Adaptive filtering for high quality HMM based speech synthesis

In this work an adaptive filtering scheme based on a dual Discrete Kalman Filtering (DKF) is proposed for Hidden Markov Model (HMM) based speech synthesis quality enhancement. The objective is to improve signal smoothness across HMMs and their related states and to reduce artifacts due to acoustic model's limitations. Both speech and artifacts are modelled by an autoregressive structure which provides an underlying time frame dependency and improves time-frequency resolution. Themodel parameters are arranged to obtain a combined state-space model and are also used to calculate instantaneous power spectral density estimates. The quality enhancement is performed by a dual discrete Kalman filter that simultaneously gives estimates for the models and the signals. The system's performance has been evaluated using mean opinion score tests and the proposed technique has led to improved results.

Effets sur l'articulation temporo-mandibulaire du Twin-Block et d'un appareil myofonctionnel de classe II.

Introduction. Ce projet de recherche consiste en une étude cohorte prospective randomisée visant à évaluer les douleurs ressenties au niveau de l’articulation temporo-mandibulaire (ATM) lors d’une thérapie d’avancement mandibulaire grâce à un appareil fixe, le correcteur de classe II (CC) et d’un appareil amovible, le Twin-Block (TB). Matériels et méthodes. Cette étude comptait 26 patients (11 hommes et 15 femmes), âge moyen de 12ans 10mois (10ans 4mois à 15ans 10mois). Les sujets devaient avoir une malocclusion de classe II et être en croissance, CVM 2 ou 3 (Cervical Vertebral Maturation). Les patients étaient divisés en deux groupes : TB et CC. La douleur était évaluée selon l’axe I de l’examen du RDC/TMD (Research Diagnostic Criteria for Temporomandibular Disorders) à 7 reprises (T0 à T6). De plus, le patient devait remplir un questionnaire, à la maison, sur la douleur ressentie et la médication prise lors des 30 premiers jours. La douleur était évaluée avant l’insertion des appareils (T0), à 1 semaine (T1) post-insertion, 4 semaines plus tard (T2), 8 semaines (T3) où une expansion de 20 tours (environs 5 mm) était débutée, ensuite (T4) (T5) et (T6) chacun à 8 semaines d’intervalles. Les tests statistiques utilisés dans cette étude : le test «Wilcoxon à un échantillon» ainsi que le test «Mann-Whitney à échantillons indépendants ». Résultats et Discussion. La douleur à l’examen clinique est variable mais tend à diminuer avec le temps. Aucune différence, statistiquement significative, ne fut observée entre les 2 groupes en ce qui à trait aux diverses palpations effectuées. Parmi les patients ayant rapporté de la douleur, 40% l’ont ressentie surtout le matin et 63,3% ont dit qu’elle durait de moins d’une heure jusqu’à quelques heures. Conclusion. D’après nos résultats, lors d’une thérapie myofonctionnnelle, il n’y a pas de différence statistiquement significative entre la douleur occasionnée par un Twin-Block et celle produite par un correcteur de classe II fixe au niveau de l’ATM et des muscles du complexe facial.

Adolescence, immigration et santé mentale : schisme et articulation des discours soignants autour des orientations et des stratégies d'intervention en contexte ethnopsychiatrique

La souffrance psychologique des adolescents fait l'objet d'une attention croissante des pouvoirs publics, des mondes associatifs, et bien sûr des cliniciens et des travailleurs psychosociaux. Pour les adolescents migrants, cette souffrance, aggravée par la fragilité, les malaises et les tensions de l'adolescence, est compliquée par l'appareil de santé publique bureaucratique complexe destiné à traduire dans le langage des soins le malaise des attitudes, des comportements et des tendances qui se dégagent de leur statut de migrant tantôt mal vécu, tantôt réprouvé par leur environnement et leur parcours personnel et familial. Ce mémoire de maîtrise a exploré l'un d'entre eux, celui de la clinique de psychiatrie transculturelle, où l'écoute de ce qui se dit importe tout autant que la traduction et la gestion de ces souffrances. En utilisant une approche anthropologique, j'analyse les discours et les attitudes de dix thérapeutes et les travailleurs sociaux qui interviennent auprès de ces familles. Cette recherche montre que le cadre de gestion sociale et culturelle qui inclut les soignants et les familles, et les ambivalences de jeunes adolescents, ne peut être analysé sans référencer le contexte plus large des différences de classes, d'accès au pouvoir, d'orientations et de bagages culturels affectant à la fois le vécu des adolescents et la manière avec laquelle leurs angoisses seront communiquées et traduites. Dans cette perspective particulière, les paradoxes et les résistances de certains professionnels médicaux et sociaux, dont les positions qu'ils occupent au sein des structures du pouvoir, autrement, pourraient les tenter de s'engager dans des diagnostics faciles liés à la socialisation prétendument difficile de ces adolescences. Au lieu de cela, ils créent des stratégies uniques qui respectent les idiomes officiels encadrant leur autorité médicale et sociale.

A socio friendly approach to the analysis of emotive speech

This paper describes certain findings of intonation and intensity study of emotive speech with the minimal use of signal processing algorithms. This study was based on six basic emotions and the neutral, elicited from 1660 English utterances obtained from the speech recordings of six Indian women. The correctness of the emotional content was verified through perceptual listening tests. Marked similarity was noted among pitch contours of like-worded, positive valence emotions, though no such similarity was observed among the four negative valence emotional expressions. The intensity patterns were also studied. The results of the study were validated using arbitrary television recordings for four emotions. The findings are useful to technical researchers, social psychologists and to the common man interested in the dynamics of vocal expression of emotions

A Hybrid Architecture for Recognising Speech Signals in Malayalam

Speech is the primary, most prominent and convenient means of communication in audible language. Through speech, people can express their thoughts, feelings or perceptions by the articulation of words. Human speech is a complex signal which is non stationary in nature. It consists of immensely rich information about the words spoken, accent, attitude of the speaker, expression, intention, sex, emotion as well as style. The main objective of Automatic Speech Recognition (ASR) is to identify whatever people speak by means of computer algorithms. This enables people to communicate with a computer in a natural spoken language. Automatic recognition of speech by machines has been one of the most exciting, significant and challenging areas of research in the field of signal processing over the past five to six decades. Despite the developments and intensive research done in this area, the performance of ASR is still lower than that of speech recognition by humans and is yet to achieve a completely reliable performance level. The main objective of this thesis is to develop an efficient speech recognition system for recognising speaker independent isolated words in Malayalam.

Perceptual Evaluation of Video-Realistic Speech

abstract With many visual speech animation techniques now available, there is a clear need for systematic perceptual evaluation schemes. We describe here our scheme and its application to a new video-realistic (potentially indistinguishable from real recorded video) visual-speech animation system, called Mary 101. Two types of experiments were performed: a) distinguishing visually between real and synthetic image- sequences of the same utterances, ("Turing tests") and b) gauging visual speech recognition by comparing lip-reading performance of the real and synthetic image-sequences of the same utterances ("Intelligibility tests"). Subjects that were presented randomly with either real or synthetic image-sequences could not tell the synthetic from the real sequences above chance level. The same subjects when asked to lip-read the utterances from the same image-sequences recognized speech from real image-sequences significantly better than from synthetic ones. However, performance for both, real and synthetic, were at levels suggested in the literature on lip-reading. We conclude from the two experiments that the animation of Mary 101 is adequate for providing a percept of a talking head. However, additional effort is required to improve the animation for lip-reading purposes like rehabilitation and language learning. In addition, these two tasks could be considered as explicit and implicit perceptual discrimination tasks. In the explicit task (a), each stimulus is classified directly as a synthetic or real image-sequence by detecting a possible difference between the synthetic and the real image-sequences. The implicit perceptual discrimination task (b) consists of a comparison between visual recognition of speech of real and synthetic image-sequences. Our results suggest that implicit perceptual discrimination is a more sensitive method for discrimination between synthetic and real image-sequences than explicit perceptual discrimination.

The intelligibility of different kinds of test material used in speech audiometry

This paper discusses several tests used to measure speech intelligibility and speech discrimination.

A comparative study of two methods of measuring loss of capacity to hear speech

This paper discusses a study to compare two tests of loss of capacity to hear speech.

Correlation of the articulation index to the MTS test under different listening conditions

This paper reviews a study to determine the relation between the aided articulation index and the aided speech recognition scores obtained with the Monosyllable, Trochee and Spondee (MTS) Test, when administered to hearing-impaired children.

Word level speech perception under four levels of simulated profound hearing loss

This paper discusses a study to determine average performance on word discrimination tests using the CID Early Speech Perception Test (ESP).

The effect of talker age and gender on speech perception of pediatric hearing aid users

Even though pediatric hearing aid (HA) users listen most often to female talkers, clinically-used speech tests primarily consist of adult male talkers' speech. Potential effects of age and/or gender of the talker on speech perception of pediatric HA users were examined using two speech tests, hVd-vowel identification and CNC word recognition, and using speech materials spoken by four talker types (adult males, adult females, 10-12 year old girls, and 5-7 year old girls). For the nine pediatric HA users tested, word scores for the male talker's speech were higher than those for the female talkers, indicating that talker type can affect word recognition scores and that clinical tests may over-estimate everyday speech communication abilities of pediatric HA users.

For better or worse: the effect of levodopa on speech in Parkinson's disease

While the beneficial effect of levodopa on traditional motor control tasks have been well documented over the decades. its effect on speech motor control has rarely been objectively examined and the existing literature remains inconclusive. This paper aims to examine the effect of levodopa on speech in patients with Parkinson's disease. It was hypothesized that levodopa would improve preparatory motor set related activity and alleviate hypophonia. Patients fasted and abstained from levodopa overnight. Motor examination and speech testing was performed the following day, pre-levodopa during their "off' state, then at hourly intervals post-medication to obtain the best "on" state. All speech stimuli showed a consistent tendency for increased loudness and faster rate during the "on" state, but this was accompanied by a greater extent of intensity decay. Pitch and articulation remained unchanged. Levodopa effectively upscaled the overall gain setting of vocal amplitude and tempo, similar to its well-known effect on limb movement. However, unlike limb movement, this effect on the final acoustic product of speech may or may not be advantageous, depending on the existing speech profile of individual patients. (C) 2007 Movement Disorder Society.

Word frequency and bigram frequency effects on linguistic processing and speech motor performance in individuals with aphasia and normal speakers

Models of normal word production are well specified about the effects of frequency of linguistic stimuli on lexical access, but are less clear regarding the same effects on later stages of word production, particularly word articulation. In aphasia, this lack of specificity of down-stream frequency effects is even more noticeable because there is relatively limited amount of data on the time course of frequency effects for this population. This study begins to fill this gap by comparing the effects of variation of word frequency (lexical, whole word) and bigram frequency (sub-lexical, within word) on word production abilities in ten normal speakers and eight mild–moderate individuals with aphasia. In an immediate repetition paradigm, participants repeated single monosyllabic words in which word frequency (high or low) was crossed with bigram frequency (high or low). Indices for mapping the time course for these effects included reaction time (RT) for linguistic processing and motor preparation, and word duration (WD) for speech motor performance (word articulation time). The results indicated that individuals with aphasia had significantly longer RT and WD compared to normal speakers. RT showed a significant main effect only for word frequency (i.e., high-frequency words had shorter RT). WD showed significant main effects of word and bigram frequency; however, contrary to our expectations, high-frequency items had longer WD. Further investigation of WD revealed that independent of the influence of word and bigram frequency, vowel type (tense or lax) had the expected effect on WD. Moreover, individuals with aphasia differed from control speakers in their ability to implement tense vowel duration, even though they could produce an appropriate distinction between tense and lax vowels. The results highlight the importance of using temporal measures to identify subtle deficits in linguistic and speech motor processing in aphasia, the crucial role of phonetic characteristics of stimuli set in studying speech production and the need for the language production models to account more explicitly for word articulation.

Speech motor control in fluent and dysfluent speech production of an individual with apraxia of speech and Broca’s aphasia

Apraxia of speech (AOS) is typically described as a motor-speech disorder with clinically well-defined symptoms, but without a clear understanding of the underlying problems in motor control. A number of studies have compared the speech of subjects with AOS to the fluent speech of controls, but only a few have included speech movement data and if so, this was primarily restricted to the study of single articulators. If AOS reflects a basic neuromotor dysfunction, this should somehow be evident in the production of both dysfluent and perceptually fluent speech. The current study compared motor control strategies for the production of perceptually fluent speech between a young woman with apraxia of speech (AOS) and Broca’s aphasia and a group of age-matched control speakers using concepts and tools from articulation-based theories. In addition, to examine the potential role of specific movement variables on gestural coordination, a second part of this study involved a comparison of fluent and dysfluent speech samples from the speaker with AOS. Movement data from the lips, jaw and tongue were acquired using the AG-100 EMMA system during the reiterated production of multisyllabic nonwords. The findings indicated that although in general kinematic parameters of fluent speech were similar in the subject with AOS and Broca’s aphasia to those of the age-matched controls, speech task-related differences were observed in upper lip movements and lip coordination. The comparison between fluent and dysfluent speech characteristics suggested that fluent speech was achieved through the use of specific motor control strategies, highlighting the potential association between the stability of coordinative patterns and movement range, as described in Coordination Dynamics theory.

Separating the contributions of hearing, lexical knowledge and speech production to speech perception scores in children with hearing impairments.

Open-set word and sentence speech-perception test scores are commonly used as a measure of hearing abilities in children and adults using cochlear implants and/or hearing aids. These tests are usually presented auditorily with a verbal response. In the case of children, scores are typically lower and more variable than for adults with hearing impairments using similar devices. It is difficult to interpret children's speech-perception scores without considering the effects of lexical knowledge and speech-production abilities on their responses. This study postulated a simple mathematical model to describe the effects of hearing, lexical knowledge, and speech production on the perception test scores for monosyllabic words by children with impaired hearing. Thirty-three primary-school children with impaired hearing, fitted with hearing aids and/or cochlear implants, were evaluated using speech-perception, reading-aloud, speech-production, and language measures. These various measures were incorporated in the mathematical model, which revealed that performance in an open-set word-perception test in the auditory-alone mode is strongly dependent on residual hearing levels, lexical knowledge, and speech-production abilities. Further applications of the model provided an estimate of the effect of each component on the overall speech-perception score for each child.