For native perceivers, seeing a native talker's face can improve the intelligibility of acoustic speech presented in noise. For this perceiver population, it is also known that the intelligibility of foreign accented English, spoken by native talkers of Mandarin Chinese, is more susceptible to the effects of noise compared to the speech of native talkers (Rogers et al, 2004). The current experiment was designed to investigate the influence of seeing a non-native talker’s face on the intelligibility of foreign accented speech presented in noise. Ten talkers (nine non-native and one native) were audio and video recorded producing 155 sentences each. Talker specific speech shaped noise was mixed with the audio of the sentences. The signal-to-noise ratio was set for each talker based on preliminary testing. Eighty-one native perceivers of English (nine per non-native talker) responded to 20 sentences spoken by the native talker and 28 by a non-native talker in each presentation condition (audio-alone, visual-alone, audiovisual). Open set identification responses were scored as percent words correct for each condition and talker. Preliminary analyses demonstrate that seeing the non-native talker increases intelligibility of speech presented in noise. However, the magnitude of the audiovisual gain is reduced in comparison to the gain observed with the native talker. Results will be discussed in terms of potential sources of audiovisual gain. These sources are hypothesized to be differentially sensitive to the influences of non-native speech. [Work supported in part by NIH/NIDCD DC04856 and the University of Kansas, UGRA].
Reference:
Rogers, C. L., Dalby, J., & Nishi, K. (2004). Effects of noise and proficiency on intelligibility of Chinese-accented English. Language and Speech, 47(Pt 2), 139-154.