Contrast enhancement in clearly spoken Mandarin sibilants

Abstract

Clearly spoken sibilants have been shown to exhibit temporal, spectral, and amplitudinal enhancement, though most existing work focuses on English. This study presents acoustics of clearly spoken sibilants in Mandarin, which exhibits a three-way place contrast between alveolar, retroflex, and alveopalatal sibilant fricatives. Sibilant-initial stimuli were elicited in a carrier phrase and were crossed according to: sibilant (alveolar, retroflex, alveopalatal), vowel context (a u), word frequency, number of syllables, and tone. The participants were eight native speakers of Mandarin with origins in mainland China. The first block was a baseline with no instruction on speech style. In the second block they were instructed to speak clearly 'as if talking to someone hard of hearing'. In the third block they were instructed to speak conversationally 'as if casually talking to a friend'. As expected, speakers enhanced duration and amplitude in the clear speech block and reduced duration and amplitude in the conversational block. Departing from previous work, spectral moments were not significantly enhanced in clear speech, which could be due to differences in type of task. However, second formants of vowels following alveopalatal sibilants were significantly higher in the clear speech block. This would be expected if clear speech is generally contrast enhancing, as the second formant of the following vowel is known to be the primary cue distinguishing the alveopalatal in perception and production.