispronunciations, truncated words and so on. The thesis is divided into three parts: PART I provides the background, both concerning scientific, personal and industrial–academic aspects in the Tuning in quotes, and the Preamble and Introduction (chapter 1). PART II consists of one chapter only, chapter 2, which dives into the etiology of disfluency. Consequently it describes previous research on disfluencies, also including areas that are not the main focus of the present tome, like stuttering, psychotherapy, philosophy, neurology, discourse perspectives, speech production, application-driven perspectives, cognitive aspects, and so on. A discussion on terminology and definitions is also provided. The goal of this chapter is to provide as broad a picture as possible of the phenomenon of disfluency, and how all those different and varying perspectives are related to each other. PART III describes the linguistic data studied and analyzed in this thesis, with the following structure: Chapter 3 describes how the speech data were collected, and for what reason. Sum totals of the data and the post-processing method are also described. Chapter 4 describes how the data were transcribed, annotated and analyzed. The labeling method is described in detail, as is the method employed to do frequency counts. Chapter 5 presents the analysis and results for all different categories of disfluencies. Besides general frequency and distribution of the different types of disfluencies, both inter- and intra-corpus results are presented, as are co-occurrences of different types of disfluencies. Also, inter- and intra-speaker differences are discussed. Chapter 6 discusses the results, mainly in light of previous research. Reasons for the observed frequencies and distribution are proposed, as are their relation to language typology, as well as syntactic, morphological and phonetic reasons for the observed phenomena. Future work is also envisaged, both work that is possible on the present data set, work that is possible on the present data set given extended labeling and work that I think should be carried out, but where the present data set fails, in one way or another, to meet the requirements of such studies. Appendices 1–4 list the sum total of all data analyzed in this thesis (apart from Tok Pisin data). Appendix 5 provides an example of a full human–computer dialogue. / The electronic version of the printed dissertation is a corrected version where typos as well as phrases have been corrected. A list with the corrections is presented in the errata list above.

An equivalence class exists if the stimuli that comprise the class are related by the properties of reflexivity, symmetry, and transitivity. Through these properties, new behavior that has not been taught emerges. For example, when taught to match Set A stimuli to Set B stimuli and to match Set A stimuli to Set C stimuli, if equivalence classes have formed, subjects will (with no explicit instruction) match Set B stimuli to Set C and Set C stimuli to Set B stimuli. Although equivalence classes have been studied extensively, few studies have considered the application of this technology to educational concerns. The purpose of this study was (a) to determine if phonics skills could effectively and efficiently be taught to young children through the formation of equivalence classes and (b) to investigate the generality of those acquired skills. Using a conditional discrimination procedure, children were taught to match printed letters to dictated phonetic sounds and to match the initial sound of pictured items to dictated phonetic sounds. Test results indicated that equivalence classes had emerged and that generalization did occur. The children could match the initial sound of pictured items to printed letters and vice versa and could name letter sounds and initial sounds of pictured items. During generality testing, each child could identify the initial sound of several novel pictured items and could sound out the letters within the words. However, reading did not occur. Only 1 of 5 children could blend the sounds of letters into recognizable words. A significant difficulty encountered throughout the study was maintaining the children's motivation, possibly due to the children's inexperience in attending to academic tasks. This study did, however, demonstrate that the formation of equivalence classes is an effective and efficient method for teaching phonics and that the formation of generalized equivalence classes is effective in extending those taught relations to novel stimuli. / Department of Special Education

Thesis (Ed.D.)--Teachers College, Columbia University, 1996. / Typescript; issued also on microfilm. Sponsor: Lenore Pogonowski. Dissertation Committee: Harold F. Abeles. Accompanying tape has Recitation of Chinese poems. Includes bibliographical references (leaves 251-257).

In this thesis, we aim to explore the combination of different lexical normalization methods and provide a practical lexical normalization pipeline for Swedish student writings within the framework of SWEGRAM(Näsman et al., 2017). An important improvement in my implementation is that the pipeline design should consider the unique morphological and phonological characteristics of the Swedish language. This kind of localization makes the system more robust for Swedish at the cost of being less applicable to other languages in similar tasks. The core of the localization lies in a phonetic algorithm we designed specifically for the Swedish language and a compound processing step for Swedish compounding phenomenon. The proposed pipeline consists of four steps, namely preprocessing, identification of out-of-vocabulary words, generation of normalization candidates and candidate selection. For each step we use different approaches. We perform experiments on the Uppsala Corpus of Student Writings (UCSW) (Megyesi et al., 2016), and evaluate the results in termsof precision, recall and accuracy measures. The techniques applied to the raw data and their impacts on the final result are presented. In our evaluation, we show that the pipeline can be useful in the lexical normalization task and our phonetic algorithm is proven to be effective for the Swedish language.

Statistical parametric text-to-speech synthesis currently relies on predefined and highly controlled prompts read in a “neutral” voice. This thesis presents work on utilising recordings of free conversation for the purpose of filled pause synthesis and as an inspiration for improved general modelling of speech for text-to-speech synthesis purposes. A corpus of both standard prompts and free conversation is presented and the potential usefulness of conversational speech as the basis for text-to-speech voices is validated. Additionally, through psycholinguistic experimentation it is shown that filled pauses can have potential subconscious benefits to the listener but that current text-to-speech voices cannot replicate these effects. A method for pronunciation variant forced alignment is presented in order to obtain a more accurate automatic speech segmentation something which is particularly bad for spontaneously produced speech. This pronunciation variant alignment is utilised not only to create a more accurate underlying acoustic model, but also as the driving force behind creating more natural pronunciation prediction at synthesis time. While this improves both the standard and spontaneous voices the naturalness of spontaneous speech based voices still lags behind the quality of voices based on standard read prompts. Thus, the synthesis of filled pauses is investigated in relation to specific phonetic modelling of filled pauses and through techniques for the mixing of standard prompts with spontaneous utterances in order to retain the higher quality of standard speech based voices while still utilising the spontaneous speech for filled pause modelling. A method for predicting where to insert filled pauses in the speech stream is also developed and presented, relying on an analysis of human filled pause usage and a mix of language modelling methods. The method achieves an insertion accuracy in close agreement with human usage. The various approaches are evaluated and their improvements documented throughout the thesis, however, at the end the resulting filled pause quality is assessed through a repetition of the psycholinguistic experiments and an evaluation of the compilation of all developed methods.

Pas de résumé disponible. / Pas de résumé disponible.

218f. / Submitted by Suelen Reis (suziy.ellen@gmail.com) on 2013-05-13T18:41:49Z No. of bitstreams: 1 dissertacao Gredson dos Santos.pdf: 1108137 bytes, checksum: a5b107ca74f32536f7d4344e97ba074c (MD5) / Approved for entry into archive by Alda Lima da Silva(sivalda@ufba.br) on 2013-05-16T17:07:24Z (GMT) No. of bitstreams: 1 dissertacao Gredson dos Santos.pdf: 1108137 bytes, checksum: a5b107ca74f32536f7d4344e97ba074c (MD5) / Made available in DSpace on 2013-05-16T17:07:24Z (GMT). No. of bitstreams: 1 dissertacao Gredson dos Santos.pdf: 1108137 bytes, checksum: a5b107ca74f32536f7d4344e97ba074c (MD5) Previous issue date: 2006 / Esta dissertação, de natureza empírica e de cunho predominantemente descritivo, buscou registrar aspectos fonéticos em variação em estudantes de duas áreas rurais do município de Catu-Ba. Procurou também verificar se e até que ponto os traços em variação na fala espontânea daqueles indivíduos se refletem na escrita monitorada dos mesmos. Além disso, tentou desenvolver uma reflexão sobre as implicações pedagógicas da variação lingüística. A metodologia utilizada permitiu a constituição de um corpus composto de duas amostras: a primeira resultante da gravação de cerca de dez horas de entrevistas com 14 estudantes da primeira e da quarta séries do Ensino Fundamental do primeiro ciclo de duas escolas municipais; a segunda resultou da aplicação de um teste com os mesmos estudantes em que eles deveriam escrever palavras sujeitas às variações que constituíram objeto da pesquisa. Os resultados sinalizaram para a ocorrência diminuta, na segunda amostra, dos fatos variáveis na fala espontânea dos estudantes em foco. / Salvador

Resumo: O foco principal deste trabalho é a análise comparativa do desempenho qualitativo da pronúncia de vinte cantores argentinos, por meio do uso da Tabela Fonética apresentada no artigo "PB cantado - normas para a pronúncia do português brasileiro no canto erudito", (KAYAMA et al. 2007), com e sem o auxílio de uma pista áudio visual em DVD, como um elemento de ensino e padronização da pronúncia e dicção do português brasileiro cantado para cantores argentinos. Foi realizada uma análise comparativa qualitativa de vinte gravações da canção "Cantar", do ciclo Seis Poemas de Helena Kolody para canto e piano, do compositor Henrique de Curitiba. Método: Os vinte cantores argentinos natos, distribuídos aleatoriamente em dois grupos, Fênix e Pégasus, receberam basicamente o mesmo material diferindo apenas quanto ao DVD entregue ao primeiro grupo, gravado em áudio e vídeo com os sons da Tabela Fonética do Português Brasileiro Cantado, realizada pelo autor deste estudo. Os cantores foram gravados em áudio e vídeo, cantando a canção em questão e esta gravação serviu como referência a seis avaliadores, doutores em suas áreas de atuação (canto, fonoaudiologia e linguística), na realização de uma análise qualitativa da dicção dos cantores argentinos, atribuindo escores à dicção das palavras cantadas pelos cantores deste estudo. Os resultados da avaliação demonstraram que os cantores do grupo Fênix (os quais receberam a pista áudio visual) tiveram uma pronúncia mais próxima do PB atingindo um escore de 77,8% neste aspecto, contra 22,2% do grupo Pégasus (cantores que não receberam a pista áudio visual). 85% dos cantores consideraram a tabela fonética do PB apenas impressa uma ferramenta insuficiente à compreensão dos sons do PB cantado e 15% consideraram a tabela suficiente. 78,6% dos cantores avaliados com a pronúncia mais próxima do ... (Resumo completo, clicar acesso eletrônico abaixo) / Abstract: The main focus in this study is the evaluation of the effectiveness of the Phonetic Chart presented in the article "PB Cantado - Normas Para a Pronúncia do Português Brasileiro no Canto Erudito" (KAYAMA et al, 2007), as an element of education and standardization of Brazilian Portuguese as sung for Argentinean singers, through comparative analysis of twenty recordings of the song "Cantar" from the cycle Seis Poemas de Helena Kolody for voice and piano, by the composer Henrique de Curitiba. Method: The twenty Argentine singers, randomly divided into two groups, Fênix and Pegasus, received basically the same material, the only difference being a DVD given to the first group, with the recording (both audio and video) of the Phonetic Chart of Brazilian Portuguese as Sung, recorded by the author of this study. The singers were videotaped singing the song in question. This tape served as a reference for the six evaluators, doctors in their areas (singing, speech therapy, linguistics), to make a qualitative analysis of the singers' diction, attributing scores for diction of the words of the song as sung by the singers who participated in this study. The results of the evaluation showed that the singers of Fênix group (which received the audiovisual track) had pronunciation more in keeping with PB (Brazilian Portuguese), reaching a 77,8% value at this point. The Pégasus group achieved a 22,2% score (singers that did not receive an audiovisual track). 85% of the singers considered that a printed version of the PB Phonetic Chart was not enough as a tool to understand the sounds of Brazilian Portuguese as Sung, and 15% considered the opposite. 78.6% of the singers evaluated with the closest pronunciation of Brazilian Portuguese were from the Fênix group. Thus, the use of audiovisual resources of the Phonetic Chart of the Brazilian Portuguese as Sung was considered essential to correct ... (Complete abstract click electronic access below) / Orientador: Martha Herr / Coorientador: Silvia Pinho / Banca: Marta de Andrada e Silva / Banca: Luciano Simões Silva / Mestre

Made available in DSpace on 2014-06-11T19:27:27Z (GMT). No. of bitstreams: 0 Previous issue date: 2010-06-25Bitstream added on 2014-06-13T19:35:38Z : No. of bitstreams: 1 pinheiro_ab_me_ia.pdf: 4234999 bytes, checksum: 6f667254e0a2c0de85410c74d0be4d00 (MD5) / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) / O foco principal deste trabalho é a análise comparativa do desempenho qualitativo da pronúncia de vinte cantores argentinos, por meio do uso da Tabela Fonética apresentada no artigo “PB cantado – normas para a pronúncia do português brasileiro no canto erudito”, (KAYAMA et al. 2007), com e sem o auxílio de uma pista áudio visual em DVD, como um elemento de ensino e padronização da pronúncia e dicção do português brasileiro cantado para cantores argentinos. Foi realizada uma análise comparativa qualitativa de vinte gravações da canção “Cantar”, do ciclo Seis Poemas de Helena Kolody para canto e piano, do compositor Henrique de Curitiba. Método: Os vinte cantores argentinos natos, distribuídos aleatoriamente em dois grupos, Fênix e Pégasus, receberam basicamente o mesmo material diferindo apenas quanto ao DVD entregue ao primeiro grupo, gravado em áudio e vídeo com os sons da Tabela Fonética do Português Brasileiro Cantado, realizada pelo autor deste estudo. Os cantores foram gravados em áudio e vídeo, cantando a canção em questão e esta gravação serviu como referência a seis avaliadores, doutores em suas áreas de atuação (canto, fonoaudiologia e linguística), na realização de uma análise qualitativa da dicção dos cantores argentinos, atribuindo escores à dicção das palavras cantadas pelos cantores deste estudo. Os resultados da avaliação demonstraram que os cantores do grupo Fênix (os quais receberam a pista áudio visual) tiveram uma pronúncia mais próxima do PB atingindo um escore de 77,8% neste aspecto, contra 22,2% do grupo Pégasus (cantores que não receberam a pista áudio visual). 85% dos cantores consideraram a tabela fonética do PB apenas impressa uma ferramenta insuficiente à compreensão dos sons do PB cantado e 15% consideraram a tabela suficiente. 78,6% dos cantores avaliados com a pronúncia mais próxima do... / The main focus in this study is the evaluation of the effectiveness of the Phonetic Chart presented in the article PB Cantado – Normas Para a Pronúncia do Português Brasileiro no Canto Erudito (KAYAMA et al, 2007), as an element of education and standardization of Brazilian Portuguese as sung for Argentinean singers, through comparative analysis of twenty recordings of the song Cantar from the cycle Seis Poemas de Helena Kolody for voice and piano, by the composer Henrique de Curitiba. Method: The twenty Argentine singers, randomly divided into two groups, Fênix and Pegasus, received basically the same material, the only difference being a DVD given to the first group, with the recording (both audio and video) of the Phonetic Chart of Brazilian Portuguese as Sung, recorded by the author of this study. The singers were videotaped singing the song in question. This tape served as a reference for the six evaluators, doctors in their areas (singing, speech therapy, linguistics), to make a qualitative analysis of the singers' diction, attributing scores for diction of the words of the song as sung by the singers who participated in this study. The results of the evaluation showed that the singers of Fênix group (which received the audiovisual track) had pronunciation more in keeping with PB (Brazilian Portuguese), reaching a 77,8% value at this point. The Pégasus group achieved a 22,2% score (singers that did not receive an audiovisual track). 85% of the singers considered that a printed version of the PB Phonetic Chart was not enough as a tool to understand the sounds of Brazilian Portuguese as Sung, and 15% considered the opposite. 78.6% of the singers evaluated with the closest pronunciation of Brazilian Portuguese were from the Fênix group. Thus, the use of audiovisual resources of the Phonetic Chart of the Brazilian Portuguese as Sung was considered essential to correct ... (Complete abstract click electronic access below)