International Conference on Auditory-Visual Speech Processing 2008

Tangalooma Wild Dolphin Resort, Moreton Island, Queensland, Australia
September 26-29, 2008

Auditory-Visual Perception of Prosodic Information: Inter-linguistic Analysis - Contrastive Focus in French and Japanese

Marion Dohen (1,2), Chun-Huei Wu (1), Harold Hill (2,3)

(1) Speech and Cognition Department, ICP, GIPSA-lab, France
(2) ATR, Cognitive Information Science Laboratories, Japan
(3) School of Psychology, University of Wollongong, Australia

Audition and vision are combined for the perception of speech segments and recent studies have shown that this is also the case for some types of supra-segmental information such as prosodic focus. The integration of vision and audition for the perception of speech segments however seems to be less important in Japanese. This study aims at comparing auditory-visual perception of prosodic contrastive focus in French and in Japanese. Two parallel focus identification tests were conducted for three modalities: AV, A and V in the two languages. Four speakers were recorded in both languages. For French, there was no AV advantage due to a ceiling effect. The same was observed for Japanese even though auditory only performances did not reach a ceiling. The results suggest that there are visual cues to prosodic focus in Japanese as well as in French but that these are not systematically combined with auditory information to enhance AV perception in Japanese. However, we also found that in some cases, especially when auditory alone perception is poor, visual and auditory information can be combined to enhance perception in Japanese.

Full Paper

Bibliographic reference.  Dohen, Marion / Wu, Chun-Huei / Hill, Harold (2008): "Auditory-visual perception of prosodic information: inter-linguistic analysis - contrastive focus in French and Japanese", In AVSP-2008, 89-94.