abstract: Audio signals, such as speech and ambient sounds convey rich information pertaining to a user’s activity, mood or intent. Enabling machines to understand this contextual information is necessary to bridge the gap in human-machine interaction. This is challenging due to its subjective nature, hence, requiring sophisticated techniques. This dissertation presents a set of computational methods, that generalize well across different conditions, for speech-based applications involving emotion recognition and keyword detection, and ambient sounds-based applications such as lifelogging. The expression and perception of emotions varies across speakers and cultures, thus, determining features and classification methods that generalize well to different conditions is strongly desired. A latent topic models-based method is proposed to learn supra-segmental features from low-level acoustic descriptors. The derived features outperform state-of-the-art approaches over multiple databases. Cross-corpus studies are conducted to determine the ability of these features to generalize well across different databases. The proposed method is also applied to derive features from facial expressions; a multi-modal fusion overcomes the deficiencies of a speech only approach and further improves the recognition performance. Besides affecting the acoustic properties of speech, emotions have a strong influence over speech articulation kinematics. A learning approach, which constrains a classifier trained over acoustic descriptors, to also model articulatory data is proposed here. This method requires articulatory information only during the training stage, thus overcoming the challenges inherent to large-scale data collection, while simultaneously exploiting the correlations between articulation kinematics and acoustic descriptors to improve the accuracy of emotion recognition systems. Identifying context from ambient sounds in a lifelogging scenario requires feature extraction, segmentation and annotation techniques capable of efficiently handling long duration audio recordings; a complete framework for such applications is presented. The performance is evaluated on real world data and accompanied by a prototypical Android-based user interface. The proposed methods are also assessed in terms of computation and implementation complexity. Software and field programmable gate array based implementations are considered for emotion recognition, while virtual platforms are used to model the complexities of lifelogging. The derived metrics are used to determine the feasibility of these methods for applications requiring real-time capabilities and low power consumption. / Dissertation/Thesis / Doctoral Dissertation Electrical Engineering 2015

abstract: The purpose of this study was to identify acoustic markers that correlate with accurate and inaccurate /r/ production in children ages 5-8 using signal processing. In addition, the researcher aimed to identify predictive acoustic markers that relate to changes in /r/ accuracy. A total of 35 children (23 accurate, 12 inaccurate, 8 longitudinal) were recorded. Computerized stimuli were presented on a PC laptop computer and the children were asked to do five tasks to elicit spontaneous and imitated /r/ production in all positions. Files were edited and analyzed using a filter bank approach centered at 40 frequencies based on the Mel-scale. T-tests were used to compare spectral energy of tokens between accurate and inaccurate groups and additional t-tests were used to compare duration of accurate and inaccurate files. Results included significant differences between the accurate and inaccurate productions of /r/, notable differences in the 24-26 mel bin range, and longer duration of inaccurate /r/ than accurate. Signal processing successfully identified acoustic features of accurate and inaccurate production of /r/ and candidate predictive markers that may be associated with acquisition of /r/. / Dissertation/Thesis / Masters Thesis Communication Disorders 2017

A diadococinesia (DDC) é a habilidade para realizar repetições rápidas de padrões relativamente simples de contrações musculares opostas, utilizada para avaliar a maturação e a integração neuromotora. A DDC oral e laríngea, associada aos demais procedimentos de avaliação fonoaudiológica, é um importante recurso na compreensão das manifestações dos distúrbios da comunicação. A partir disso, objetivou-se estabelecer valores de referência quanto à normalidade em relação aos resultados da avaliação da DDC oral e laríngea nos diferentes gêneros e faixas etárias de crianças falantes do português brasileiro, bem como analisar a diferença entre os gêneros e faixas etárias. Participaram 150 crianças, distribuídas nas faixas de oito, nove e dez anos de idade. A DDC oral foi avaliada por meio da repetição de \"pa\", ta\", \"ca\" e \"pataca\" e a DDC laríngea, pela repetição de \"a\" e \"i\". Foram utilizados os programas Motor Speech Profile Advanced e Mult Speech Main Program, da Kay Elemetrics Corp. Os parâmetros da DDC foram apresentados como média, mediana e percentil para cada emissão. A comparação entre gênero e idade foi realizada por meio da Análise de Variância a dois critérios e do teste de Tuckey. Quanto à DDC oral, a análise estatística dos resultados demonstrou que, com o avanço da idade: houve aumento do número de emissões de monossílabas por segundo, redução do tempo médio entre essas emissões; houve aumento do coeficiente de variação do período durante a sílaba \"ca\" e aumento do coeficiente de variação do pico da intensidade para a sílaba \"ta\". O número de emissões por segundo da monossílaba \"ta\" foi maior para as meninas que para os meninos. Na emissão da trissílaba, o número de emissões por segundo foi diferente entre os gêneros e, considerando-se os subgrupos de idade e gênero, as meninas de oito anos apresentaram menor número de emissões que todos os demais subgrupos. Quanto à DDC laríngea, com o avanço da idade houve aumento do número de emissões por segundo e períodos mais curtos da vogal \"i\" para as meninas; menor valor do desvio padrão do período e da perturbação do período para essa mesma vogal. Conclui-se que foi possível estabelecer os valores de normalidade da DDC oral e laríngea para o grupo de crianças estudado e que houve diferenças quanto ao gênero e à idade, o que demonstra que o desenvolvimento da DDC oral e laríngea deve ser considerado na avaliação da comunicação oral de crianças. / Diadochokinesis (DDK) is the ability to perform fast repetitions of relatively simple patterns of opposite muscle contractions and it is employed for the evaluation of the neuromotor maturation and integration. The oral and laryngeal DDK, associated with other procedures for the speech evaluation, are important resources in the understanding of communication disorders. Thus, this study was conducted to establish reference values of normality of the outcomes of oral and laryngeal DDK for the different genders and age ranges of Brazilian Portuguese-speaking children, as well as to analyze the presence of difference between genders and among age ranges. The study sample was composed of 150 children aged 8, 9 and 10 years. The oral DDK was evaluated by repetition of \"pa\", \"ta\", \"ka\" and \"pataka\", and laryngeal DDK was assessed by repetition of \"a\" and \"i\". The softwares Motor Speech Profile Advanced and Mult Speech Main Program, of Kay Elemetrics Corp, were employed analysis. The DDK parameters were presented as mean, median and percentile for each emission. Comparison among genders and age ranges was performed by twoway analysis of variance and the Tukey test. Statistical analysis of oral DDK revealed with an increase in age: there were an increase in the number of emissions of monosyllables per second, a reduced mean time between emissions; an increase in the coefficient of variation of the period during the syllable \"ka\", and an increase in the coefficient of variation of the peak intensity for the syllable \"ta\". The number of emissions of the syllable \"ta\" per second was higher for females than to males. In the trisyllabic emission, the number of emissions per second was different among the genders. Besides, regarding the subgroups of age and gender, the 8 year-old-girls showed a decreased number of emissions than the other subgroups. With regard to laryngeal DDK, there was an increased number of emissions per second and shorter periods of vowel \"i\" for females with the increase in age; there was also a smaller standard deviation and perturbations of the period on for this same vowel. It was concluded that it was possible to establish values of normality of oral and laryngeal DDK for the group of children investigated; and that there were differences as to gender and age, which demonstrates that the development of oral and laryngeal DDK should be considered in the evaluation of oral communication of children.

Quando a cidade de Corumbá foi fundada no século XVIII demorou muito tempo para encontrar sua vocação econômica. Assim, dentro das referências revisadas para a elaboração deste trabalho desta tese, não houve constatação de uma vocação econômica específica que tenha relevância pelo menos até os 50 primeiros anos do século XIX, só a partir daí, é que a economia corumbaense teve um rumo, quando prevalece o comércio para a extração de excedente e na qual existe uma articulação devido à navegação e a comunicação com os portos principais da Bacia do Prata. Fato diferente ao acontecido com a cidade logo após o conflito bélico. A delimitação temporal estabelecida para este projeto de pesquisa obedece a alguns critérios adotados para sua definição. A delimitação inicial do período, 1870, se justifica em razão do fim da guerra da Tríplice Aliança, evento este que teve uma relevância marcante para esta região do Brasil, porque foi ocupada pelas tropas paraguaias deixando um rastro de destruição e desolação por toda Corumbá, isso ocorreu justamente no momento em que a cidade estava definindo a sua vocação econômica. Já a escolha do ano de 1914 se deve a alguns fatos como a chegada ao Pantanal a estrada de ferro Noroeste do Brasil, que à época se estendia até as margens do Rio Paraguai, a 70 quilômetros de Corumbá. O clima pantaneiro favorece uma adaptação à atividade pecuária, assim a introdução do gado no início do século XVIII, a atividade pecuária encontrou no sul de Mato Grosso as condições climáticas e ecológicas propícias à sua reprodução e proporcionou a fixação da população em torno das grandes fazendas de criação. Desta maneira, a grande parte riqueza desta parte do Mato Grosso foi o gado, base de sua economia no início do povoamento e também foi o fator de articulação da economia incipiente no Pantanal em Corumbá, quando a atividade agrícola ainda era restrita à zona litoral do país. Esta articulação consiste na ligação com outros setores econômicos. Desta forma, o setor primário consiste na própria exploração da atividade pecuária. No setor secundário estavam as charqueadas que, apesar de não apresentarem uma transformação completa da matéria-prima em outro produto, propiciam agregação de valor à carne. E no setor terciário, a distribuição dos produtos que se daria, num primeiro momento, pela via fluvial e mais adiante, pela ferroviária nos principais centros consumidores. Outros produtos passaram pelo porto corumbaense, mas o mais importante foi borracha, ambos comercializados tanto visando o mercado externo, assim como a importação de produtos para toda a região do Mato Grosso. / When the city of Corumba was founded in the 18th century, it took a long time for finding its economic vocation. Thus, among the references herein revised, we found no evidence of a specific and relevant vocation at least until the first 50 years of the 19th century. From that moment on, the corumbaense economy took a direction towards the commerce of the extractions, articulated by the navigation growth and the communication among the main harbors of the Prata river basin, what differs from what happened to the city right after the war. The temporal limits herein established followed some criteria and the starting point of 1870 was chosen for marking the end of the War of Triple Alliance, whose event was strongly relevant for this Brazilian region because the troops occupation of the territories brought together a trail of destruction and desolation all over Corumba city, what happened when the economic vocation was being chosen. The year of 1914 closed the period of research. It coincided with the arrival of the railroad Noroeste do Brasil [Northeast of Brazil], which, at that time, reached the margins of Paraguay river, 70 km away from Corumba. The pantanalian climate favored the adaptation of livestock activity, which dated the beginning of the 18th century, especially in south of the Mato Grosso state, whose climate and ecologic conditions were propitious to reproduction, proportioning the formation of a new villages around the big livestock farms. Thus, the great wealth of this part of the state was based on livestock, which was the basis of the economy during the population settlement and were the main economic factor of the in both Pantanal and Corumba regions in a period which the agriculture was still restricted to the coastal areas of Brazil,whose articulation was bonded to other economic sectors. Thus, the primary sector consisted on the livestock itself. The secondary sector was formed by the charqueadas [area destined to jerk beef maturation] which, although showed no modification on the raw materials, added financial value to the meat. The tertiary sector was the distribution of the products to consume centers, firstly done by waterways and later by railways. Other products were commercialized on corumbaense harbor the most important one was the rubber and their destination were both the foreign market and the importation around Mato Grosso state.

Interspeech posture (ISP) is a term used to define the position of a person’s articulators when they are preparing to speak. Research suggests that ISP may be representative of a speaker’s phonological knowledge in a particular language, as determined empirically with ultrasound measures of the tongue in English-French bilinguals (Wilson & Gick, 2014). It is possible, therefore, that measuring ISP could be a diagnostic tool for determining phonological knowledge in bilingual speakers. However, more information on ISP in typical adult bilingual speakers is needed before diagnostic claims can be made. For example, ISP is believed to be language specific, and the typical ISP for each language must be determined. Therefore, the purpose of this study was to extend the research by Wilson and Gick (2014) to investigate ISP in Spanish-English speaking adults. To this end, 13 bilingual Spanish-English adults were asked to produce 30 sentences while speaking in monolingual and bilingual modes. While they were speaking, ultrasound images of the oral cavity were obtained by placing a probe sub-mentally and analyzing the position of the tongue using Articulate Assistant Advanced 2.0 software (Articulate Instruments, 2012). Tongue and palate contour measurements were made by using a curved tongue spline that was manually drawn and semi-automatically fit to each speaker’s tongue/palate contour. ISP was measured using the participant’s tongue tip height along a reference angle from the probe to the alveolar ridge. Additionally, monolingual English speaking adults were asked to rate the accentedness of each bilingual’s speech in English as a behavioral correlate of language proficiency. Overall results of this study were non-significant; bilingual Spanish-English speakers utilized similar postures in monolingual Spanish and English modes, and in bilingual mode, in contrast with the findings of Wilson and Gick (2014). Accentedness ratings in English v indicated that the bilingual speakers were relatively uniform in their lack of accentedness. Although overall results from this study differ from those of Wilson and Gick (2014) a subset of their participants- speakers that were rated as having non-native accents- had similar results in that they also showed no difference in ISP. Related ISP’s across languages may be due to participants having native sounding English but non-native Spanish. Due to contrasting findings from Wilson and Gick (2014), further investigation with accented speakers is needed to determine if distinct ISPs exist for bilingual Spanish-English speakers.

The purpose of this investigation was to examine the utility of nonlinear phonological frameworks for designing and executing an intervention program with phonologically disordered children. Six such children between the ages of 3 and 6 years participated in the study three times a week over three consecutive six-week blocks. The following general questions were addressed: 1. Will nonlinear phonological frameworks help to predict logical and attainable intervention goals for phonologically disordered children? 2. Are the separate prosodic and segmental levels of representation of nonlinear phonology psychologically real? 3. If the 'prosodic tier' has some observable clinical reality, will there be a difference in proportion and rate of syllable/word shapes acquired as a result of intervention methods that contrast the onset and rime versus those that utilize the mora a constituent? 4. If the 'segmental/melodic tier' has some observable independence, is there any advantage to be gained from targeting specified features at 'higher' versus lower' levels in the feature hierarchy in phonemic inventory intervention? An alternating block, mulitiple baseline design (counterbalanced over six single subjects) provided an opportunity to investigate the above questions. Within each six-week block, three week periods were devoted in turn to prosodic (syllable structure) training and segmental training. Prosodic subblocks were divided into two four-session sunblocks to contrast developmental change for targets presented as moraic constituents versus onset-rime constituents. Segmental periods were divided into two four-session subblocks to contrast developmental change for features from higher and lower levels in the feature hierarchy. Analyses during and after the study demonstrated the following with respect to the four research questions: 1. The nonlinear frameworks provided a logical model for deriving attainable intervention goals. All of the children became intelligible by the end of the project as a result of attaining the goals determined by nonlinear phonological theory. 2. Rate of attainment of syllabic and segmental goals differed, with a faster rate of change for syllabic goals overall, suggesting independence of segmental and prosodic tiers, and possible dominance of the prosodic tier. Interactions between tiers were also observed, suggesting that they are interdependent as well as autonomous. 3. Moraic and onset-rime condition quantitative results were virtually equivalent, but some qualitative differences appeared which had relevance for the each of the theories with respect to status of the onset, word-final consonants, and epenthesis. 4. Higher level features in the feature hierarchy tended to be acquired before lower level features. The nonlinear phonological frameworks stimulated a successful intervention study. Evidence gained through this study in turn contributes to the understanding of the nonlinear constructs. / Medicine, Faculty of / Audiology and Speech Sciences, School of / Graduate

Dans cette thèse sera traité le thème de l’intégration et de l’articulation entre urbanisme et transport, avec le but d’évaluer et étudier le sens de l’application d’un plan régional de Transit Oriented Development (TOD) ou d’urbanisme des transports en commun en Nord-Pas-de-Calais. À cet égard nous avons fait le choix d’utiliser l’outil de la modélisation intégrée d’usage du sol et transport et notamment le logiciel de simulation Tranus, pour implémenter un modèle de simulation capable de répondre aux plusieurs questionnements à la base de cette recherche.On propose dans ce texte une perspective intégrée, inclusive et interactive sur les problématiques et les enjeux qui concernent les politiques d’usage du sol et des transports à l’échelle d’une région. Selon une approche multidisciplinaire et multi-échelle, qui suit les principes d’interdépendances entre les nombreux éléments du territoire, que l’on retrouve dans l’urbanisme des réseaux. Il s’agit d’aborder les thématiques de la mobilité et des transports, selon un nouveau paradigme, basé sur les concepts d’accessibilité, de connectivité et de multimodalité et donc selon l’idée de concevoir un urbanisme et un développement non plus auto-centré, mais orientés vers l’usage des transports en commun et des modes de transport non motorisés. La thèse s’inscrit dans le cadre d’un travail de recherche doctorale en aménagement et transport, déroulé à l’Université des Sciences et Technologies de Lille 1 à travers une cotutelle entre le Laboratoire Ville Mobilité et Transport (LVMT – IFSTTAR) et l’Université de Calabre (Italie) et une collaboration scientifique avec le bureau d’étude d’ingénierie Vénézuélien Modelistica. / This thesis discuss the theme of integration and articulation between urban and transportation planning, with the aim of evaluate and studyi the sense and potentialities of the application of a Regional Plan for Transit Oriented Development (TOD) in Nord-Pas-de-Calais. In this regard we have chosen to use the tool of land use and transport integrated modeling (LUTI), and in particular the integrated simulation software Tranus, to implement a model which could answer to several research questions.We propose in this paper an integrated, inclusive and interactive perspective about problems and issues concerning land-use and transport policies at a regional level. A multidisciplinary and multi-scalar approach, following the principles of interdependence between all different elements of the territory, which is found in the concept of "networked city" of Dupuy. The aim is to address the themes of mobility and transport, according to a new paradigm, based on the concepts of accessibility, connectivity and multimodality and therefore according to the idea of an urbanism and a development oriented to transit and non-motorized transport, rather than car-oriented. The thesis is part of a phd research in urban and transportation planning, held at the University of Science and Technology of Lille 1, through a joint supervision between the Laboratoire Ville MObilité et Transport (LVMT - IFSTTAR) and the University of Calabria (Italy) and a scientific collaboration with Venezuelan engineering firm Modelistica.

The purpose of this dissertation was to investigate the influences of auditory information provided by the cochlear implant (CI) on the readings skills of children born with profound deafness. I investigated the relationship of access to the sound signal provided by the CI on a constellation of skills related to word-reading. In a preliminary study, I examined the relationship between the early speech production and perception skills of 72 CI users on later reading skills. Using regression analysis, I found I could explain 59% of the variance of later reading skills by early speech perception and production performance. Secondly, I examined the phonological processing skills of 29 children with prelingual, profound hearing loss with at least 4 years of CI experience. I compared this performance with 29 children with normal hearing, matched with regard to word-reading ability and Socio-Economic-Status. I also compared speech production and perception skills with phonological processing and reading skills. Results revealed that children with CIs were able to complete tasks measuring phonological processing, but there were performance differences between the two groups. Although the children with CIs had mean standard reading achievement standard scores that were about 12 points lower than the children with normal hearing, the mean standard scores for both groups was within the normal range. Finally, a regression analysis revealed that the Phonological Processing skills accounted for 50%, and 75% of the variance in word and paragraph reading scores for all the children. In conclusion early speech perception and production skills of children with profound hearing loss who receive CIs predict future reading achievement skills. Better early speech perception and production skills result in higher reading achievement. Furthermore, the early access to sound helps to build better phonological processing skills, which is one of the likely contributors to eventual reading success. Thus, it is reasonable, possible and important to assess the early speech production perception and subsequent phonological processing in children with profound hearing loss who receive CIs.

Actors often receive training to develop effective strategies for using the voice on stage. Arthur Lessac developed a training approach that concentrated on three energies: structural action, tonal action, and consonant action. Together, these energies help to create a more resonant voice, which is characterized by a fuller sound that carries well over noise and distance. In Lessac-Based Resonant Voice Therapy, voice clinicians help clients achieve a resonant voice through structural posturing and awareness of tonal changes. However, LBRVT does not include the third component of Lessac's approach: consonant action. This study examines the effect that increased consonant energy has on the speaking voice--particularly regarding loudness. Audio samples were collected from eight actor participants who read a monologue using three distinct styles: normal articulation, poor articulation (elicited using a bite block), and over-articulation (elicited using a Lessac-based training intervention). Participants learned about the "consonant orchestra," practiced producing each sound in a consonant cluster word list, and practiced linking the consonants in short phrases. Twenty graduate students of speech-language pathology listened to speech samples from the different conditions, and made comparative judgments regarding articulation, loudness, and projection. Group results showed that the over-articulation condition was selected as having the greatest articulation, loudness, and projection in comparison to the other conditions, although vocal intensity (dB SPL) was not statistically different. These findings indicate that articulation treatment may be beneficial for increasing perceived vocal loudness.

It would be useful to obtain information about social and emotional characteristics in children who are receiving articulation/phonological intervention in order to assess the effectiveness of various treatment approaches from a social/emotional perspective. The purpose of this study was to determine whether or not articulation and phonological intervention influences children's social and emotional characteristics as perceived by their parents and, if so, which articulation approach (traditional vs. phonological cycling) results in more improvement in different domains of social and emotional characteristics. The specific social and emotional characteristics explored in this study are social skills, communication, independence, self-esteem, and domestic responsibility as assessed by the Affective Behavior Scales for the Disabled-Modified (ABSD-Modified, Brannan, 1991). In this study, each of the subject's parents completed a rating scale of social and emotional characteristics of their child at the beginning of intervention and again after 20 weeks of intervention. The scores for the five social and emotional domains were compared for differences prior to and following intervention. Additionally, the amount of improvement for those social and emotional characteristics was compared between the two groups, one group receiving traditional articulation intervention and the other group receiving a phonological cycling approach. Data analysis revealed no statistically significant difference between pre- and post-intervention subscale scores for the traditional articulation intervention group and for the phonological cycling intervention group combined. The results also indicated no statistically significant difference in the amount of change in social and emotional characteristics between the two groups of subjects. However, the research data did show trends toward the statistically significant level of .05 in the social/emotional domains of self-esteem (p = .097) and communication (p = .091) for the phonological cycling group. Trends toward the statistically significant level in the two domains of self-esteem and communication suggest that articulation/phonological intervention may influence other areas in the individual's life. Therefore, further investigation of the research questions posed for this study is warranted.