The traditional way of estimating the formant frequencies from articulatory data presupposes knowledge of how the vocal tract cross-sectional area varies for a given articulatory shape (Fant 1960/1970). Accordingly, in order to derive the formant pattern of a given articulation, the three-dimensional shape of the vocal tract (VT) needs to be known. In the past cross-sectional areas have typically been derived by means of ‘d-to-A rules’ that use the mid-sagittal cross-distance d at each point along the VT to produce a corresponding cross-sectional area A. X-ray and MRI data have been used to calibrate such rules (Heinz & Stevens 1964, Sundberg et al. 1987, Ericsdotter 2005). Although this procedure has produced many useful results it is time consuming and laborious. It is speaker-specific. It presupposes access to information on the three-dimensional shape of the VT, which is not experimentally readily accessible. Such observations raise the question whether sufficiently accurate alternative approaches can be developed. Is it possible to go straight from articulatory data to formant frequencies without having to construct a cross-sectional area function? If such methods could be developed it would have many uses both in phonetics and practical applications. This paper reports an attempt to map the time variations of selected articulatory parameters from X-ray profiles directly on the formant tracks using multiple regression analysis. Preliminary results for F1 indicate that multiple regression analysis can indeed be useful for making such predictions. The prospects of extending the present analyses to other formants are discussed.

Selon la littérature, la dyslexie est une difficulté d’apprentissage de la lecture, non liée à un retard mental, à un déficit sensoriel et ni à l’environnement social ou familial défavorisé. L’enfant dyslexique présente un écart de 18 mois à 24 mois par rapport aux réalisations scolaires d'un sujet en lecture, et ses potentialités intellectuelles mesurées par une échelle d'intelligence (Q.I.). Il souffre d’un dysfonctionnement des structures cérébrales entravant ses capacités cognitives avec une mauvaise identification des mots. Le diagnostic repose sur l’évaluation de l’habileté de la parole, de la lecture et de l’écriture. Dans ce travail, nous nous sommes intéressés aux enfants présentant une dyslexie phonologique. Cette dernière se caractérise principalement par une altération de la voie phonologique. Ces enfants présentent une sensibilité intracatégorielle supérieure à leur sensibilité intercatégorielle. Si nous acceptons l’hypothèse du couplage entre la perception et la production de la parole, nous pouvons parler d’une perturbation de l’exécution articulatoire de leurs productions qui explique leur retard de parole et leur trouble du langage. En effet, la parole obéit à certaines règles phonétiques et phonologiques ; la variation d’un phonème coarticulé est la source de la difficulté de sa discrimination chez l’enfant dyslexique, car un phonème présente plusieurs allophones. Nous avons supposé que cette variabilité articulatoire d’un seul phonème est la cause principale de leur retard de production et de perception de la parole. De ce fait, nous avons procédé à une analyse de leur production verbale à la recherche des marqueurs articulatoires qui diffèrent de ceux des enfants du groupe de contrôle. Quatre expériences ont été effectuées traitant des séquences de types (CV, CVCV, VCVCV et VCVCCV) en lecture de mots et de pseudo mots, à haute voix et à débit normal. Les paramètres temporels étudiés en ce qui concerne les indices acoustiques pour le segment consonantique ont été pour la consonne occlusive sourde, la durée du VTT, le silence acoustique et le VOT et, pour la consonne occlusive sonore, l’occlusion et le VOT. Nous avons aussi mesuré les segments vocaliques adjacents. Ces différentes analyses révèlent, lors de l’identification de mots, que la voie d’assemblage peut être partiellement opérationnelle chez les enfants dyslexiques, une déformation de mot partielle ou totale, des oppositions de sonorité ou de lieu d’articulation problématiques. A noter aussi des durées segmentales plus élevées chez l’enfant dyslexique que chez l’enfant normo-lecteur traduisant une lenteur dans la progression de la réalisation articulatoire nécessitant une durée supplémentaire. Ces paramètres ne suivent pas souvent une cohérence dans leurs proportions par rapport aux exigences imposées par les caractéristiques du lieu d’articulation et de la qualité du signal. Cependant nous constatons qu’ils peuvent préserver leur intelligibilité. Cette étude expérimentale a permis de mettre en évidence la notion de couplage entre la perception et le trouble articulatoire, voire le retard de la parole et du langage. / Available literature suggests that dyslexia is a learning disability that is neither linked to mental retardation, not to sensory disorder nor to deprived social background. The dyslexic child shows a difference of 18 to 24 months between the academic achievement of a reading subject and his or her intellectual potentials as measured by an intelligence scaling (IQ). The child suffers from brain structure dysfunction which impedes his cognitive abilities resulting in wrong word identification. We also know that the diagnosis is based on the evaluation of speech, reading and writing abilities. In our work, we focus on child phonological dyslexia. The latter is mainly characterized by impaired phonological route due to poor development of phonological ability. In reality, dyslexic children have a higher intra-categorical sensitivity rather than an inter-categorical perception. If we accept the idea of coupling between speech production and perception, we can posit the existence of a perturbation in the articulatory realisation of speech produced by dyslexic children which explains both their delayed speech and language disorder. Indeed, speech is governed by a number of phonetic and phonological rules; the variation of a coarticulated phoneme is at the root of difficulty of phoneme discrimination encountered by dyslexic children. This coarticulated phoneme has several allophones. We suggest that the articulatory variability of a single phoneme is the main cause of speech impairment. We performed an analysis of their verbal production in search of the articulatory markers that differ from children with no learning disability. With this in mind, we conducted four studies dealing with different sequences (CV, CVCV, and VCVCV VCVCCV). Tasks included reading words and pseudo-words, aloud and at normal speech rate. The measured parameters studied as regards acoustic cues for stops were VTT, the acoustic silent phase and VOT for voiceless stops; occlusion and VOT for the voiced stops, and duration for flanking vowels. During word identification, analyses reveal that the phonological route may be partially operational for our subjects. They also reveal partial or total word deformation; voicing contrast and place of articulation contrasts were also affected. Note should be taken of the higher measures found in dyslexic children than in normal readers, reflecting a slower gestures in their articulatory realizations due to additional time needed in these children. In addition, these parameters do not often follow a coherent pattern of proportional values relative to the requirements imposed by the characteristics of place of articulation and signal quality. However, we find that word intelligibility could be maintained. This experimental study has demonstrated the link between perception and articulation disorder, and to some extent speech and language production impairment.

Articulatory synthesis is based on modeling various physical phenomena of speech production, including sound radiation from the mouth. With regard to sound radiation, the most common approach is to approximate it in terms of a simple spherical source of strength equal to the mouth volume velocity. However, because this approximation is only valid at very low frequencies and does not account for the diffraction by the head and the torso, we simulated two alternative radiation characteristics that are potentially more realistic: the radiation from a vibrating piston in a spherical baffle, and the radiation from the mouth of a detailed model of the human head and torso. Using the articulatory speech synthesizer VocalTractLab, a corpus of 10 sentences was synthesized with the different radiation characteristics combined with three different phonation types. The synthesized sentences were acoustically compared with natural recordings of the same sentences in terms of their long-term average spectra (LTAS), and evaluated in terms of their naturalness and intelligibility. The intelligibility was not affected by the type of radiation characteristic. However, it was found that the more similar their LTAS was to real speech, the more natural the synthetic sentences were perceived to be. Hence, the naturalness was not directly determined by the realism of the radiation characteristic, but by the combined spectral effect of the radiation characteristic and the voice source. While the more realistic radiation models do not per se improve synthesis quality, they provide new insights in the study of speech production and articulatory synthesis.