Search

International audience; The design of pianos is mainly based on empirical knowledge due to the lack of a simple tool that could predict sound changes induced by modifications of the geometry and/or the mechanical properties of the soundboard. We introduce the concept of Sound Computer-Aided Design through the framework of a program that is intended to simulate the acoustic results of virtual pianos. The calculation of the sound is split into four modules that compute respectively the modal basis of the stiffened soundboard, the string dynamics excited by the hammer, the soundboard dynamics excited by the string vibration, and the sound radiation. The exact resemblance between synthesis and natural sounds is not the primary purpose of the software. However, sound synthesis of real and modified pianos are used as reference tests to assess our main objective, namely to reflect faithfully structural modifications in the produced sound, and thus to make this tool helpful for both instrument makers and researchers of the musical acoustics community.

International audience; Wearing an oxygen mask changes the speech production of speakers. It indeed modifies the vocal apparatus and perturbs the articulatory movements of the speaker. This paper studies the impact of the oxygen mask of military aircraft pilots on formant trajectories, both dynamically (variations of the formants at a utterance level) and globally (mean value at the utterance level) for 12 speakers. A comparative analysis of speech collected with and without an oxygen mask shows that the mask has a significant impact on the formant trajectories, both on the mean values and on the formant variations at the utterance level. This impact is strongly dependent on the speaker and also on the mask model. These observations suggest that the articulatory movements of the speaker are modified by the presence of the mask. These observations are validated via a preliminary ASR experiment that uses a data augmentation technique based on articulatory perturbations that are driven by our experimental observations.

La phonétique expérimentale, comme de nombreux autres domaines scientifiques, a profité des émulations philosophiques du siècle des Lumières. Les avancées techniques de l’époque, elles-mêmes sous-tendues par le courant mécaniste qui à cette même époque influence l’ensemble des domaines scientifiques, sont venues nourrir la discipline. Pour s’en convaincre, on peut par exemple s’intéresser au cas des machines parlantes (speaking machine) qui ont fleuri au cours du XVIIIe siècle, en profitant notamment des développements scientifiques de l’époque. Ces machines, dont les premières versions remontent probablement au XVIIIe siècle (Heudin 2008), visaient à reproduire, de façon artificielle, la parole naturelle, en recréant les phénomènes articulatoires, mécaniques, physiques et acoustiques à l’origine de la production de signaux sonores. Le défi était de taille puisqu’il était alors impossible de générer, de contrôler ou de manipuler des signaux électriques. Ceci ne sera rendu possible qu’un siècle plus tard lorsque les méthodes d’enregistrement et de restitution du signal acoustique auront été inventées puis améliorées (voir Pagani-Naudet et al. dans cet ouvrage). Il s’agissait donc nécessairement de produire, depuis la source, une parole artificielle qui soit le plus proche possible, d’un point de vue perceptif de la parole naturelle. De plus, compte tenu du niveau de connaissances anatomiques, physiologiques mais aussi physiques sur le fonctionnement acoustique de la production de la parole, la plupart des concepteurs de machines parlantes ont dû faire confiance à leur intuition, plus ou moins bonne pour arriver à leurs fins.Même s’il est difficile de connaître les réelles motivations des concepteurs de ces machines puisqu’ils n’ont pas toujours documenté ces aspects de leur recherche, il semble toutefois que les premières tentatives ont été effectuées dans l’unique but d’imiter la voix humaine, et ce quel que soit le réalisme du modèle physique employé. Il faudra attendre le milieu du XIXe siècle pour que des scientifiques prennent conscience des potentialités de tels dispositifs et les utilisent non seulement pour l’étude de la production de la parole mais puissent également concevoir des dispositifs expérimentaux permettant de réaliser de réelles avancées sur la connaissance des mécanismes de production de la parole humaine. Différents dispositifs de machines parlantes seront passés en revue dans ce chapitre en suivant cette distinction de finalité. Une tentative de classement entre les machines ayant pour but l’imitation de la voix, la reproduction de la parole et la synthèse de la parole sera donc proposée en suivant autant que possible l’ordre chronologique. Il est également intéressant de noter que l’existence de ces multiples approches subsiste encore aujourd’hui en synthèse de parole ; ce qui permettra en fin de chapitre de discuter brièvement de ce parallèle essentiel entre les machines parlantes d’hier et celles d’aujourd’hui, qu’elles soient expérimentales ou numériques.

International audience; The design of pianos is mainly based on empirical knowledge due to the lack of a simple tool that could predict sound changes induced by changes of the geometry and/or the mechanical properties of the soundboard. We present the framework of a program for the Computer-Aided Design of piano soundboards that is intended to bridge that gap by giving piano makers a tool to synthesize tones of virtual pianos. The sound synthesis is solely based on physical models of the instrument in playing situation. The calculation of the sound is split into several modules: computation of the modal basis of the stiffened soundboard, computation of the string dynamics, simulation of the soundboard dynamics excited by the string vibration, and calculation of the sound radiation. Reference tests of sound synthesis of real pianos as well as sound synthesis of modified pianos are used to assess our main objective, namely to reflect faithfully structural modifications in the produced sound, and thus to make this tool helpful for both piano makers and researchers of the musical acoustics community.

International audience; This work concerns the construction of articulatory models for synthesis of speech, and more specifically the velum and epiglottis. The direct application of principal component analysis to the contours of these articulators extracted from MRI images results in unrealistic factors due to delineation errors. The approach described in this paper relies on the application of PCA to the centerline of the articulator and a simple reconstruction algorithm to obtain the global articulator contour. The complete articulatory model was constructed from static Magnetic Resonance (MR) images because their quality is much better than that of dynamic MR images. We thus assessed the extent to which the model constructed from static images is capable of approaching the vocal tract shape in MR images recorded at 55 Hz for continuous speech. The analysis of reconstruction errors shows that it is necessary to add dynamic images to the database of static images, in particular to approach the tongue shape for the /l/ sound.

Revised Selected Papers of the 11th International Seminar, ISSP 2017, Tianjin, China, October 16-19, 2017; International audience; The aim of this work is to develop an algorithm for controlling the articulators (the jaw, the tongue, the lips, the velum, the larynx and the epiglottis) to produce given speech sounds, syllables and phrases. This control has to take into account coarticulation and be flexible enough to be able to vary strategies for speech production. The data for the algorithm are 97 static MRI images capturing the articulation of French vowels and blocked consonant-vowel syllables. The results of this synthesis are evaluated visually, acoustically and perceptually, and the problems encountered are broken down by their origin: the dataset, its modeling, the algorithm for managing the vocal tract shapes, their translation to the area functions, and the acoustic simulation. We conclude that, among our test examples, the articulatory strategies for vowels and stops are most correct, followed by those of nasals and fricatives. Improving timing strategies with dynamic data is suggested as an avenue for future work.

International audience; The paper presents a numerical study about the acoustic impact of the glottal chink opening on the production of fricatives. Sustained fricatives are simulated by using classic lumped circuit element methods to compute the propagation of the acoustic wave along the vocal tract. A recent glottis model is connected to the wave solver to simulate a partial abduction of the vocal folds during their self-oscillating cycles. Area functions of fricatives at the three places of articulation of French (palato-alveolar, alveolar, and labiodental) have been extracted from static MRI acquisitions. Simulations highlight the existence of three distinct regimes, named A, B, and C, depending on the chink opening. They are characterized by the frication noise level: A exhibits a low frication noise level, B is a mixed noise/voice signal, and C contains only frication noise. They have significant impacts on the first spectral moments. Boundaries of these regimes are defined in terms of minimal abduction of the vocal folds, and simulations show that they depend on articulatory and glottal configurations. Regime B is shown to be unstable: it requires very specific configurations in comparison with other regimes, and acoustic features are very sensitive in this regime.

International audience; This paper is an attempt to solve two problems related to musical acoustics. The first one consists in defining a signature of an instrument, namely, summarizing its vibroacoustical behavior. The second one deals with the existing relationship between the musical sound and the vibroacoustic properties of the instrument body. The violin is the application of this paper. A proposed solution for the first problem consists in an estimation of the bridge transfer mobility and the mean-value of the lateral bridge transfer mobility. The second problem is studied via the comparison between the amplitudes of harmonics, extracted from a glissando audio signal, and the lateral bridge transfer mobility: Both curves exhibit similar features. This is the main result of the paper. This is evidenced by studying the effect of a violin mute on both the lateral bridge transfer mobility and the produced sound. Finally, this is evidenced by successfully identifying which violin is played in an audio recording, using the computation of the Pearson distance between the distribution of the amplitude of harmonics and a database of measured mobilities.

International audience; Acoustic Black Hole (ABH) effect is a passive vibration reduction technique which takes advantage of properties of wave propagation in thin structures of varying thickness. A practical implementation of ABH on a uniform beam consists in an extremity whose thickness follows a power-law profile covered by a very thin layer of additional damping material. A modal analysis based on a High Resolution technique shows that the ABH significantly increases the Modal Overlap Factor (MOF) of the beam, thus reducing the resonant behaviour of the structure. Further investigations, including a two dimensional numerical model of the structure based on the finite difference method, show that this MOF can be explained by an increase of the modal density and a high damping of a number of modes of the structure due to the ABH.

International audience; This paper investigates the possibility of reproducing the self-sustained oscillation of the tongue tip in alveolar trills. The interest is to study the articulatory and phonatory configurations that are required to produce alveolar trills. Using a realistic geometry of the vocal tract, derived from cineMRI data of a real speaker, the paper studies the mechanical behavior of a lumped two-mass model of the tongue tip. Then, the paper proposes a solution to simulate the incomplete occlusion of the vocal tract during linguopalatal contacts by adding a lateral acoustic waveguide. Finally, the simulation framework is used to study the impact of a set of parameters on the characteristic features of the produced alveolar trills. It shows that the production of trills is favored when the distance between the equilibrium position of the tongue tip and the hard palate in the alveolar zone is less than 1 mm, but without linguopalatal contact, and when the glottis is fully adducted.

International audience; The paper presents a method to acquire articulatory data from a sequence of MRI images at a high framerate. The acquisition rate is enhanced by partially collecting data in the kt-space. The combination of compressed sensing technique, along with homodyne reconstruction, enables the missing data to be recovered. The good reconstruction is guaranteed by an appropriate design of the sampling pattern. It is based on a pseudo-random Cartesian scheme, where each line is partially acquired for use of the homodyne reconstruction, and where the lines are pseudo-randomly sampled: central lines are constantly acquired and the sampling density decreases as the lines are far from the center. Application on real speech data show that the framework enables dynamic sequences of vocal tract images to be recovered at a framerate higher than 30 frames per second and with a spatial resolution of 1 mm. A method to extract articulatory data from contour identification is presented. It is intended, in fine, to be used for the creation of a large database of articulatory data.

International audience; Speech production is a dynamic process which relies on the realization of articulatory gestures by the speaker. Its study thus necessitates technologies to acquire articulatory data, which often come from medical imaging. Xrays have been used until the end of last century. These data are films which require processing before being exploited. Xarticulators software was developed for this purpose and it enables the manual or automatic extraction of articulator contours. This communication presents the database collected at IPS, the protocol used to process data and the construction of articulatory models intended to approximate the vocal tract shape for vowels and consonants. We detail the improvement of the model in the cases of consonants so as to offer a good geometric precision at the place of articulation, and thus a correct acoustic signal with articulatory synthesis. This communications ends with the presentation of articulatory copy synthesis tools which enable realistic speech to be synthesized from images of an Xray film of the vocal tract.; La production de la parole est un phénomène dynamique qui repose sur la réalisation de gestes articulatoires par le locuteur. Son étude nécessite donc des moyens d'acquisition, souvent issus de l'imagerie médicale comme c'est le cas avec l'imagerie par rayons X qui a été utilisée pour construire des bases de données jusqu'à la fin du siècle dernier. Ces données se présentent sous la forme de cinéradiographies qui requièrent un dépouillement préalable à leur exploitation. Le logiciel Xarticulators a été développé à cette fin et il permet notamment d'extraire manuellement ou automatiquement les contours des articulateurs. Cette communication présente la constitution des bases de données de l'IPS, le protocole de dépouillement et la construction de modèles articulatoires destinés à approcher la forme du conduit vocal pour les voyelles et les consonnes. Nous détaillons en particulier l'amélioration du modèle dans le cas consonnes pour qu'il puisse être précis au point d'articulation et donner une acoustique correcte lors de la synthèse. Cette communication se termine avec la présentation des outils de synthèse articulatoire par copie qui permettent de synthétiser de la parole réaliste à partir des images d'un film aux rayons X du conduit vocal.

The mechanical characteristics of wood panels used by instrument makers are related to numerous factors, including the nature of the wood or characteristic of the wood sample (direction of fibers, micro-structure nature). This leads to variations in the Young's modulus, the mass density, and the damping coefficients. Existing methods for estimating these parameters are not suitable for instrument makers, mainly because of the need of expensive experimental setups, or complicated protocols, which are not adapted to a daily practice in a workshop. In this paper, a method for estimating the Young's modulus, the mass density, and the modal loss factors of flat panels, requiring a few measurement points and an affordable experimental setup, is presented. It is based on the estimation of two characteristic quantities: the modal density and the mean mobility. The modal density is computed from the values of the modal frequencies estimated by the subspace method ESPRIT (Estimation of Signal Parameters via Rotational Invariance Techniques), associated with the signal enumeration technique ESTER (ESTima-tion of ERror). This modal identification technique is proved to be robust in the low-and the mid-frequency domains, i.e. when the modal overlap factor does not exceed 1. The estimation of the modal parameters also enables the computation of the modal loss factor in the low-and the mid-frequency domains. An experimental fit with the theoretical expressions for the modal density and the mean mobility enables an accurate estimation of the Young's modulus and the mass density of flat panels. A numerical and an experimental study show that the method is robust, and that it requires solely a few measurement points.

Since the 1960s and 1970s, researchers have proposed simplified models using only a few parameters to describe the vibro-acoustical behavior of string instruments in the low-frequency range. This paper presents a method for deriving and estimating a few important parameters or features describing the mechanical behavior of classical guitars over a broader frequency range. These features are selected under the constraint that the measurements may readily be made in the workshop of an instrument maker. The computations of these features use estimates of the modal parameters over a large frequency range, made with the high-resolution subspace ESPRIT algorithm (Estimation of Signal Parameters via Rotational Invariant Techniques) and the signal enumeration technique ESTER (ESTimation of ERror). The methods are applied to experiments on real metal and wood plates and numerical simulations of them. The results on guitars show a nearly constant mode density in the mid- and high-frequency ranges, as it is found for a flat panel. Four features are chosen as characteristic parameters of this equivalent plate: Mass, rigidity, characteristic admittance, and the mobility deviation. Application to a set of 12 guitars indicates that these features are good candidates to discriminate different classes of classical guitars.

The “gothic” style of harp was popular across most of Europe from the late medieval period to the Renaissance. These harps have a one-piece, longitudinally oriented hardwood soundboard, as opposed to the transverse-oriented bonded softwood structure on a modern instrument. In addition, the one-piece back is flexible, whereas the back of a modern instrument is a rigid molded shell. To study the gothic harp, one was constructed from plans created by the Boston Museum of Fine Arts from a late German model in their collection. The vibrational behaviors of the soundboard and soundbox were measured at various stages of construction. The completed instrument was subjected to modal analysis and radiativity measurements. The sound radiation is dominated by two breathing modes at 188 and 273 Hz, each with strong motion of the back, and modes around 350 Hz. Taken together, these modes function like the A0/T1 resonance pairs seen in the soundboxes of other instruments, and a comparison is made with the guitar. Also observed is that as the frequency increases, radiation is emitted from higher up the soundboard, and from higher soundholes. This feature has been observed in other harps, and is a consequence of the harp family’s unique geometry.

Flexural waves propagating in a beam can be efficiently absorbed if one extremity is tapered with a power law profile and covered by a very thin viscoelastic layer (V. Krylov, JSV 274 (2004), 605-619). Such a terminaison induces an effect known as ?the acoustic black hole effect? (ABH), which is resulting from properties of propagation of flexural wave in beams having non homogeneous thicknesses: if the thickness decreases locally, flexural waves slow down and the amplitude of the displacement field increases, leading to efficient energy dissipation if an absorbing layer is placed where the thickness is minimum (V. Georgiev et al., JSV 330 (2011), 2497-2508). Absorption of the ABH terminaison is estimated thanks to the direct measurement of the reflexion coefficient, using a wave decomposition technique. Experimental modal Analysis of a ABH beam can be performed using a high resolution technique which permits to estimate the modal density. Analysis of these experimental results is performed thanks to a model based on the finite difference method. It is shown that local transverse modes are playing an important role in the absorption properties of ABH.

The gothic style of harp was popular across most of Europe from the late medieval period to the Renaissance. To study the vibroacoustic behavior of gothic harps, one was constructed from plans created by the Boston Museum of Fine Arts from a late German model in their collection. The vibrational behaviors of the soundboard and soundbox were measured at various stages of construction of the instrument. Once complete, the instrument was subjected to modal analysis and radiativity measurements using Weinreich’s method. The sound radiation of this harp is dominated by two breathing modes at 188 and 273 Hz and higher modes around 350 Hz, which together function like the A0/T1 resonance pairs seen in the soundboxes of many other instruments including modern concert harps. As the frequency increases, radiation is emitted from higher up the soundboard and from higher soundholes, as has been observed in other harps. Unlike modern instruments, the gothic harp’s thin back plays a large role in sound production.