Search

In speech production, the anatomical morphology forms the substrate on which the speakers build their articulatory strategy to reach specific articulatory-acoustic goals. The aim of this study is to characterize morphological inter-speaker variability by building a shape model of the full vocal tract including hard and soft structures. Static magnetic resonance imaging data from 41 speakers articulating altogether 1947 phonemes were considered, and the midsagittal articulator contours were manually outlined. A phoneme-independent average-articulation representative of morphology was calculated as the speaker mean articulation. A principal component analysis-driven shape model was derived from average-articulations, leading to five morphological components, which explained 87% of the variance. Almost three-quarters of the variance was related to independent variations of the horizontal oral and vertical pharyngeal lengths, the latter capturing male-female differences. The three additional components captured shape variations related to head tilt and palate shape. Plane wave propagation acoustic simulations were run to characterize morphological components. A lengthening of 1 cm of the vocal tract in the vertical or horizontal directions led to a decrease in formant values of 7%–8%. Further analyses are required to analyze three-dimensional variability and to understand the morphological-acoustic relationships per phoneme. Average-articulations and model code are publicly available ( https://github.com/tonioser/VTMorphologicalModel ).

The animation of articulated characters is a central problem in the computer graphics field. Skeletal animation techniques define a workflow which has proven to be effective for boundary representations (B-Reps). This paper extends the classical skeletal animation pipeline to deal with characters internal tissues. In contrast to most common approaches, the proposed technique automates all the stages of this workflow. Well known skinning algorithms, such as Linear Blending Skinning, Dual Quaternion Skinning or Optimized Centers of Rotation were adapted to allow the use of our technique in applications where interactivity is required. The pipeline proposed in this paper can be used in many computer graphics systems such as games or educational applications to visualize and animate the internal anatomy of a virtual character at interactive rates.

The various speech sounds of a language are obtained by varying the shape and position of the articulators surrounding the vocal tract. Analyzing their variations is crucial for understanding speech production, diagnosing speech disorders and planning therapy. Identifying key anatomical landmarks of these structures on medical images is a pre-requisite for any quantitative analysis and the rising amount of data generated in the field calls for an automatic solution. The challenge lies in the high inter- and intra-speaker variability, the mutual interaction between the articulators and the moderate quality of the images. This study addresses this issue for the first time and tackles it by means of Deep Learning. It proposes a dedicated network architecture named Flat-net and its performance are evaluated and compared with eleven state-of-the-art methods from the literature. The dataset contains midsagittal anatomical Magnetic Resonance Images for 9 speakers sustaining 62 articulations with 21 annotated anatomical landmarks per image. Results show that the Flat-net approach outperforms the former methods, leading to an overall Root Mean Square Error of 3.6 pixels/0.36 cm obtained in a leave-one-out procedure over the speakers. The implementation codes are also shared publicly on GitHub.

Speech communication relies on articulatory and acoustic codes shared between speakers and listeners despite inter-individual differences in morphology and idiosyncratic articulatory strategies. This study addresses the long-standing problem of characterizing and modelling speaker-independent articulatory strategies and inter-speaker articulatory variability. It explores a multi-speaker modelling approach based on two levels: statistically-based linear articulatory models, which capture the speaker-specific articulatory variability on the one hand, are in turn controlled by a speaker model, which captures the inter-speaker variability on the other hand. A low dimensionality speaker model is obtained by taking advantage of the inter-speaker correlations between morphology and strategy. To validate this approach, contours of the vocal tract articulators were manually segmented on midsagittal MRI data recorded from 11 French speakers uttering 62 vowels and consonants. Using these contours, multi-speaker models with 14 articulatory components and two morphology and strategy components led to overall variance explanations of 66%–69% and root-mean-square errors of 0.36–0.38 cm obtained in leave-one-out procedure over the speakers. Results suggest that inter-speaker variability is more related to the morphology than to the idiosyncratic strategies and illustrate the adaptation of the articulatory components to the morphology.Speech communication relies on articulatory and acoustic codes shared between speakers and listeners despite inter-individual differences in morphology and idiosyncratic articulatory strategies. This study addresses the long-standing problem of characterizing and modelling speaker-independent articulatory strategies and inter-speaker articulatory variability. It explores a multi-speaker modelling approach based on two levels: statistically-based linear articulatory models, which capture the speaker-specific articulatory variability on the one hand, are in turn controlled by a speaker model, which captures the inter-speaker variability on the other hand. A low dimensionality speaker model is obtained by taking advantage of the inter-speaker correlations between morphology and strategy. To validate this approach, contours of the vocal tract articulators were manually segmented on midsagittal MRI data recorded from 11 French speakers uttering 62 vowels and consonants. Using these contours, multi-speaker mode...

International audience; Speech production can be analysed in terms of universal articulatory-acoustic phonemic units shared between speakers. However, speakers’ morphological differences and idiosyncratic articulatory strategies lead to large inter-speaker articulatory variability. Relationships between strategy and morphology have already been pinpointed in the literature. This study aims thus at generalising existing results on a larger database for the entire vocal tract (VT) and at quantifying phoneme-specific inter-speaker articulatory invariants. Midsagittal MRI of 11 French speakers for 62 vowels and consonants were recorded and VT contours manually edited. A procedure of normalisation of VT contours between speakers, based on principal component analysis of mean VT contours, led to an overall reduction of inter-speaker VT contours variance of 84%. On the opposite, the sagittal function (i.e. the transverse sagittal distance along the VT midline), which is the main determinant of the acoustic output, had an overall amplitude variance decrease of only 18%, suggesting that the speakers adapt their strategy to their morphology to achieve proper acoustic goals. Moreover, articulatory invariants were identified on the sagittal variance distribution along the VT as the regions with lower variability. These regions correspond to the classical places of articulation and are associated with higher acoustic sensitivity function levels.

This paper analyzes the influence of pregnancy stage and fetus position on the whole-body and brain exposure of the fetus to radiofrequency electromagnetic fields. Our analysis is performed using semi-homogeneous pregnant woman models between 8 and 32 weeks of amenorrhea. By analyzing the influence of the pregnancy stage on the environmental whole-body and local exposure of a fetus in vertical position, head down or head up, in the 2100 MHz frequency band, we concluded that both whole-body and average brain exposures of the fetus decrease during the first pregnancy trimester, while they advance during the pregnancy due to the rapid weight gain of the fetus in these first stages. From the beginning of the second trimester, the whole-body and the average brain exposures are quite stable because the weight gains are quasi proportional to the absorbed power increases. The behavior of the fetus whole-body and local exposures during pregnancy for a fetus in the vertical position with the head up were found to be of a similar level, when compared to the position with the head down they were slightly higher, especially in the brain.

International audience; Fetal dosimetry studies require the development of accurate numerical 3Dmodels of the pregnant woman and the fetus. This paper proposes a 3D articulatedfetal growth model covering the main phases of pregnancy and a pregnantwoman model combining the utero-fetal structures and a deformable non-pregnantwoman body envelope. The structures of interest were automatically or semi-automatically (depending on the stage of pregnancy) segmented from a databaseof images and surface meshes were generated. By interpolating linearly betweenfetal structures, each one can be generated at any age and in any position.A method is also described to insert the utero-fetal structures in the maternalbody. A validation of the fetal models is proposed comparing a set of biometricmeasurements to medical reference charts. The usability of the pregnant womanmodel in dosimetry studies is also investigated, with respect to the inuence ofthe abdominal fat layer.

Three-dimensional (3D) reconstruction of lower limbs is essential for surgical planning and clinical outcome evaluation. 3D reconstruction from biplanar calibrated radiographs may be an alternative to irradiation issues of CT-scan. A previous study proposed a two-step reconstruction method based on parametric models and statistical inferences leading to a fast Initial Solution (IS) followed by manual adjustments. This study aims to improve the IS using a new 3D database, a novel parametric model of the tibia and a different regression approach. The IS was evaluated in terms of shape accuracy on 9 lower limbs and reproducibility of clinical measurements on 22 lower limbs. Reconstruction time was also evaluated. Comparison to the previous method showed an improvement of the IS in terms of shape accuracy (1.3 vs. 1.6 and 2 mm respectively for both femur and tibia) and reproducibility of clinical measurements (i.e. 3.1° vs. 8.3° for neck-shaft-angle; 4.2° and 5° vs. 5° and 6° for tibial and femoral torsion respectively). The proposed approach constitutes a considerable step towards an automatic 3D reconstruction of lower limb.

Fetal dosimetry studies require the development of accurate 3D models of the fe- tus. This paper proposes a 3D articulated fe- tal growth model including skeleton, body en- velope, brain and lungs based on medical im- ages of 10 dierent fetuses acquired in clinical routine. The structures of interest were semi- manually segmented from the images and sur- face meshes were generated. A generic mesh of each structure has been deformed towards the segmented ones. By interpolating linearly be- tween the subjects of the database, each struc- ture can be estimated at any age and in any position. This process results in an automated model, the operator being only required to specify the age and position of the desired es- timated fetus.

Scientists seek to use fossil and archaeological evidence to constrain models of the coevolution of human language and tool use. We focus on Neanderthals, for whom indirect evidence from tool use and ancient DNA appears consistent with an adaptation to complex vocal-auditory communication. We summarize existing arguments that the articulatory apparatus for speech had not yet come under intense positive selection pressure in Neanderthals, and we outline some recent evidence and analyses that challenge such arguments. We then provide new anatomical results from our own attempt to reconstruct vocal tract (VT) morphology in Neanderthals, and document our simulations of the acoustic and articulatory potential of this reconstructed Neanderthal VT. Our purpose in this paper is not to polarize debate about whether or not Neanderthals were human-like in all relevant respects, but to contribute to the development of methods that can be used to make further incremental advances in our understanding of the evolution of speech based on fossil and archaeological evidence.

International audience; An original three-dimensional 3D linear articulatory model of the velum and nasopharyngeal wall has been developed from magnetic resonance imaging MRI and computed tomography images of a French subject sustaining a set of 46 articulations, covering his articulatory repertoire. The velum and nasopharyngeal wall are represented by generic surface triangular meshes fitted to the 3D contours extracted from MRI for each articulation. Two degrees of freedom were uncovered by principal component analysis: first, VL accounts for 83% of the velum variance, corresponding to an oblique vertical movement seemingly related to the levator veli palatini muscle; second, VS explains another 6% of the velum variance, controlling a mostly horizontal movement possibly related to the sphincter action of the superior pharyngeal constrictor. The nasopharyngeal wall is also controlled by VL for 47% of its variance. Electromagnetic articulographic data recorded on the velum fitted these parameters exactly, and may serve to recover dynamic velum 3D shapes. The main oral and nasopharyngeal area functions controlled by the articulatory model, complemented by the area functions derived from the complex geometry of each nasal passage extracted from coronal MRIs, were fed to an acoustic model and gave promising results about the influence of velum movements on the spectral characteristics of nasals.

Visualising volumetric medical images such as computed tomography and magnetic resonance imaging (MRI) on picture archiving and communication systems (PACS) clients is often achieved by image browsing in sagittal, coronal or axial views or three-dimensional (3D) rendering. This latter technique requires fine thresholding for MRI. On the other hand, computing virtual radiograph images, also referred to as digitally reconstructed radiographs (DRR), provides in a single two-dimensional (2D) image a complete overview of the 3D data. It appears therefore as a powerful alternative for MRI visualisation and preview in PACS. This study describes a method to compute DRR from T1-weighted MRI. After segmentation of the background, a histogram distribution analysis is performed and each foreground MRI voxel is labeled as one of three tissues: cortical bone, also known as principal absorber of the X-rays, muscle and fat. An intensity level is attributed to each voxel according to the Hounsfield scale, linearly related to the X-ray attenuation coefficient. Each DRR pixel is computed as the accumulation of the new intensities of the MRI dataset along the corresponding X-ray. The method has been tested on 16 T1-weighted MRI sets. Anterior-posterior and lateral DRR have been computed with reasonable qualities and avoiding any manual tissue segmentations. This proof-of-concept holds for research application for use in clinical PACS.

International audience; In this paper, we describe a comparison of the acoustic characteristics of one-dimensional and three-dimensional models of vocal tracts with nasal coupling. One-dimensional acoustic propagation is computed using an electric analog model. A finite element method is used for three-dimensional acoustic simulation. The comparison of these two approaches involves the vocal-tract shape of two subjects, one Japanese male and one French male pronouncing the vowel /a/ in their native language. Results show that the pole/zero pairs ascribed to the nasal coupling for both simulations appeared at almost the same frequency, at least below 2 kHz. Little difference between the one-dimensional and three-dimensional simulations in the transfer functions for the French subject is observed, since the three-dimensional mesh for the French subject is smoother. An extra pole exists in the transfer function of the three-dimensional model for the Japanese subject, possibly caused by the asymmetric structure of the laryngeal cavity. In the three-dimensional distribution of the active sound intensity vectors for the French subject, sound energy fluxes circulate between oral and nasal cavities coupled in the vicinity of the lips and nostrils.

International audience

International audience; In the context of patient-specific 3D bone reconstruction, enhancing the surface with cortical thickness (COT) opens a large field of applications for research and medicine. This functionality calls for database analysis for better knowledge of COT. Our study provides a new approach to reconstruct 3D internal and external cortical surfaces from computer tomography (CT) scans and analyses COT distribution and variability on a set of asymptomatic femurs. The reconstruction method relies on a short (∼5 min) initialisation phase based on 3D reconstruction from biplanar CT-based virtual X-rays and an automatic optimisation phase based on intensity-based cortical structure detection in the CT volume, the COT being the distance between internal and external cortical surfaces. Surfaces and COT show root mean square reconstruction errors below 1 and 1.3 mm. Descriptions of the COT distributions by anatomical regions are provided and principal component analysis has been applied. The first mode, 16–50% of the variance, corresponds to the variation of the mean COT around its averaged shape; the second mode, 9–28%, corresponds to a fine variation of its shape. A femur COT model can, therefore, be described as the averaged COT distribution in which the first parameter adjusts its mean value and a second parameter adjusts its shape.

La fin du XXe et le tout début de ce siècle révèlent une véritable réarticulation des recherches dans le domaine de l’émergence de la parole et du langage. Le naturalisme, qui est au centre de cette approche, se propose de décrire les relations entre la biologie (au sens très large du terme) d’une part, la parole et le langage, d’autre part, par une accumulation d’hypothèses, de données et de preuves formulées et établies grâce à de multiples collaborations interdisciplinaires. Comme pour les travaux sur l’origine de l’Homme (la découverte d’un nouveau fossile entraînant souvent une remise en question des théories précédentes), on assiste à un foisonnement théorique qui entraîne parfois des développements très hypothétiques, s’appuyant sur des résultats fragiles et sur trop peu de données, proposés dans des disciplines connexes mais non maîtrisées ou trop simplifiées. C’est pourquoi les bilans réguliers, les mises en perspectives critiques ne nous semblent pas superflus.Dans un premier temps nous proposerons une classification qui permet une lecture des différentes théories proposées depuis un demi-siècle (théorie push-pull). Dans l’état actuel des connaissances, il n’est pas possible d’inférer quand nos ancêtres, voire nos lointains ont acquis la faculté de langage et de parole : le contrôle des articulateurs, la coordination entre le larynx et le conduit vocal, la phonologie, la syntaxe, la sémantique et la récursivité. Parmi les questions qui se posent, il en est une qui reste sans réponse : pourquoi notre espèce est actuellement la seule à posséder langue et parole ? De nombreuses questions font partie des problèmes mal posés, comme le sont les questions du type: Quelle(s) langue(s) parlaient nos prédécesseurs ? Possédaient-ils une langue unique ? En effet, on ne dispose pas (pour le moment) de suffisamment de données pour pouvoir y répondre. Peut-être même que ces questions ne trouveront pas de solution. Actuellement, il est quand même possible de répondre à la question suivante : si nous supposons que nos ancêtres (et cousins lointains) contrôlaient leur larynx et leur conduit vocal de la même manière que les Hommes actuels, est-ce que la géométrie de leur conduit leur permettaient de produire les structures sonores qui sont pratiquement présentes dans toutes les langues du monde ?Nous présenterons ensuite nos travaux qui participent à la nouvelle réarticulation avec une approche véritablement axée sur la pluridisciplinarité. Ils s’inscrivent dans le domaine des relations entre la morphologie des organes de la production de la parole et son contrôle. Nous présentons de nouveaux résultats concernant la croissance du conduit vocal de la naissance à l’âge adulte puis des reconstructions du conduit vocal pour des fossiles qui couvrent la période de 10.000 ans à un million et demi d’années BP (Before Present). À partir du crâne, de la mandibule et des vertèbres cervicales nous essaierons de manière plausible (1) de localiser l’os hyoïde, support de la langue, et la position de la glotte, (2) de reconstituer un conduit vocal, à l’aide d’un modèle articulatoire, (3) d’induire les possibilités acoustiques de tous ces conduits. En combinant phylogenèse et ontogenèse il est possible de représenter l’anatomie du tractus en synthétisant deux remodelages qui renvoient à l’ontogenèse et à la phylogenèse. Nous montrerons que tous ces conduits ont les mêmes potentialités acoustiques, ils peuvent produire les voyelles /i a u/ qui sont pratiquement présentes dans toutes les langues du monde : un triangle à l’intérieur duquel se situent toutes les autres voyelles. Quand aux consonnes les plus fréquentes /p t k/, /b d g/ elles sont aussi à la portée de tous ces conduits vocaux à partir de gestes de fermeture dans des régions précises (lèvres, zone alvéodentale, zone vélaire). De nouvelles pistes de recherche sont proposées qui tendraient à montrer qu’il y a vraisemblablement plusieurs centaines de milliers, voire plusieurs millions d’années que le conduit vocal présente une morphologie favorable à l’émergence et à la production de la parole. Un cadre est posé dans lequel les capacités orofaciales nécessaire à la parole pourraient être reliées au mécanisme précurseur d’ingestion (mastication-déglutition).The end of the XXth century and the beginning of this century saw a reorganization of the researches in the field of speech and language emergence (SLE). Naturalism is the kernel of this new approach. It consists in describing the relations between biological aspects (in every sense of the word) on the one hand and speech and language, on the other hand, by an accumulation of hypotheses and evidence derived from a huge range of data collected thanks to interdisciplinary collaborations.As is the case for researches on the origin of Man, a theoretical profusion of hypotheses has arisen which sometimes leads to very hypothetical developments, based on fragile results and on too little data, and proposed in related but not fully mastered or too much simplified disciplines. This is why regular critical overviews do not seem superfluous. First, we propose a classification (push and pull theory) that provides a new reading of the various theories which have been proposed for half a century. In the present state of knowledge it is not possible to infer when our ancestors acquired the FacultyofSpeechandLanguageandSpeech: control of speech articulators, coordination between larynx and vocal tract, phonology, syntax, semantic and recursivity. Among old unsolved questions: Why is our species alone in having speech and language? Many others questions are (for the moment?) ill posed problems: we do not have sufficiently data to answer. Perhaps these questions will remain unsolved. But we think that the following question can be solved: If we suppose that our ancestors (and distant cousins) controlled their larynx and vocal tract in the same way as present-day humans, did the geometry of their vocal tract allow them to produce the universal sound structures of the languages spoken today? We analyzed 31 skulls from now to 1.5 Ma (millions years) BP (Before Present) for fossil hominids available at the Muséedel’Homme in Paris or in the literature: (1) 10-30 ka BP: modern humans: Paleolithic; (2) 90-200 ka BP: anatomically modern humans; (3) 45-90 ka BP: Neanderthals; (4) 1.5 Ma BP: Homoergaster; These skulls are all well kept and possess a jaw in the majority of cases but the vertebral column has been reconstituted. We attempt to: (1) Localize hyoid bone and then glottis position; (2) Reconstitute a vocal tract model in a plausible way using an articulatory model; (3) Quantify the acoustic capabilities of this reconstituted vocal tract. For this purpose, we combine phylogenesis and ontogenesis. We are in a position to state that our ancestors and distant cousins were equipped with a vocal tract that could produce the same variety of vowel sounds as we can today: the vowels /i a u/. The vocal tract morphology has been favorable to the emergence and production of speech since several hundreds of thousands of years. But how to know to what extent they mastered the control skills needed to produce speech? New lines of research are proposed in which orofacial abilities necessary to the emergence of speech are linked to a precursor mechanism dedicated to feeding (masticating-swallowing movements).

Three-dimensional (3D) reconstruction of the skeleton from biplanar X-rays relies on scarce information digitalised by an operator on both frontal and lateral radiographs. In clinical routine, difficulties occur for non-skilled operators to discriminate the medial from the lateral femur condyle on the lateral view. Our study proposes an algorithm able to detect automatically a possible inversion of the two condyles by the operator at an early stage of the reconstruction process. It relies on the computation of two 3D femur surfaces, one directly from the operator digitalisation and the other from the same digitalisation with medial and lateral condyles automatically swapped. Pairs of virtual biplanar X-rays are computed for both reconstructions and the closest pair to the original X-rays is selected on the basis of similarity measures, pointing the correct 3D surface. The algorithm shows a success rate higher than 85% for both asymptomatic and pathological femurs whatever the initial condyle digitalisation of the operator, bringing automatically non-skilled operators acting in clinical routine to the level of skilled operators. This study validates moreover the proof-of-concept of automatic shape adjustments of a 3D surface on the basis of similarity measures in the process of 3D reconstruction from biplanar X-rays.

La fin du XXe et le tout debut de ce siecle revelent une veritable rearticulation des recherches dans le domaine de l’emergence de la parole et du langage. Le naturalisme, qui est au centre de cette approche, se propose de decrire les relations entre la biologie (au sens tres large du terme) d’une part, la parole et le langage, d’autre part, par une accumulation d’hypotheses, de donnees et de preuves formulees et etablies grâce a de multiples collaborations interdisciplinaires. Comme pour les travaux sur l’origine de l’Homme (la decouverte d’un nouveau fossile entrainant souvent une remise en question des theories precedentes), on assiste a un foisonnement theorique qui entraine parfois des developpements tres hypothetiques, s’appuyant sur des resultats fragiles et sur trop peu de donnees, proposes dans des disciplines connexes mais non maitrisees ou trop simplifiees. C’est pourquoi les bilans reguliers, les mises en perspectives critiques ne nous semblent pas superflus.Dans un premier temps nous proposerons une classification qui permet une lecture des differentes theories proposees depuis un demi-siecle (theorie push-pull). Dans l’etat actuel des connaissances, il n’est pas possible d’inferer quand nos ancetres, voire nos lointains ont acquis la faculte de langage et de parole : le controle des articulateurs, la coordination entre le larynx et le conduit vocal, la phonologie, la syntaxe, la semantique et la recursivite. Parmi les questions qui se posent, il en est une qui reste sans reponse : pourquoi notre espece est actuellement la seule a posseder langue et parole ? De nombreuses questions font partie des problemes mal poses, comme le sont les questions du type: Quelle(s) langue(s) parlaient nos predecesseurs ? Possedaient-ils une langue unique ? En effet, on ne dispose pas (pour le moment) de suffisamment de donnees pour pouvoir y repondre. Peut-etre meme que ces questions ne trouveront pas de solution. Actuellement, il est quand meme possible de repondre a la question suivante : si nous supposons que nos ancetres (et cousins lointains) controlaient leur larynx et leur conduit vocal de la meme maniere que les Hommes actuels, est-ce que la geometrie de leur conduit leur permettaient de produire les structures sonores qui sont pratiquement presentes dans toutes les langues du monde ?Nous presenterons ensuite nos travaux qui participent a la nouvelle rearticulation avec une approche veritablement axee sur la pluridisciplinarite. Ils s’inscrivent dans le domaine des relations entre la morphologie des organes de la production de la parole et son controle. Nous presentons de nouveaux resultats concernant la croissance du conduit vocal de la naissance a l’âge adulte puis des reconstructions du conduit vocal pour des fossiles qui couvrent la periode de 10.000 ans a un million et demi d’annees BP (Before Present). A partir du crâne, de la mandibule et des vertebres cervicales nous essaierons de maniere plausible (1) de localiser l’os hyoide, support de la langue, et la position de la glotte, (2) de reconstituer un conduit vocal, a l’aide d’un modele articulatoire, (3) d’induire les possibilites acoustiques de tous ces conduits. En combinant phylogenese et ontogenese il est possible de representer l’anatomie du tractus en synthetisant deux remodelages qui renvoient a l’ontogenese et a la phylogenese. Nous montrerons que tous ces conduits ont les memes potentialites acoustiques, ils peuvent produire les voyelles /i a u/ qui sont pratiquement presentes dans toutes les langues du monde : un triangle a l’interieur duquel se situent toutes les autres voyelles. Quand aux consonnes les plus frequentes /p t k/, /b d g/ elles sont aussi a la portee de tous ces conduits vocaux a partir de gestes de fermeture dans des regions precises (levres, zone alveodentale, zone velaire). De nouvelles pistes de recherche sont proposees qui tendraient a montrer qu’il y a vraisemblablement plusieurs centaines de milliers, voire plusieurs millions d’annees que le conduit vocal presente une morphologie favorable a l’emergence et a la production de la parole. Un cadre est pose dans lequel les capacites orofaciales necessaire a la parole pourraient etre reliees au mecanisme precurseur d’ingestion (mastication-deglutition).

National audience; L'article s'intéresse au thème de l'exception et des universaux linguistiques au niveau des tendances générales des structures phonologiques dans les langues du monde. S'appuyant sur les résultats déjà obtenus dans le domaine des systèmes vocaliques (plus les voyelles sont distinctes, plus elles ont de chance de survivre), l'article soulève le problème des consonnes qui semblent ne pas obéir à ce principe de distinctivité acoustique maximale. Pour intégrer les structures vocaliques et consonantiques dans une théorie plus générale, l'article utilise des résultats mis en évidence dans les communications en situation de handicap auditif et/ou visuel, afin de montrer que les systèmes consonantiques semblent se constituer sur la base d'un principe de distinctivité bimodale acoustique et visuelle.

Both ontogenetically and phylogenetically feeding tasks precede speaking tasks, two of the major functions of the human vocal tract (VT). Researchers have therefore suggested that speech cyclicity, in particular tongue movements, derives from feeding cyclicity (MacNeilage, 1998; Hiiemae and Palmer 2003). Our study explores this hypothesis using articulatory modelling to determine if the tongue movements in speech are a subset of those in feeding. In a previous study (Serrurier et al., 2008) we used ElectroMagnetic Articulography, however this is an invasive technique that further gives no VT boundaries. To overcome these limitations we here use Digital VideoFluoroscopy (DVF). We wish to explore if we can extract from the raw data and from an articulatory model based on the data, articulations which are geometrically and acoustically close to the quantal vowels /a i u/. Their typical F1-F2 values and their typical geometric shape constitute respectively our acoustic and articulatory targets.

International audience; Volume images of tongue were acquired by MRI from one subject uttering a corpus representative of French allophone articulations. Supplementary images of hard palate, jaw, and hyoid bone were acquired by CT. The three-dimensional tongue surface outline was represented, for each of the 46 articulations of the corpus, by a mesh obtained by fitting a generic mesh to the set of tongue contours traced from the MR images. Jaw and hyoid bone positions were also determined. The set of the 3D coordinates of all vertices of the tongue mesh constituted the variables on which linear component analysis was applied. Six linearly independent components were found to explain 87 % of the variance of the tongue data. The associated parameters that control the linear articulator tongue model are related to jaw and hyoid positions, and to the actions of tongue muscles such as the genioglossus, the hyoglossus or the styloglossus. In addition, it was shown that the full 3D tongue surface is predictable from its 2D midsagittal contour with a mere 13.6 % increase in the overall full 3D reconstruction RMS error, which confirms quantitatively previous results. Finally, the tongue volume was found to depart by at most ±5% from its mean over the corpus, which supports the hypothesis of tongue tissue incompressibility for speech.

In the framework of studies on nasality, we have attempted to develop a 3-D articulatory model of velum. Sets of 25 sagittal MRI images have been collected for one French subject sustaining 46 articulations. Contours of the velum have been manually drawn from these images; a common 3-D triangular mesh has then been fitted to these contours in order to constitute 3-D representations of the velum outline. This resulted in 46 meshes of 5239 vertices defined by their 3-D coordinates. In order to develop a velum model free of the influence of the tongue, the analysis was first limited to the 28 articulations where no tongue - velum contact occurs: it was found that the first component extracted from these data by Principal Component Analysis can explain as much as 79% of the variance of the complete 3-D shape of the velum, while the second component was meaningless in terms of articulatory movements. It was also found that the same model could explain 71% of the variance for the corpus of the 18 articulations where tongue - velum contact occurs. It was finally shown that the vertical coordinate of a single flesh point attached to the downward face of the velum is enough to predict the complete 3-D shape of the free velum.