The relationship between vowel identification in noise and vowel intelligibility for Chinese and Korean speakers.

Abstract

The present study aims to investigate the effects of background noise on English vowel perception for non-native listeners and to examine the relationship between vowel perception and vowel intelligibility for non-native people. For experiment 1, vowel identification was measured in two noise types: multi-talker babble and long-term speech shaped noise at various signal-to-noise ratios (SNRs) for English-native (EN), Chinese-native (CN), and Korean-native (KN) listeners. For experiment 2, English vowels spoken by CN and KN speakers who served as listeners in experiment 1 were recorded. Preliminary data of vowel perception showed that at low SNRs (−18 to −15 dB SNR), all three groups of listeners performed poorly, and at relatively middle and high SNRs, vowel identification for both CN and KN listeners was significantly poorer than EN listeners. Even at +3 dB SNR, the average identification scores for non-native listeners were more than 20% less than those for EN listeners. Furthermore, there was significant individual variability in vowel identification within each non-native group. This variability in vowel perception might be related with variability in vowel intelligibility for the non-native individuals. Vowel intelligibility for the non-native speakers will be evaluated by EN listeners. The implication on bilingual speech perception and production will be discussed.