Dans ce travail de recherche nous avons abordé la question de la perception et de la production des voyelles orales du français par seize enfants tunisiens (8-9 ans) néo-apprenants du français à trois moments différents de leur apprentissage : au début de leur première année de français, après 9 mois d’apprentissage et suite à une période d’entraînements phonétiques. Les résultats obtenus montrent que (1) la perception et la production des voyelles sont fortement influencées par la L1 des apprenants aussi bien au début qu’après 9 mois d’apprentissage ; (2) les performances en perception et en production des néo-apprenants dépendent plus des voyelles elles-mêmes que de leur statut en L2 par rapport à la L1 (voyelles nouvelles, similaires, ou identiques) ; (3) les entraînements phonétiques que nous avons administrés n’apportent pas de bénéfice sur la perception et la production des voyelles orales françaises par les néo-apprenants tunisien. / In this thesis, we studied the perception and production of the French oral vowels by sixteen Tunisian children (8-9 years old) learning French. Data are collected at three different periods: at the beginning of their first year of learning, after nine months of learning, and after a period of phonetic training. Main results are: (1) the perception and production of the vowels are greatly influenced by the L1 of the learners, both at the beginning and after nine months of learning; (2) the perception and production performances of the learners are better predicted by the identity of the vowel rather than its status in L2 compared to L1 (new, similar or identical vowels); (3) the phonetic training we gave showed no benefit on the perceptual or production performances of the children.

Cette étude acoustico-perceptive concerne les limites de la perception et de la production des voyelles orales du français [i, e, ɛ, a, u, o, ɔ, y, ø, œ], en isolation et en contextes consonantiques divers, chez dix tchécophones, futures enseignantes de Français Langue Étrangère (FLE). Les résultats montrent que (1) La maîtrise phonétique des voyelles dépend de leurs graphies et de l’entourage consonantique. (2) Les voyelles fermées [i, y, u] et le [a] sont globalement maîtrisées avec authenticité. (3) Les capacités de perception des contrastes entre les voyelles moyennes e/ɛ, ø/œ et o/ɔ ainsi que leur production sont limitées. Ces résultats ne sont que partiellement en accord avec les prédictions établies à partir du Speech Learning Model (SLM) de Flege (1995), basé sur la notion de similarité phonétique qui existe entre la langue maternelle (LM) et la langue étrangère (LE). / This acoustic-perceptual study concerns the limits of perception and production of French oral vowels [i, e, ɛ, a, u, o, ɔ, y, ø, œ], in isolation and in different consonantal contexts, in ten pre-service Czech teachers of French as a Foreign Language (FFL). The results show that (1) Phonetic proficiency in vowels depends on their spellings and consonantal context. (2) Vowels [i, y, u] and [a] are generally mastered with authenticity. (3) The ability to hear contrasts between the vowels e/ɛ, ø/œ and o/ɔ and pronounce them is limited. These results are only partially consistent with the predictions established in the Speech Learning Model (SLM) by Flege (1995), based on the notion of phonetic similarity between the mother tongue (MT) and the foreign language (FL).

Cette étude acoustico-perceptive concerne les limites de la perception et de la production des voyelles orales du français [i, e, ɛ, a, u, o, ɔ, y, ø, œ], en isolation et en contextes consonantiques divers, chez dix tchécophones, futures enseignantes de Français Langue Étrangère (FLE). Les résultats montrent que (1) La maîtrise phonétique des voyelles dépend de leurs graphies et de l’entourage consonantique. (2) Les voyelles fermées [i, y, u] et le [a] sont globalement maîtrisées avec authenticité. (3) Les capacités de perception des contrastes entre les voyelles moyennes e/ɛ, ø/œ et o/ɔ ainsi que leur production sont limitées. Ces résultats ne sont que partiellement en accord avec les prédictions établies à partir du Speech Learning Model (SLM) de Flege (1995), basé sur la notion de similarité phonétique qui existe entre la langue maternelle (LM) et la langue étrangère (LE). / This acoustic-perceptual study concerns the limits of perception and production of French oral vowels [i, e, ɛ, a, u, o, ɔ, y, ø, œ], in isolation and in different consonantal contexts, in ten pre-service Czech teachers of French as a Foreign Language (FFL). The results show that (1) Phonetic proficiency in vowels depends on their spellings and consonantal context. (2) Vowels [i, y, u] and [a] are generally mastered with authenticity. (3) The ability to hear contrasts between the vowels e/ɛ, ø/œ and o/ɔ and pronounce them is limited. These results are only partially consistent with the predictions established in the Speech Learning Model (SLM) by Flege (1995), based on the notion of phonetic similarity between the mother tongue (MT) and the foreign language (FL).

Atualmente desenvolvem-se técnicas para a análise, identificação e o reconhecimento da fala. As mais eficientes mostram-se matematicamente complicadas, baseadas em análise estatísticas de dados, o que torna o sistema moroso, necessitando uma grande quantidade de dados para amostras. Este trabalho tem como objetivo apresentar a possibilidade do uso de Estruturas Neurais Artificiais Paraconsistentes no aprendizado e reconhecimento de sinais de fala, independentemente de análise estatística, ou número de amostras. A partir de um estudo piloto, identificou-se a necessidade de um aprofundamento no estudo dos Traços Formantes dos Fones. Com os Formantes dos Fones pode-se criar um sistema capaz de reconhecer sons produzidos em qualquer língua, pelas combinações da produção de sons através da emissão simultânea de um conjunto de Formantes. Como possível solução para a identificação dos Formantes dos Fones propõe-se neste trabalho a criação do conceito de Constelação Fônica, que consiste no reconhecimento de combinações de características matemáticas identificadas nos sinais sonoros de fala. Como uma forma de reconhecer estas Constelações, apresentam-se as Redes Neurais Artificiais Paraconsistentes, eficientes no reconhecimento de padrões por proximidade e com capacidade para tratamento de sinais contraditórios e paracompletos. Para a viabilização desta solução, criou-se um Programa de Computador (Sistema de Análise da Produção da Fala - SIAPF) capaz de promover os tratamentos necessários em um sinal falado, gerando assim a sua Constelação Fônica e sua respectiva Rede Neural Artificial Paraconsistente. A partir da Rede Neural Artificial Paraconsistente correspondente ao sinal em questão, pode-se estudá-lo de uma forma interpretativa, com menor acumulo de dados e tratamentos estatísticos do que com as ferramentas tradicionais. O SIAPF passa a ser mais uma ferramenta para análise de produção de fala, viabilizando a criação de novas formas de medidas quantitativas e normatizadas para os Formantes da Fala. / Currently, many diferent techniques are developing for the analysis identification and recognition of speech. The most efficient are shown mathematically complicated, based on analysis statistical of data, that actually makes systems very slow, creating the nessecity for a great amount of data for samples. The aim of this work is to present the possibility of the use of Paraconsistent Artificial Neural Network Structures, in the learnig and recognition of speech signals, independent of statistical analysis, or the number of samples. Starting from a pilot study, the need of a deep study of the Formants of Phones was identified. With the Formants of Phones, a system capable to recognize sounds produced in any language, through conbinations of the production of sounds and simultaneous emission of a group of Formants was created. A possible solution for the identification of Formants of Phones proposed in this work is the creation of Phonic Constellation concept, witch uses the recognition of combinations of identified mathematical characteristcs in the resoant signal of speech. As a form of recognizing these Constellation, this work present the Paraconsistent Artificial Neural Network, efficient in the recognition of patterns for proximity and with capacity to process contractory and paracomplet signals. To develop this solution a Computer Program (System of Analysis of the Production of Speech - SIAPF) capable to process the necessary calculations of a speech signal, generating its Phonic Constelattion and its respective Paraconsistent Artificial Neural Network was created. The subject signal can be studied in an interpretative way using the Paraconsistent Artificial Neural Network with a minimal accumulation of data and statistical calculations compared to traditional tools. The SIAPF is another new tool for analysis of speech production, creating new ways for quantitative and standard measures of Formants of Speech.

Atualmente desenvolvem-se técnicas para a análise, identificação e o reconhecimento da fala. As mais eficientes mostram-se matematicamente complicadas, baseadas em análise estatísticas de dados, o que torna o sistema moroso, necessitando uma grande quantidade de dados para amostras. Este trabalho tem como objetivo apresentar a possibilidade do uso de Estruturas Neurais Artificiais Paraconsistentes no aprendizado e reconhecimento de sinais de fala, independentemente de análise estatística, ou número de amostras. A partir de um estudo piloto, identificou-se a necessidade de um aprofundamento no estudo dos Traços Formantes dos Fones. Com os Formantes dos Fones pode-se criar um sistema capaz de reconhecer sons produzidos em qualquer língua, pelas combinações da produção de sons através da emissão simultânea de um conjunto de Formantes. Como possível solução para a identificação dos Formantes dos Fones propõe-se neste trabalho a criação do conceito de Constelação Fônica, que consiste no reconhecimento de combinações de características matemáticas identificadas nos sinais sonoros de fala. Como uma forma de reconhecer estas Constelações, apresentam-se as Redes Neurais Artificiais Paraconsistentes, eficientes no reconhecimento de padrões por proximidade e com capacidade para tratamento de sinais contraditórios e paracompletos. Para a viabilização desta solução, criou-se um Programa de Computador (Sistema de Análise da Produção da Fala - SIAPF) capaz de promover os tratamentos necessários em um sinal falado, gerando assim a sua Constelação Fônica e sua respectiva Rede Neural Artificial Paraconsistente. A partir da Rede Neural Artificial Paraconsistente correspondente ao sinal em questão, pode-se estudá-lo de uma forma interpretativa, com menor acumulo de dados e tratamentos estatísticos do que com as ferramentas tradicionais. O SIAPF passa a ser mais uma ferramenta para análise de produção de fala, viabilizando a criação de novas formas de medidas quantitativas e normatizadas para os Formantes da Fala. / Currently, many diferent techniques are developing for the analysis identification and recognition of speech. The most efficient are shown mathematically complicated, based on analysis statistical of data, that actually makes systems very slow, creating the nessecity for a great amount of data for samples. The aim of this work is to present the possibility of the use of Paraconsistent Artificial Neural Network Structures, in the learnig and recognition of speech signals, independent of statistical analysis, or the number of samples. Starting from a pilot study, the need of a deep study of the Formants of Phones was identified. With the Formants of Phones, a system capable to recognize sounds produced in any language, through conbinations of the production of sounds and simultaneous emission of a group of Formants was created. A possible solution for the identification of Formants of Phones proposed in this work is the creation of Phonic Constellation concept, witch uses the recognition of combinations of identified mathematical characteristcs in the resoant signal of speech. As a form of recognizing these Constellation, this work present the Paraconsistent Artificial Neural Network, efficient in the recognition of patterns for proximity and with capacity to process contractory and paracomplet signals. To develop this solution a Computer Program (System of Analysis of the Production of Speech - SIAPF) capable to process the necessary calculations of a speech signal, generating its Phonic Constelattion and its respective Paraconsistent Artificial Neural Network was created. The subject signal can be studied in an interpretative way using the Paraconsistent Artificial Neural Network with a minimal accumulation of data and statistical calculations compared to traditional tools. The SIAPF is another new tool for analysis of speech production, creating new ways for quantitative and standard measures of Formants of Speech.

The human voice might have been shaped by sexual selection. Hence, voice fundamental (F0, or pitch) and formant frequencies (Fn, or timbre) are proposed to convey fitness cues germane to rivals and potential mates. First, I confirm the independent effects of F0 and Fn on listeners’ assessments of speaker size, masculinity, and attractiveness. Second, I quantify the just-noticeable differences in both vocal features and then place F0 and Fn cues in conflict by equally discriminable amounts to test their relative influence on such voice-based social judgments. Results revealed a greater relative role of Fn in listeners’ ratings of all three dimensions, suggesting that these dimensions might all be cued more reliably by Fn than F0. Alternatively, given post-hoc principal component analyses that revealed considerable overlap in ratings of size, masculinity, and attractiveness, listeners’ conceptions of these dimensions may not be independent despite a research tradition that assumes they are. / xi, 102 leaves : ill. ; 29 cm

(English) The dissertation explores the vowel system of Czech L3 Spanish, focusing on three facets of its architecture: vowel quality (following SLM by Flege 1995, 1999, 2003), vowel quantity and vocalic sequences between words. Preliminary studies (Čechová 2013, 2014) suggest that there might be some evidence for the Mechanism of Equivalence in Czech L3 Spanish, since the vowel spaces of both languages consist of the same categories, with different phonetic realizations. Hence, the more similar the sounds are, the harder it is to capture the relevant difference, necessary to attain native-like pronunciation. Vowel quantity, being traditionally associated with the phonological feature of length in Czech (Palková 1994), is not present in Spanish, and compensating for that property, some Czech speakers tend to pronounce accented syllables with extraordinarily longer durations. Finally, vowel sequences in Spanish are usually subjetct to resyllabification, whereas Czech prefers glottalization to keep morphemes of words separated. These predictions were tested in 22 university Czech students with advanced level of Spanish (C1-C1). Subsequent analysis revealed consistent inclination towards L1 in terms of vowel quality, in less extent in vowel quantity, and although the prevalecent strategy for majority...

The current study investigated vowel quality in Miami Cuban Spanish, looking specifically at the influence of English on the Spanish vowel system. The vowel production of eleven Miami Cubans from three generations is investigated. Subjects include six males and five females. Three different elicitation instruments were used. The first was a brief sociolinguistic interview, the second was a story that the participants were asked to read aloud. Carrier words were embedded into the story in order to obtain multiple samples of both stressed and unstressed vowels. For the third instrument, subjects were asked to read a list of words with careful attention to their pronunciation. The reading list contained the same words that were embedded in the story of the first task. These three instruments were repeated in both Spanish and in English in order to investigate possible English influence in the Spanish vowel system of these bilinguals.

The mismatch negativity (MMN) response, a passively-elicited component of the auditory event-related potential (ERP), reflects preattentive identification of infrequent changes in acoustic stimuli. In the current study, the MMN response was examined closely to determine what extent natural speech sounds evoke the MMN. It was hypothesized that a significant MMN response results during the presentation of deviant stimuli from which spectral energy within formant bands critical to vowel identification has been removed. Localizations of dipoles within the cortex were hypothesized to yield information pertaining to the processing of formant-specific linguistic information. A same/different discrimination task was administered to 20 adult participants (10 female and 10 male) between the ages of 18 and 26 years. Data from behavioral responses and ERPs were recorded. Results demonstrated that the MMN may be evoked by natural speech sounds. Grand-averaged brain maps of ERPs created for all stimulus pairs showed a large preattentive negativity. Additionally, amplitudes of the MMN were greatest for pairs of auditory stimuli in which spectral energy not corresponding to formant frequencies was digitally eliminated. Dipoles reconstructed from temporal ERP data were located in cortical areas known to support language and auditory processing. Significant differences between stimulus type and reaction time were also noted. The current investigation confirms that the MMN response is evoked by natural speech sounds and provides evidence for a theory of preattentive formant-based processing of speech sounds.

Questions of speaker’s recognition by voice are investigated in this dissertation. Speaker recognition systems, their evolution, problems of recognition, systems of features, questions of speaker modeling and matching used in text-independent and text-dependent speaker recognition are considered too. The text-independent speaker recognition system has been developed during this work. The Gaussian mixture model approach was used for speaker modeling and pattern matching. The automatic method for voice activity detection was proposed. This method is fast and does not require any additional actions from the user, such as indicating patterns of the speech signal and noise. The system of the features was proposed. This system consists of parameters of excitation source (glottal) and parameters of the vocal tract. The fundamental frequency was taken as an excitation source parameter and four formants with three antiformants were taken as parameters of the vocal tract. In order to equate dispersions of the formants and antiformants we propose to use them in mel-frequency scale. The standard mel-frequency cepstral coefficients (MFCC) for comparison of the results were implemented in the recognition system too. These features make baseline in speech and speaker recognition. The experiments of speaker recognition have shown that our proposed system of features outperformed standard mel-frequency cepstral coefficients. The equal error rate (EER) was equal to 5.17% using proposed... [to full text] / Disertacijoje nagrinėjami kalbančiojo atpažinimo pagal balsą klausimai. Aptartos kalbančiojo atpažinimo sistemos, jų raida, atpažinimo problemos, požymių sistemos įvairovė bei kalbančiojo modeliavimo ir požymių palyginimo metodai, naudojami nuo ištarto teksto nepriklausomame bei priklausomame kalbančiojo atpažinime. Darbo metu sukurta nuo ištarto teksto nepriklausanti kalbančiojo atpažinimo sistema. Kalbėtojų modelių kūrimui ir požymių palyginimui buvo panaudoti Gauso mišinių modeliai. Pasiūlytas automatinis vokalizuotų garsų išrinkimo (segmentavimo) metodas. Šis metodas yra greitai veikiantis ir nereikalaujantis iš vartotojo jokių papildomų veiksmų, tokių kaip kalbos signalo ir triukšmo pavyzdžių nurodymas. Pasiūlyta požymių vektorių sistema, susidedanti iš žadinimo signalo bei balso trakto parametrų. Kaip žadinimo signalo parametras, panaudotas žadinimo signalo pagrindinis dažnis, kaip balso trakto parametrai, panaudotos keturios formantės bei trys antiformantės. Siekiant suvienodinti žemesnių bei aukštesnių formančių ir antiformančių dispersijas, jas pasiūlėme skaičiuoti melų skalėje. Rezultatų palyginimui sistemoje buvo realizuoti standartiniai požymiai, naudojami kalbos bei asmens atpažinime – melų skalės kepstro koeficientai (MSKK). Atlikti kalbančiojo atpažinimo eksperimentai parodė, kad panaudojus pasiūlytą požymių sistemą buvo gauti geresni atpažinimo rezultatai, nei panaudojus standartinius požymius (MSKK). Gautas lygių klaidų lygis, panaudojant pasiūlytą požymių... [toliau žr. visą tekstą]