Numéro
J. Phys. IV France
Volume 02, Numéro C1, Avril 1992
Deuxième Congrès Français d'Acoustique / Second French Conference on Acoustics
Page(s) C1-319 - C1-322
DOI http://dx.doi.org/10.1051/jp4:1992168
Deuxième Congrès Français d'Acoustique / Second French Conference on Acoustics

J. Phys. IV France 02 (1992) C1-319-C1-322

DOI: 10.1051/jp4:1992168

IDENTIFICATION VISUELLE DES GESTES DE PROTRUSION ET DE RÉTRACTION DES LÈVRES AU COURS DES PAUSES ACOUSTIQUES : Les performances de sujets français et grecs

M.A. CATHIARD1, A. CIROT-TSEVA2 and M.T. LALLOUACHE2

1  Laboratoire de Psychologie Expérimentale, URA CNRS N° 665, Université Pierre Mendès-France, BP 47X, F-38040 Grenoble cedex, France
2  Institut de la Communication Parlée, URA CNRS N° 368, INPG-Université Stendhal, BP. 25X, F-38040 Grenoble cedex, France


Abstract
Since coarticulation is one of the basic phenomena in speech production, the issue that speech perception could take advantage of it, is of course a very fundamental one. The main questions we address in this paper are, more specifically : (i) Could speech perception benefit from anticipatory - and perseverative - coarticulation when it is clearly visible, but inaudible, as in the case of vowel rounding (protrusion/retraction) along acoustic silent pauses, which occur naturally in discourse ? (ii) Could speakers of a language like Greek, who do not have linguistically significant rounding contrasts, as in French, take the same visual advantage of coarticulation? Our results show that : (i) Rounding anticipation can be identified reliably (95% correct), only by eye, up to 120 milliseconds before any perceivable sound ; but this identification perseverate no more than 40-60 ms after sound utterance. (ii) The comparison between French and Greek subjects did not reveal significant differences in their identification behaviour, neither for anticipatory, nor perseverative coarticulation. Wether these results argue for a universal lipreading skill, remains of course an open quest.



© EDP Sciences 1992