René Carré

Distinctive regions and modes: a new theory of speech production

Abstract A new explanatory and formal theory in speech production is presented. This theory is founded on perturbation theory and supported by computer simulation. Sensitivity functions which relate variations in formant frequency to small variations in area functions are revised in a manner that allows for their application to large variations in area functions. Available speech data are shown to confirm this theory, and known phonetic, articulatory and acoustic phenomena can be explained and embedded in a simple, global and formal model. The theory provides new insights into some principles of articulatory-acoustic relations and the application of these principles to phonetic theory. In this paper, several implications are studied: the articulatory-acoustic relation is simplified, the quantal nature of speech is confirmed, phonetic universals and phonetic systems are considered from a new point of view, formant transitions are explained and normalized, and an easy to control 9-parameter model proposed. We also consider other speech phenomena as interpreted by this theory such as compensatory and symmetry effects. The theory can also be used to formulate and calculate basic speech dynamic patterns. Several applications and research perspectives are proposed.

The French language database: Defining, planning, and recording a large database

A database of spoken French sounds is described with the methodology of the way in which it was built up (speaker selection, recording conditions, computer control of data acquisition,...). The base will contain various corpus for evaluation and training of recognition and synthesis systems, acoustic, articulatory and prosodic studies. The preliminary labeling and further exploitations of this database are starting in several laboratories.

From an acoustic tube to speech production

Abstract The general aim of this paper is to find out if speech production characteristics may be explained by the physical and permanent properties of an acoustic tube that is 18 cm long. In a way, we intend to generalize the deductive approach used by Lindblom [Phonetic Universal in Vowel Systems, in: J.J. Ohala, J.J. Jaeger (Eds.), Experimental Phonology, Academic Press, Orlando, p. 13] to explain vowel systems from production and perception properties. However, the question here is not to account for vowel systems but rather to explain certain characteristics of the speech production system. In the present research, these characteristics are not observed per se and used as unquestionable “constraints”. In order to answer this question related to the characteristics of the speech production system, the properties of the acoustic tube are first studied to build an acoustic production model having the following specific property: the shape of the tube is deformed to perform maximum acoustic changes, i.e., a minimum area deformation provokes a maximum acoustic variation. Following this approach, a set of distinctive deformation gestures involving corresponding distinctive acoustic changes is obtained and used to set up the intrinsic “phonological” system of the tube designed for communication needs. Then, it is shown that the Distinctive Region Model (DRM) summarizes the main results obtained without any constraints. Finally, model and speech production characteristics are compared. With a limited number of constraints, which can be explained, included into the model (such as a fixed larynx cavity), they fit surprisingly well. Thus, can these speech production characteristics be explained by the proposed deductive approach, i.e., are the main characteristics of the vocal tract, and of speech production in general, consequences of specific deformations of the shape of the tube to perform maximum acoustic changes? Implications of the findings are discussed.

Locus equations in the light of articulatory modeling

This paper examines the significance of the so-called “locus equation” by means of articulatory simulations of V1CV2 utterances with different intergestural timing, and, therefore, with varying degrees of consonant–vowel coarticulation. Movement toward the vowel V2 started (i) at the beginning of the transition to the consonant, (ii) at the beginning of the complete consonant closure, or (iii) at the beginning of the release of the consonant. For each combination of vowels and each consonant, F2 onset of V2 as a function of F2 of V2 was adequately described by straight lines corresponding to locus equations (referred to as first-order locus equations). The findings show that the derived locus equations depend on the consonant place on the one hand and on the degree of coarticulation on the other. The effect of varying intergestural timing was compared with published data on locus equation coefficients for individual speakers using the format of y-intercept plotted versus slope for each place of articulati...

Speech analyzer and synthesizer using vocal tract simulation

A speech analyzer and synthesizer uses simulation of the acoustic behavior of a tube divided into portions having variable sections. The section variations of the various portions of the tube generate sounds corresponding to voiced phonemes when an air flow and pressure source is positioned in analogy with human vocal cords. Using simulation techniques, it is possible to generate the phonemes in the form of electric signals supplied to a loud-speaker. The selection of tube portion lengths correlates to the accuracy of the approximation desired. For a three-formant approximation (formants are the tube resonance frequencies), the tube is divided into eight portions having successive lengths, L/10, L/15, 2L/15, 3L/15, 3L/15, 2L/15, L/15 and L/10, where L is the overall length of the tube.

Vowel-consonant-vowel modeling bysuperposition of consonant closure on vowel-to-vowel gestures

Abstract The aim of the paper is to describe the behavior of the Distinctive Region Model for the production of V 1 CV 3 utterances. The model, deduced from acoustic theory, is characterized by simplicity and efficiency in its handling of the articulatory-acoustic relation. A superposition principle that governs vowel-vowel transitions and consonant closures is tested. In the experiments, the duration of the closure is equal to the duration of the vowel-vowel transition and the movements between area targets are cosine transitions. The present work will report some attempts at modeling a well-known body of data on V 1 CV 3 utterances ( Ohman, 1966 ).

Perception of Vowel-to-Vowel Transitions with Different Formant Trajectories

In this paper, the perceptual effects of vowel-to-vowel transitions determined by different temporal variations of model parameters which specify the shapes of the vocal tract area function are investigated. It is shown that, (a) the method of deformation of the vocal tract area function between two targets can be perceptually important and (b) conversely, within certain limits, the time course of parameters from one state to another, and the precise synchronization of two parameters is not important for the correct identification of a vowel series. These characteristics are necessary but not sufficient to prove the existence of a phonetic gesture percept.

Relations entre la forme du conduit vocal et les caractéristiques acoustiques des voyelles françaises

The distribution inside the vocal tract of certain acoustic parameters (volume velocity, acoustic pressure, stored energy) is calculated for the French vowels. Calculations are made using a simulated

Modeling and Perception of ‘Gesture Reduction’

The phenomenon of vowel reduction is investigated by modeling ‘gesture reduction’ with the use of the Distinctive Region Model (DRM). First, a definition is proposed for the term gesture, i.e. an acoustically efficient command aimed at deforming, in the time domain, the area function of the vocal tract. Second, tests are reported on the perception of vowel-to-vowel transitions obtained with reduced gestures. These tests show that a dual representation of formant transitions is required to explain the reduction phenomenon: the trajectory in the F1–F2 plane and the time course of the formant changes. The results also suggest that time-domain integration of the trajectories constitutes an integral part of the auditory processing of transitions. Perceptual results are also discussed in terms of the acoustic traces of DRM gestures.

‘Speaker’ and ‘Speech’ Characteristics: A Deductive Approach

In this paper, a deductive approach to the study of vocal tract characteristics, vowel, vowel-to-vowel and vowel-consonant-vowel productions is presented. The distinctive region model, deduced from ac