Vocal Tract

In this work, we present a silent speech system that is able to generate audible speech from captured movement of speech articulators. Our goal is to help laryngectomy patients, i.e. patients who have lost the ability to speak following surgical removal of the larynx most frequently due to cancer, to recover their voice. In our system, we use a magnetic sensing technique known as Permanent Magnet Articulography (PMA) to capture the movement of the lips and tongue by attaching small magnets to the articulators and monitoring the magnetic field changes with sensors close to the mouth. The captured sensor data is then transformed into a sequence of speech parameter vectors from which a time-domain speech signal is finally synthesised. The key component of our system is a parametric transformation which represents the PMA-to-speech mapping. Here, this transformation takes the form of a statistical model (a mixture of factor analysers, more specifically) whose parameters are learned from simultaneous recordings of PMA and speech signals acquired before laryngectomy. To evaluate the performance of our system on voice reconstruction, we recorded two PMA-and-speech databases with different phonetic complexity for several non-impaired subjects. Results show that our system is able to synthesise speech that sounds as the original voice of the subject and also is intelligible. However, more work still need to be done to achieve a consistent synthesis for phonetically-rich vocabularies.

High vertical laryngeal position (VLP), pharyngeal constriction, and laryngeal compression are common features associated with hyperfunctional voice disorders. The present study aimed to observe the effect on these variables of different semioccluded vocal tract postures in 20 subjects diagnosed with hyperfunctional dysphonia. During observation with flexible endoscope, each participant was asked to produce eight different semioccluded exercises: lip trills, hand-over-mouth technique, phonation into four different tubes, and tube phonation into water using two different depth levels. Participants were required to produce each exercise at three loudness levels: habitual, soft, and loud. To determine the VLP, anterior-to-posterior (A-P) compression, and pharyngeal width, a human evaluation test with three blinded laryngologists was conducted. Judges rated the three endoscopic variables using a five-point Likert scale. An intraclass correlation coefficient to assess intrarater and interrater agreement was performed. A multivariate linear regression model considering VLP, pharyngeal width, and A-P laryngeal compression as outcomes and phonatory tasks and intensity levels as predictive variables were carried out. Correlation analysis between variables was also conducted. Results indicate that all variables differ significantly. Therefore, VLP, A-P constriction, and pharyngeal width changed differently throughout the eight semioccluded postures. All semioccluded techniques produced a lower VLP, narrower aryepiglottic opening, and a wider pharynx than resting position. More prominent changes were obtained with a tube into the water and narrow tube into the air. VLP significantly correlated with pharyngeal width and A-P laryngeal compression. Moreover, pharyngeal width significantly correlated with A-P laryngeal compression.

Formant frequency of vowels in a language is considered as one of the important acoustical parameter of speech signal. This parameter can be seen as acoustic resonance of human vocal tract. Although formant frequencies which is changeable across genders, age, and languages have been studied for various purposes by many researchers in some languages, alteration effects of stop consonants on adjacent vowels hasn’t been worked yet for Turkish language. In this study, formant frequency values (F1, F2 and F3) of eight isolated vowels (/a/, /e/, /i/, /i/, /o/, /o/, /u/, /u/) have been compared to formant frequency values of vowels that come after any stop consonant (/p/, /t/, /k/, /b/, /d/, /g/) to detect any changes caused by stop consonants in formant frequencies. Totally 48 meaningful Turkish syllables (combinations of all stop consonants and all vowels) and 8 isolated vowels have been uttered by 10 male speakers three times repeatedly for each unit. At the end of this study, the plosi...

Emphasis (contrastive pharyngealization of coronals) in Arabic spreads from an emphatic consonant to neighboring segments. Previous research suggests that in addition to changing spectral characteristics of adjacent segments, emphasis might affect voice onset time (VOT) of voiceless stops because emphatic stops in Arabic dialects have considerably shorter VOT than their plain cognates. No study investigated whether emphatic coarticulation could shorten VOT in plain stops produced in emphatic environment. The present study investigates changes in VOT in syllable-initial /t/ using production data from sixteen speakers of Qatari Arabic, who read non-word syllables with initial plain and emphatic stops /t/ and /t. / adjacent to another plain or emphatic consonant. The results show that emphasis spread is a gradient process that affects only spectral characteristics of segments, causing changes in vowel formants and spectral centre of gravity of stops. Long-lag VOT in plain /t/, however, was not shortened in emphatic syllables. The findings suggest that shorter VOT in voiceless emphatic stops in Qatari Arabic is not a mechanical aftermath of pharyngealization but, rather, a phonological requirement to maintain contrast between long-lag and short-lag VOT in plain and emphatic stops.

This paper presents a sensor-augmented saxophone mouthpiece which promotes data collection from musicians. The collected data aims to unveil the mechanics of saxophone tone formation from the embouchure and air flow control—for which only limited knowledge exists due to its occlusion concealed by the players’ mouth. Ultimately, by grasping and modeling how musicians play, we aim to develop intelligent music tutoring systems. Towards this aim, we provide the first steps of a sensor-augmented mouthpiece from which we compute air pressure, air flow, and mouthpiece-reed tip opening indicators of a performance. Di↵erent strategies to capture these indicators have been studied under controlled lab conditions as a basis to develop a prototype mouthpiece which is evaluated by an expert saxophonist.

This paper examines the articulatory correlates of the Hero and Villain Voice Types, which were auditorily identified in a separate study on cartoon voices, using the magnetic resonance imaging (MRI) technique. In general, the MRI images were in good agreement with the previous auditory analysis results; the major characteristic difference between voice quality settings of heroes and those of villains and between two villainous voice types was found in the supraglottal states and the pharyngeal cavity. Auditory analysis can be as valid as acoustic or any other analysis method depending on the level of training in a commonly accepted system such as Laver's framework for voice quality description. However, the MRI technique also allowed us to see what would not be observed otherwise, e.g., larynx height, pharyngeal cavity, vocal tract length, and the position of the hyoid bone. Auditory and physiological methods should be used in combination in order to further our understanding of the larynx and the pharynx.

Based on the assumption that the goals of phonemic speech movements are both auditory and somatosensory in nature, a biomechanical model of the vocal tract in conjunction with an adaptive controller inspired by the DIVA model of speech production were employed to investigate how the feedforward and feedback mechanisms operate in the planning for the production of the vowel [iy] with the presence of a bite-block and/or masking noise. Simulations of the vocal tract model corroborate the hypotheses predicted by the DIVA model, suggesting that: a) in the absence of auditory feedback, the subject used his learned feedforward motor command for producing a vowel-to-vowel sequence; b) in the presence of a bite-block and auditory feedback, the subject relied on the acoustic feedback mechanism during his productions; c) in the absence of auditory feedback and with a biteblock present, the subject relied on his feedforward motor plans initially, and then later on, incorporated corrective commands from the somatosensory feedback system into his productions.

Alcohol is known to impair fine articulatory control and movements. In drunken speech, incomplete closure of the vocal tract can result in deaffrication of the English affricate sounds /tʃ/ and /ʤ/, spirantization (fricative-like production) of the stop consonants and palatalization (retraction of place of articulation) of the alveolar fricative /s/ (produced as /ʃ/). Such categorical segmental errors have been well-reported. This study employs a phonologically-informed neural network approach to estimate degrees of deaffrication of /tʃ/ and /ʤ/, spirantizationof /t/ and /d/ and place retraction for /s/ in a corpus of intoxicated English speech. Recurrent neural networks were trained to recognize relevant phonological features [anterior], [continuant] and [strident] in a control speech corpus. Their posteriorprobabilities were computed over the segments produced under intoxication. The results obtained revealed both categorical and gradient errors and, thus, suggested that this new ...

Best Tree Encoding (BTE) is a promising feature extraction technique based on wavelet packet decomposition that is utilized in Automatic Speech Recognition (ASR). This research introduces an enhancement of Wavelet Packet Best Tree (WPBT) Calculations. The standard features BTE encodes the tree structure using a mathematical model into a features vector of 4 components. The best tree structure has been calculated using the entropy function. In the standard version of BTE, Shannon entropy has been chosen as the entropy function. In this research, Shannon Entropy (SE), Renyi Entropy (RE), and Tsallis Entropy (TE) are used to construct the Best Tree. The encoding of the Best Tree has been done using the same mathematical model approach in the standard 4-Point BTE. The proposed model is tested and Verified against the most widely used feature Mel Frequency Cepstral Coefficient (MFCC) plus delta and delta-delta coefficients (39 parameters) to evaluate its performance. The TIMIT database is used in this research. All phones are divided into five classes: Vowels, Fricatives, Silences, Nasals, and Plosives. The acoustical model has been implemented using Hidden Markov Model (HMM). No language model has been applied. The HMM Tool Kit (HTK) software is used for model implementation. The experiments show that BTE using Tsallis entropy yields the highest overall success rate of 75.85% which is better than MFCC's overall success rate of 71.76%. Comparing the vector of 4 components of BTE to the 39 components vector of MFCC makes it a very promising feature vector to be considered for research and development.

Conduction aphasia is a language disorder characterized by frequent speech errors, impaired verbatim repetition, a deficit in phonological short-term memory, and naming difficulties in the presence of otherwise fluent and grammatical speech output. While traditional models of conduction aphasia have typically implicated white matter pathways, recent advances in lesions reconstruction methodology applied to groups of patients have implicated left temporoparietal zones. Parallel work using functional magnetic resonance imaging (fMRI) has pinpointed a region in the posterior most portion of the left planum temporale, area Spt, which is critical for phonological working memory. Here we show that the region of maximal lesion overlap in a sample of 14 patients with conduction aphasia perfectly circumscribes area Spt, as defined in an aggregate fMRI analysis of 105 subjects performing a phonological working memory task. We provide a review of the evidence supporting the idea that Spt is an interface site for the integration of sensory and vocal tract-related motor representations of complex sound sequences, such as speech and music and show how the symptoms of conduction aphasia can be explained by damage to this system.

The basic goal of the voice conversion system to mimics the characteristics of the target speaker voice by keeping the linguistic and paralinguistic information intact. The characteristics of a speaker in speech reflect at different level such as vocal tract, excitation and prosodic parameters. This propose work based on cepstrum which represents the vocal tract and excitation parameters of the speech. This paper proposes the decomposition of the cepstrum by wavelet and mapped the source cepstrum features in to target cepstrum features using Radial

The basic goal of the voice conversion system to mimics the characteristics of the target speaker voice by keeping the linguistic and paralinguistic information intact. The characteristics of a speaker in speech reflect at different level such as vocal tract, excitation and prosodic parameters. This propose work based on cepstrum which represents the vocal tract and excitation parameters of the speech. This paper proposes the decomposition of the cepstrum by wavelet and mapped the source cepstrum features in to target cepstrum features using Radial basis function neural network. The results are evaluated using subjective and objective measures based on voice quality method and the listening tests prove that the proposed algorithm converts speaker individuality while maintaining high speech quality..

The complex cepstrum vocoder is used to modify the speaker specific characteristics of the source speaker speech to that of the target speaker speech. The low time and high time liftering are used to split the calculated cepstrum into the vocal tract and the source excitation parameters. The obtained mixed phase vocal tract and source excitation parameters with finite impulse response preserve the phase properties of the resynthesized speech frame. The radial basis function is explored to capture the nonlinear mapping function for modifying the complex cepstrum based real and imaginary components of the vocal tract and source excitation of the speech signal. The state-of-the-art Mel cepstrum envelope and the fundamental frequency (F0) are considered to represent the vocal tract and the source excitation of the speech frame, respectively. Radial basis function is used to capture and formulate the nonlinear relations between the Mel cepstrum envelope of the source and target speakers....

In speaker transformation, the speaker dependent spectral parameters are generally characterized by single scale features. These features approximate the vocal tract, but produce artifacts during speech signal reconstruction. In this paper, multi-resolution wavelet based feature set is proposed, which finely tunes the speaker specific characteristics of the speech signal. The Radial Basis Function is used to propose the mapping function for modifying these characteristics. The performance of the proposed system is evaluated using different objective and subjective measures. Evaluation results illustrate that the proposed algorithm maintains target voice individuality while maintaining the quality and naturalness of the speech signal.

The objective of voice conversion system is to formulate the mapping function which can transform the source speaker characteristics to that of the target speaker. In this paper, we propose the General Regression Neural Network (GRNN) based model for voice conversion. It is a single pass learning network that makes the training procedure fast and comparatively less time consuming. The proposed system uses the shape of the vocal tract, the shape of the glottal pulse (excitation signal) and long term prosodic features to carry out the voice conversion task. In this paper, the shape of the vocal tract and the shape of source excitation of a particular speaker are represented using Line Spectral Frequencies (LSFs) and Linear Prediction (LP) residual respectively. GRNN is used to obtain the mapping function between the source and target speakers. The direct transformation of the time domain residual using Artificial Neural Network (ANN) causes phase change and generates artifacts in consecutive frames. In order to alleviate it, wavelet packet decomposed coefficients are used to characterize the excitation of the speech signal. The long term prosodic parameters namely, pitch contour (intonation) and the energy profile of the test signal are also modified in relation to that of the target (desired) speaker using the baseline method. The relative performances of the proposed model are compared to voice conversion system based on the state of the art RBF and GMM models using objective and subjective evaluation measures. The evaluation measures show that the proposed GRNN based voice conversion system performs slightly better than the state of the art models.

In this paper we present results of three distinct studies addressing European Portuguese Nasal Vowels height. studies contemplated: analysis of EMMA data for one male speaker, analysis of first formant values for nasal vowels after stops in a corpus with contextual and regional coverage, and a perceptual experiment using articulatory synthesizer generated stimuli.

Patients with larynx cancer often lose their voice following total laryngectomy. Current methods for postlaryngectomy voice restoration are all unsatisfactory due to different reasons: requires frequent replacement due to biofilm growth (tracheo-oesoephageal valve), speech sounds gruff and masculine (oesophageal speech) or robotic (electro-larynx) and, in general, are difficult to master (oesophageal speech and electro-larynx). In this work we investigate an alternative approach for voice restoration in which speech articulator movement is converted into audible speech using a speaker-dependent transformation learned from simultaneous recordings of articulatory and audio signals. To capture articulator movement, small magnets are attached to the speech articulators and the magnetic field generated while the user 'mouths' words is captured by a set of sensors. Parallel data comprising articulatory and acoustic signals recorded before laryngectomy are used to learn the mapping between the articulatory and acoustic domains, which is represented in this work as a mixture of factor analysers. After laryngectomy, the learned transformation is used to restore the patient's voice by transforming the captured articulator movement into an audible speech signal. Results reported for normal speakers show that the proposed system is very promising.

They use optical sensors and artificial intelligence methods for process supervision and diagnostics. Research is aimed to develop a system allowing a parametric evaluation of the quality of pulverized coal burner operation. Due to the highly nonlinear nature of dependencies and lack of an analytical model, the artificial intelligence methods were used to estimate and classify the selected parameter, including a relatively new class of classification methodsartificial immunology algorithms. The article shows results for coal-shredded straw blends, yet the methodology may be applied for other types of blends.

Extracting reliable 3D facial deformation parameters from static facial postures is a major component of our system for audiovisual synthesis. This paper describes several important improvements to that process, including reduction of position alignment errors, simplification of the generic face mesh and, most important, increasing the range and variety of static postures used.

When creating realistic talking head animations, accurate modeling of speech articulators is important for speech perceptibility. Previous lip modeling methods such as simple numerical lip modeling focus on creating a general lip model without incorporating lip speaker variations. Here we present a method for creating accurate speaker-specific lip representations that retain the individual characteristics of a speaker's lips via an adaptive numerical approach using 3D scanned surface and MRI data. By adjusting spline parameters automatically to minimize the error between node points of the lip model and the raw 3D surface, new 3D lips are created efficiently and easily. The resulting lip models will be used in our talking head animation system to evaluate auditory-visual speech perception, and to analyze our 3D face database for statistically relevant lip features.

À primeira vista, flauta e saxofone parecem instrumentos muito diferentes: um é metálico com palheta, outro é um tubo aberto soprando diretamente. Mas ambos funcionam com o mesmo princípio físico: a vibração de uma coluna de ar dentro de um tubo. O que muda é a forma como essa coluna de ar vibra, e isso explica por que cada instrumento tem um timbre tão característico. Quando tocamos uma nota, não ouvimos apenas uma única frequência. O som é formado pela nota fundamental (a base) e por uma série de sons mais agudos chamados harmônicos. -Esses harmônicos são como "cores" que se misturam à nota principal. -O conjunto deles é que dá a "assinatura" do timbre de cada instrumento. s É por isso que, mesmo tocando a mesma nota, um saxofone e uma flauta soam completamente diferentes. A flauta é um tubo aberto nos dois lados. Isso significa que a onda estacionária forma sempre pontos de vibração máxima (antinódulos) nas extremidades. -Resultado: todos os harmônicos (pares e ímpares) podem soar com liberdade. -Efeito no som: o timbre fica claro, brilhante, cristalino, cheio de "cintilação". O saxofone, apesar de ter campana aberta, funciona como tubo fechado em uma extremidade (onde está a palheta). -Isso, em teoria, deveria privilegiar apenas os harmônicos ímpares, como acontece no clarinete. -Mas aqui entra a diferença: o saxofone tem tubo cônico (vai se alargando progressivamente). -Essa conicidade reorganiza os nós e antinós, permitindo que todos os harmônicos estejam presentes. -Efeito no som: o timbre do sax é rico, cheio, uniforme em todos os registros. -Flauta: soa mais "aberta", com brilho e leveza. -Clarinete: reforça apenas os ímpares → som mais "oco" ou fechado. -Saxofone: mistura todos os harmônicos → som encorpado, flexível, que se adapta bem tanto ao jazz quanto à música erudita. s Se a flauta fosse uma "luz branca com muito brilho", o sax seria uma "luz quente e colorida". Entender o papel dos harmônicos nos ajuda a perceber que cada instrumento tem uma identidade sonora construída pela física do ar dentro do tubo. -A flauta encanta pelo brilho cristalino. -O sax fascina pelo corpo e calor do som. E no final, ambos nos mostram como a ciência do som e a arte da música caminham juntas, transformando física em emoção.

A articulação e a postura determinam a configuração do trato vocal, definindo a produção da voz. O objetivo foi caracterizar o perfil vocal de professores com e sem queixa vocal; descrever e comparar, por meio de análise objetiva e subjetiva, os comportamentos articulatórios e posturais. Portanto, foi aplicado um protocolo e os professores foram audiogravados dentro e fora do contexto profissional. As imagens da articulação foram analisadas de forma objetiva (software Image Tool) e subjetiva (com observação da postura por quatro juízes). A partir dos dados, sugere-se que o aumento da articulação tenha relação com o processo de alfabetização e não como prevenção às alterações vocais. Estas, associadas às mudanças posturais, podem levar ao aparecimento ou ao agravamento de um problema vocal. A metodologia proposta permite outras aplicabilidades na fonoaudiologia.

é considerado o pai do saxofone clássico americano. Foi o primeiro professor universitário de saxofone nos Estados Unidos, lecionando na Universidade de Michigan de 1953 a 1974. Sua obra *The Art of Saxophone Playing*, publicada em 1963, tornou-se referência mundial e continua sendo adotada por músicos e professores em diferentes estilos e países.

Nos últimos anos, proliferaram acessórios que prometem "melhorar" o som do saxofone por meio do aumento de massa em pontos específicos: tudel, anéis/cintas de fixação, parafusos especiais, campana e até bases de apoio. A alegação recorrente é que esses reforços "potencializam a ressonância", "estabilizam harmônicos" e "corrigem" afinação. Mas afinal, o que realmente muda no som quando se adiciona massa? E, principalmente, o que não muda? A resposta exige olhar para dois pilares: (1) o que a ciência acústica já demonstrou e (2) o que a prática de oficina e escuta crítica comprova há décadas.

Recent techniques of evaluation of vocal tract acoustic transfer functions are based on the external excitation of the tract at the thyroid cartilage level. This paper presents further developments of methods using Gaussian white noise or pseudo-random sequences. In addition to the traditional non-parametric spectral characterisation, an automatic estimation of formants and bandwidths has been developed using model identification procedures. A user-friendly protocol has been implemented to perform quasi simultaneous recording of transfer functions and radiated sound. This method has been used to study the influence of glottis conditions on formants and bandwidths for the French vowels, and an experimental formant-cavity aflliation index has been defined as the relative bandwidth variation between closed glottis and sound production conditions. Finally a first attempt to characterise the aspiration noise spectrum for whispered vowels has shown that the ovexall source spectral tilt increase with overall SPL.

How are glottal source and vocal tract involved in the sopranos' passagio around E5? This pilot study investigates a legit soprano singing C5 to C6 with a lyrical technique and C5 to D7 in a "light" voice. On (de)crescendi and glissandi over the passagio, we observed clear bifurcations from "light" to lyrical mode, with discontinuous changes in f0, larynx vertical position, amplitude and shape of the electroglottographic signal. No such discontinuities occurred from C5 to D7 within the "light" mode. We established phonetograms of both modes and characterised differences in voice quality (Spectrum, Harmonicity, Open quotient). For [oe] in both modes, this singer tunes the first vocal tract resonance (R1) to f0 over the passagio. For a free vowel, a single and strong resonance is observed in the "light" mode, which may correspond to the merging of R1 and R2 and which is tuned to 2 f0 below C6, then to f0 until F6. Above F#6, two resonances a...

In closely musical environment the consciousness of the influence of the vocal tract on the clarinet sound is consolidated enough; indeed some clarinet players, by means of little alterations of their own vocal tract, are able to obtain various sounds, and slights corrections of tuning. ...

In this paper we examine accounts of "exotic" components of the sound systems of Iroquoian, Polynesian and Khoisan languages, and their implications for the history of phonetic studies and linguistics in general. On the basis of examples from European and American scholarship between the 17th and early 20th century we demonstrate recurring misconceptions in the description of consonant inventories, phonotactic structures as well as intra-and inter-speaker variation and change. Further, we examine their influence on the interpretation of other components of language and their role in the construction of a biased image of the languages and their speakers.

HAL is a multi-disciplinary open access archive for the deposit and dissemination of scientific research documents, whether they are published or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers. L'archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d'enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

This work is part of a project aiming at defining possible speech prerequisites in the geometry, the musculature and the control of the vocal tract. In this paper we intend to reconstruct anatomical and geometrical landmarks of the vocal tract from the bony structure of modern man. Thereafter, we will try to extrapolate our procedure to fossilized skull and mandible. We constituted a head and neck xeroradiographic database for 10 speakers. Using bony references and positional angles we intend to predict two dimensions: palatal distance and larynx height. These dimensions can be used to quantify growth differences (from birth to adulthood) and gender differences, as well as differences between species. They are important in that they lead to the prediction of maximal acoustic space of the vocal tract. Then we evaluated the range of accuracy of these predictions. This work has been carried out by a multidisciplinary team comprised of specialists in radiology, anthropology, informatics...

This work is part of a project in a quest of the origin of speech. From classical bony landmarks of the head and jaw used in anthropology, and using a generic model of the vocal tract we attempted to apply the prediction of geometric limits of the vocal tract for modern man to fossils covering a period from 10 ka until 2 Ma (Paleolithic period). We can infer that all the reconstituted vocal tracts could produce the same variety of speech sounds as we can today. But we do not know to what extent they mastered the control skills needed to produce speech.

The objective of this work is twofold. First, a model of the vocal tract is positioned into the bony architecture of the male and female skulls from birth to adulthood. Second, vowel spaces are determined and vowel prototypes, for the cardinal vowels, are synthesized using a simulation of the laryngeal source. Results of this modeling study during ontogeny allow for a better understanding of speech acquisition processes in infants and vocal tract reconstruction of fossils' Hominids.

There has been a lack of objective data on the singing voice registers, particularly on the so called "whistle" register, occurring in the top part of the female pitch range, which is accessible only to some singers. This study offers unique strobolaryngoscopic and high-speed (7812.5 images/s) videokymographic data on the vocal fold behavior of an untrained female singer capable of producing three distinct voice qualities, i.e., the chest, head and whistle registers. The sound was documented spectrographically. The transition from chest to head register, accompanied by pitch jumps, occurred around tones B4-C#5 (500-550 Hz) and was found to be associated with a slight decrease in arytenoids adduction, resulting in decrease of the closed quotient. The register shifts from head to whistle, also accompanied by pitch jumps, occurred around tones E5-B5 (670-1000 Hz) without any noticeable changes in arytenoids adduction. Some evidence was found for the vocal tract influence on t...

There has been a lack of objective data on the singing voice registers, particularly on the so called “whistle” register, occurring in the top part of the female pitch range, which is accessible only to some singers. This study offers unique strobolaryngoscopic and high-speed (7812.5images∕s) videokymographic data on the vocal fold behavior of an untrained female singer capable of producing three distinct voice qualities, i.e., the chest, head and whistle registers. The sound was documented spectrographically. The transition from chest to head register, accompanied by pitch jumps, occurred around tones B4–C#5 (500–550Hz) and was found to be associated with a slight decrease in arytenoids adduction, resulting in decrease of the closed quotient. The register shifts from head to whistle, also accompanied by pitch jumps, occurred around tones E5-B5 (670–1000Hz) without any noticeable changes in arytenoids adduction. Some evidence was found for the vocal tract influence on this transitio...

We address the hypothesis that postures adopted during grammatical pauses in speech production are more ''mechanically advantageous'' than absolute rest positions for facilitating efficient postural motor control of vocal tract articulators. We quantify vocal tract posture corresponding to inter-speech pauses, absolute rest intervals as well as vowel and consonant intervals using automated analysis of video captured with real-time magnetic resonance imaging during production of read and spontaneous speech by 5 healthy speakers of American English. We then use locally-weighted linear regression to estimate the articulatory forward map from low-level articulator variables to high-level task/goal variables for these postures. We quantify the overall magnitude of the first derivative of the forward map as a measure of mechanical advantage. We find that postures assumed during grammatical pauses in speech as well as speech-ready postures are significantly more mechanically advantageous than postures assumed during absolute rest. Further, these postures represent empirical extremes of mechanical advantage, between which lie the postures assumed during various vowels and consonants. Relative mechanical advantage of different postures might be an important physical constraint influencing planning and control of speech production.

-time magnetic resonance imaging (RT-MRI) of human speech production is enabling significant advances in speech science, linguistics, bio-inspired speech technology development, and clinical applications. Easy access to RT-MRI is however limited, and comprehensive datasets with broad access are needed to catalyze research across numerous domains. The imaging of the rapidly moving articulators and dynamic airway shaping during speech demands high spatio-temporal resolution and robust reconstruction methods. Further, while reconstructed images have been published, to-date there is no open dataset providing raw multi-coil Rt-MRI data from an optimized speech production experimental setup. Such datasets could enable new and improved methods for dynamic image reconstruction, artifact correction, feature extraction, and direct extraction of linguistically-relevant biomarkers. The present dataset offers a unique corpus of 2D sagittal-view RT-MRI videos along with synchronized audio for 75 participants performing linguistically motivated speech tasks, alongside the corresponding public domain raw RT-MRI data. The dataset also includes 3D volumetric vocal tract MRI during sustained speech sounds and high-resolution static anatomical T2-weighted upper airway MRI for each participant.

A real-time MRI examination of retroflex stops and rhotics in Tamil reveals that in some contexts these consonants may in fact be achieved with little or no retroflexion of the tongue tip. Rather, maneuvering and shaping of the tongue in order to achieve post-alveolar contact varies across vowel contexts. Between back vowels /a/ and /u/, post-alveolar constriction involves curling back of the tongue tip, but in the context of high front vowel /i/, the same constriction is achieved by bunching of the tongue. It appears that though there is a stable constriction target in the post-alveolar region, its achievement is not fixed but is instead a consequence of the variable state of the vocal tract in different vowel contexts. Articulatory configurations of the tongue across these vowel contexts were examined by comparing measures of Gaussian curvature at evenly spaced points along the vocal tract. The results support the notion that so-called retroflex consonants have a specified target constriction in the post-alveolar region, but that the specific articulations employed to achieve this constriction are not fixed, in keeping with the task dynamic model of speech production.

The English past tense allomorph following a coronal stop (e.g., /bɑndəd/) includes a vocoid that has traditionally been transcribed as a schwa or as a barred i. Previous evidence has suggested that this entity does not involve a specific articulatory gesture of any kind. Rather, its presence may simply result from temporal coordination of the two temporally adjacent coronal gestures, while the interval between those two gestures remains voiced and is acoustically reminiscent of a schwa. The acoustic and articulatory characteristics of this vocoid are reexamined in this work using real-time MRI with synchronized audio which affords complete midsagittal views of the vocal tract. A novel statistical analysis is developed to address the issue of articulatory targetlessness based on previous models that predict articulatory action from segmental context. Results reinforce the idea that this vocoid is different, both acoustically and articulatorily, than lexical schwa, but its targetless...

We address the hypothesis that postures adopted during grammatical pauses in speech production are more ''mechanically advantageous'' than absolute rest positions for facilitating efficient postural motor control of vocal tract articulators. We quantify vocal tract posture corresponding to inter-speech pauses, absolute rest intervals as well as vowel and consonant intervals using automated analysis of video captured with real-time magnetic resonance imaging during production of read and spontaneous speech by 5 healthy speakers of American English. We then use locally-weighted linear regression to estimate the articulatory forward map from low-level articulator variables to high-level task/goal variables for these postures. We quantify the overall magnitude of the first derivative of the forward map as a measure of mechanical advantage. We find that postures assumed during grammatical pauses in speech as well as speech-ready postures are significantly more mechanically advantageous than postures assumed during absolute rest. Further, these postures represent empirical extremes of mechanical advantage, between which lie the postures assumed during various vowels and consonants. Relative mechanical advantage of different postures might be an important physical constraint influencing planning and control of speech production.

Speech production can be described in multiple coordinate frames: articulatory configurations, gestural tasks, and acoustic patterns. Examination of the achievement of retroflex stops and liquids in Tamil suggests that we must consider separately the gestural task of apical pre-palatal constriction and the articulatory maneuver to achieve the task. The maneuver of the tongue during retroflex consonants varies across vowel contexts. Specifically, in the symmetrical intervocalic contexts between back vowels /a/ and /u/, an apical pre-palatal constriction is achieved by curling back the tongue. In the context of high front vowel /i/, a laminal pre-palatal constriction is achieved by bunching the tongue. However, the location of retroflex consonant constriction within the vocal tract is consistent across all of these vowel contexts, suggesting that the constriction task remains the same. Variation in the articulatory configuration of the retroflex in the two contexts was quantified thro...

This paper presents an automatic procedure to analyze articulatory setting in speech production using real-time magnetic resonance imaging of the moving human vocal tract. The procedure extracts frames corresponding to inter-speech pauses, speech-ready intervals and absolute rest intervals from magnetic resonance imaging sequences of read and spontaneous speech elicited from five healthy speakers of American English and uses automatically extracted image features to quantify vocal tract posture during these intervals. Statistical analyses show significant differences between vocal tract postures adopted during inter-speech pauses and those at absolute rest before speech; the latter also exhibits a greater variability in the adopted postures. In addition, the articulatory settings adopted during inter-speech pauses in read and spontaneous speech are distinct. The results suggest that adopted vocal tract postures differ on average during rest positions, ready positions and inter-speec...

The original task-dynamic model of speech production incorporated the theoretical tenets of Articulatory Phonology and provided a dynamics of inter-articulator coordination for single and co-produced constriction gestures, given a gestural score that specifies a time-dependent vector of gestural activations for a given utterance. More recently, the model has been significantly extended to provide a framework for investigating the higher order dynamics of prosodic phrasing, syllable structure, lexical stress, and the prominence (accentual) properties associated with higher level prosodic constituents (e.g., foot, word, phrase, sentence). There are two new components in the model. The first is an ensemble of gestural planning oscillators that defines a dynamics of gestural score formation in that, once the ensemble reaches an entrained steady-state of relative phasing, the waveform of each oscillator is used to specify the activation function of that oscillator's associated const...

Speech is produced when time varying vocal tract system is excited with time varying excitation source. Therefore, the information present in a speech such as message, emotion, language, speaker is due to the combined effect of both excitation source and vocal tract system. However, there is very less utilization of excitation source features to recognize emotion. In our earlier work, we have proposed a novel method to extract glottal closure instants (GCIs) known as epochs. In this paper, we have explored epoch features namely instantaneous pitch, phase and strength of epochs for discriminating emotions. We have combined the excitation source features and the well known Male-frequency cepstral coefficient (MFCC) features to develop an emotion recognition system with improved performance. DNN-HMM speaker adaptive models have been developed using MFCC, epoch and combined features. IEMOCAP emotional database has been used to evaluate the models. The average accuracy for emotion recognition system when using MFCC and epoch features separately is 59.25% and 54.52% respectively. The recognition performance improves to 64.2% when MFCC and epoch features are combined.

The paper describes advances in the development of an ultrasound silent speech interface for use in silent communications applications or as a speaking aid for persons who have undergone a laryngectomy. It reports some first steps towards making such a device lightweight, portable, interactive, and practical to use. Simple experimental tests of an interactive silent speech interface for everyday applications are described. Possible future improvements including extension to continuous speech and real time operation are discussed.

BackgroundThe astonishing variety of sounds that birds can produce has been the subject of many studies aiming to identify the underlying anatomical and physical mechanisms of sound production. An interesting feature of some bird vocalisations is the simultaneous production of two different frequencies. While most work has been focusing on songbirds, much less is known about dual-sound production in non-passerines, although their sound production organ, the syrinx, would technically allow many of them to produce “two voices”. Here, we focus on the king penguin, a colonial seabird whose calls consist of two fundamental frequency bands and their respective harmonics. The calls are produced during courtship and for partner and offspring reunions and encode the birds’ identity. We dissected, μCT-scanned and analysed the vocal tracts of six adult king penguins from Possession Island, Crozet Archipelago.ResultsKing penguins possess a bronchial type syrinx that, similarly to the songbird’s...

Active control is widely used in industry. However, there have been relatively few applications to musical instruments, particularly wind instruments. The aim of this study is to attempt to control the sound quality and playability of wind instruments, using active control. Active control makes it possible to modify the input impedance (amplitude and frequency) of an instrument and to modify the instrument's quality. Simulations and first experiments on a cylindrical tube, which is considered to be a simple wind instrument, with embedded microphone and speaker are presented. Finally, the effects on the sound and the input impedance of the instrument are studied.