ISCA Archive SpeechProsody 2024
ISCA Archive SpeechProsody 2024

Perceiving the social meanings of creaky voice in Mandarin Chinese

Yao Yao, Meixian Li, Shiyue Li, Charles B. Chang

While there is a growing literature on the social meanings of nonmodal voice qualities, most of the existing studies focus on English and use either naturally produced speech stimuli (which are hard to control acoustically) or a small set of fully synthesized stimuli. This paper reports a perceptual study of the social meanings of creaky voice in Mandarin Chinese, using a large set of resynthesized stimuli featuring 38 talkers (19F) and 6–10 pairs of sentences per talker that differed only in voice quality (creaky vs. modal). Sixty listeners (33F) answered 4 questions about the talker’s demographic profile (age, gender, sexuality, education) and gave 19 ratings of personality traits (e.g., confident, professional, charismatic) and interactive potential (e.g., engagingness). Using factor analysis and mixed-effects modeling, our results showed that for male listeners, creaky voice significantly decreased the perceived warmth of male talkers but increased the perceived warmth of female talkers; creaky voice also led to more gender identification errors on female talkers by female listeners and made male talkers sound older. These findings point toward multifaceted social meanings of creaky voice in Mandarin, which extend beyond talker attractiveness and are closely linked to gender, both the talker’s and the listener’s.