Hardy Teresa L D, Boliek Carol A, Wells Kristopher, Dearden Carol, Zalmanowitz Connie, Rieger Jana M
Am J Speech Lang Pathol. 2016 May 1;25(2):125-37. doi: 10.1044/2015_AJSLP-14-0098.
The purpose of this study was to describe the pretreatment acoustic characteristics of individuals with male-to-female gender identity (IMtFGI) and investigate the ability of the acoustic measures to predict ratings of gender, femininity, and vocal naturalness.
This retrospective descriptive study included 2 groups of participants. Speakers were IMtFGI who had not previously received communication feminization treatment (N = 25). Listeners were members of the lay community (N = 30). Acoustic data were retrospectively obtained from pretreatment recordings, and pretreatment recordings also served as stimuli for 3 perceptual rating tasks (completed by listeners).
Acoustic data generally were within normal limits for male speakers. All but 2 speakers were perceived to be male, limiting information about the relationship between acoustic measures and gender perception. Fundamental frequency (reading) significantly predicted femininity ratings (p = .000). A total of 3 stepwise regression models indicated that minimum frequency (range task), second vowel formant (sustained vowel), and shimmer percentage (sustained vowel) together significantly predicted naturalness ratings (p = .005, p = .003, and p = .002, respectively).
Study aims were achieved with the exception of acoustic predictors of gender perception, which could be described for only 2 speakers. Future research should investigate measures of prosody, voice quality, and other aspects of communication as predictors of gender, femininity, and naturalness.
本研究旨在描述具有男性向女性性别认同的个体（IMtFGI）的治疗前声学特征，并研究声学测量预测性别、女性气质和声音自然度评分的能力。
这项回顾性描述性研究包括两组参与者。说话者为未曾接受过沟通女性化治疗的IMtFGI（N = 25）。听众为普通社区成员（N = 30）。声学数据是从治疗前的录音中回顾性获取的，治疗前的录音也用作3项感知评分任务的刺激材料（由听众完成）。
声学数据总体上在男性说话者的正常范围内。除2名说话者外，所有说话者都被感知为男性，这限制了有关声学测量与性别感知之间关系的信息。基频（朗读）显著预测了女性气质评分（p = .000）。总共3个逐步回归模型表明，最低频率（范围任务）、第二共振峰（持续元音）和闪烁百分比（持续元音）共同显著预测了自然度评分（分别为p = .005、p = .003和p = .002）。
除了仅能描述2名说话者的性别感知声学预测指标外，研究目标均已实现。未来的研究应调查韵律、语音质量和沟通的其他方面作为性别、女性气质和自然度预测指标的情况。
