Phonemic Perception Training in Quiet and Noise of English Speech for Chinese College Students

date April 24, 2015
comments no comments.

Friday, April 24, 2015 4:00 p.m. in ETC 4.150

Professor Chang Liu
Department of Communication Science and Disorders
The University of Texas at Austin

The goal of this study is to examine two hypotheses of our previous studies on English speech perception of Chinese college students, who had similar performance in English vowel identification in quiet and stationary noise, but lower scores in multi-talker babble, compared to their peers in the US.  The 1-2 year experience of living in the US may promote Chinese-native speakers’ capacity to use the temporal cues in noise and get less impact by informational masking of babble.  Several experiments of English vowel perception were conducted before and after phonemic perception training. Chinese college students were divided into three training groups: control (no specific phonemic training), training in quiet (TIQ), and training in noise (TIN). Five English back vowels were used as the training stimuli and training was conducted daily for one hour for six consecutive days. Preliminary results showed that English vowel perception in quiet was improved after training for all three groups with the training effect highest for TIN, lowest effect for Control, and TIQ in between. Vowel perception in noise was improved the most for the TIN, followed by TIQ, while no improvement was found for Control. In particular, perceptual weight was shifted after training. That is, the three groups used vowel duration cues similarly before training, whereas TIN and TIQ did not use duration cues, but Control still used it after training, implying that phonemic training improved listeners’ use of phonetic cues. Training effect was also extended to new talkers for TIN and TIQ, especially for TIN in noise.  In addition, training effect was observed for five front vowels that are not included in the training vowels. Post-training results indicate that TIN had better capacity against informational masking of babble than TIQ, but did not differ in the capacity of using temporal cues of noise. Altogether, English phonemic training may improve non-native listeners’ phonetic processing and the ability to resist informational masking.