Acoustic analysis of Mandarin Chinese vowels produced by young adults

Abstract

Objectives: Acoustic analysis is a kind of objective assessment of speech-sound which can offer a relatively simple and visual way to examine production of a vowel. Mandarin Chinese is a tone language in which the same phonetic segment carries a different meaning when produced at different lexical tones. Previous studies have examined American English vowels produced by native adult speakers. The aim of the present study was to establish the vowel formant space by examining the formant frequencies associated with Mandarin vowels produced by Chinese young adults aged 23 to 33 years old and identify the characteristics of vowels in Mandarin Chinese at 4 different lexical tones. Methods: Acoustic signals of the six Mandarin vowels (/a/, /i/, /e/, /i/, /u/, /y/) produced by native young adult speakers of Mandarin Chinese (n=11) were recorded, and the participants were instructed to produce all the speech samples using a comfortable loudness level and speech rate. The speech samples were produced was randomly. The first two formants were analyzed by using a professional acoustic measurement system (Multi-Speech, KayPentax, USA). Multi-Speech provides a time-domain waveform and a frequency-domain wide-band spectrogram (filter bandwidth=300 Hz). Results: The results showed that, generally speaking, the vowel /a/ exhibited the highest mean F1 value whereas vowel /i/ had the lowest F1. The vowel /i/ showed the highest F2 while /u/ showed the lowest F2 value. The vowel of /蘅/ and /e/ with different four tones in F1 and F2 had significant differences (P<0.05). However, no significant differences (P>0.05) in F1 and F2 were observed in the other vowels across tones, like /a/, /i/, /u/, /y/. Conclusion: An acoustic vowel space was established in which /i/, /u/, and /a/ were corner vowels, and /蘅/ and /e/ were found centrally, contributing to the generally triangular vowel space associated with these six core vowels of Mandarin Chinese. The investigation of the first two formant frequencies of vowels across tones shows significant differences were found in the vowels /蘅/ and /e/ with different tones in F1 and F2. When /蘅/ was produced with level tone, the tongue was more advanced and depressed than with other three tones. For /e/ with rising tone the tongue was more retracted, compared to the tongue for /e/ with dipping tone.