Acoustic Analysis of Mandarin-Speaking Transgender Women.

Abstract

Objectives: This study aims to investigate the speech characteristics and assess the potential risk of voice fatigue and voice disorders in Chinese transgender women (TW).

Methods: A case-control study was conducted involving TW recruited in Shanghai, China. The participants included 15 TW, 20 cisgender men (CISM), and 20 cisgender women (CISW). Acoustic parameters including formants (F₁, F₂, F₃, F₄), cepstral peak prominence (CPP), jitter, shimmer, harmonic-to-noise ratio (HNR), noise-to-harmonics (NHR), fundamental frequency (f₀), and intensity, across vowels, passages, and free talking. Additionally, the Voice Handicap Index-10 (VHI-10) and the Voice Fatigue Index were administered to evaluate voice-related concerns.

Results: (1) The F₁ of TW was significantly higher than that of CISW for the vowels /i/ and /u/, and significantly higher than that of CISM for the vowels /a/, /i/, and /u/. The F₂ of TW was significantly lower than CISW for the vowels /i/, significantly higher than CISW for the vowels /u/, and significantly higher than CISM for the vowels /a/ and /u/. F₃ was significantly lower in TW than in CISW for the vowels /a/ and /i/. The F4 formant was significantly lower in TW than in CISW for the vowels /a/ and /i/, but significantly higher than in CISM for the vowel /u/. (2) The f₀ of TW was significantly lower than that of CISW for the vowels /a/, /i/, /u/, during passage reading, and in free speech, but was significantly higher than CISM during passage reading and free talking. Additionally, TW exhibited significantly higher intensity compared with CISW for the vowel /a/ and during passage reading. (3) Jitter in TW was significantly higher than in CISW for the vowels /i/ and /u/, and significantly lower than in CISM during passage reading and free talking. Shimmer was significantly higher in TW compared with both CISW and CISM across the vowels /a/, /i/, during passage reading, and in free talking. The HNR in TW was significantly lower than in both CISW and CISM across all vowels, during passage reading, and in free talking. The NHR was significantly higher in TW than in CISW across all vowels, during passage reading, and in free talking, and significantly higher than in CISM for the vowels /a/, /i/, during passage reading, and in free talking. The CPP in TW was significantly lower than in CISW during passage reading and free talking, and significantly lower than in CISM across all vowels, during passage reading, and in free speech. (4) The VHI-10 scores were significantly higher in TW compared with both CISM and CISW.

Conclusions: TW exhibit certain acoustic parameters, such as f₀ and some of the formants, that fall between those of CISW and CISM without undergoing phonosurgery or voice training. The findings suggest a potential risk for voice fatigue and the development of voice disorders as TW try to modify their vocal characteristics to align with their gender identity.