ASA 127th Meeting M.I.T. 1994 June 6-10

2pSP11. Exploring the Japanese McGurk effect using various S/N ratios and speakers.

Kaoru Sekiyama

MIT Res. Lab. of Electron., Rm. 36-761, Cambridge, MA 02139

Keiko Nishino

Kanazawa Univ., Kanazawa, 920-11 Japan

The McGurk effect demonstrates that perceivers integrate auditory and lip-read speech even when two sources of information conflict. Whereas this effect is robust in English, it was found [K. Sekiyama and Y. Tohkura, J. Phon. 21, 427--444 (1993)] that it is much weaker among Japanese perceivers. This was further examined. In experiment 1, Japanese subjects were tested under various S/N conditions produced by combining four speech (65, 58, 51, and 44 dB) and four white noise (56, 48, 40, and 32 dB) levels. The audio-visual speech stimuli were those used in the earlier study and were created from a Japanese speaker's utterances of ten monosyllables (/ba,ma,pa,wa,da,na,ta,ga,ra,ka/). Subjects were asked to report what they heard. When the noise level was lower than the speech level, the McGurk effect was not as strong as the one reported for English. In experiment 2, Japanese subjects were tested with a new stimulus set created from four Japanese speaker's utterances of eight monosyllables. The stimuli (58 dB) were presented without adding noise. The weak McGurk effect was replicated. From these results, it is generalized that the Japanese tend to be less subject to the McGurk effect.