Hardy Teresa L D, Rieger Jana M, Wells Kristopher, Boliek Carol A
Faculty of Rehabilitation Medicine, University of Alberta, Edmonton, Alberta, Canada; Glenrose Rehabilitation Hospital, Alberta Health Services, Edmonton, Alberta, Canada.
Faculty of Rehabilitation Medicine, University of Alberta, Edmonton, Alberta, Canada.
J Voice. 2020 Mar;34(2):300.e11-300.e26. doi: 10.1016/j.jvoice.2018.10.002. Epub 2018 Nov 28.
This study aimed to identify the most salient set of acoustic predictors of (1) gender attribution; (2) perceived masculinity-femininity; and (3) perceived vocal naturalness amongst a group of transgender and cisgender speakers to inform voice and communication feminization training programs. This study used a unique set of acoustic variables and included a third, androgynous, choice for gender attribution ratings.
Data were collected across two phases and involved two separate groups of participants: communicators and raters. In the first phase, audio recordings were captured of communicators (n = 40) during cartoon retell, sustained vowel, and carrier phrase tasks. Acoustic measures were obtained from these recordings. In the second phase, raters (n = 20) provided ratings of gender attribution, perceived masculinity-femininity, and vocal naturalness based on a sample of the cartoon description recording.
Results of a multinomial logistic regression analysis identified mean fundamental frequency (f) as the sole acoustic measure that changed the odds of being attributed as a woman or ambiguous in gender rather than as a man. Multiple linear regression analyses identified mean f, average formant frequency of /i/, and mean sound pressure level as predictors of masculinity-femininity ratings and mean f, average formant frequency, and rate of speech as predictors of vocal naturalness ratings.
The results of this study support the continued targeting of f and vocal tract resonance in voice and communication feminization/masculinization training programs and provide preliminary evidence for more emphasis being placed on vocal intensity and rate of speech. Modification of these voice parameters may help clients to achieve a natural-sounding voice that satisfactorily represents their affirmed gender.
本研究旨在确定一组最显著的声学预测指标，用于（1）性别归因；（2）感知到的男性气质-女性气质；以及（3）在一组跨性别者和顺性别者中感知到的声音自然度，以为语音和沟通女性化培训项目提供参考。本研究使用了一组独特的声学变量，并在性别归因评分中增加了第三种中性选择。
数据分两个阶段收集，涉及两组不同的参与者：沟通者和评分者。在第一阶段，在卡通复述、持续元音和载体短语任务期间录制了沟通者（n = 40）的音频。从这些录音中获取声学测量数据。在第二阶段，评分者（n = 20）根据卡通描述录音样本对性别归因、感知到的男性气质-女性气质和声音自然度进行评分。
多项逻辑回归分析结果确定平均基频（f）是唯一能改变被归为女性或性别模糊而非男性的几率的声学测量指标。多元线性回归分析确定平均f、/i/的平均共振峰频率和平均声压级作为男性气质-女性气质评分的预测指标，平均f、平均共振峰频率和语速作为声音自然度评分的预测指标。
本研究结果支持在语音和沟通女性化/男性化培训项目中继续以f和声道共振为目标，并为更多地强调声音强度和语速提供了初步证据。修改这些语音参数可能有助于客户获得一种自然的声音，令人满意地体现他们所认同的性别。
