The purpose of this study is to identify the accuracy with which graduate students in a department of communication sciences and disorders identify modal register, vocal fry, and uptalk presented in audio samples of female celebrity speakers, and to report these listeners' perceptual responses to a variety of attributes (eg, trustworthy, competent, educated). This investigation was an anonymous online survey study. As part of an anonymous online survey, graduate students in a department of communicative sciences and disorders listened to training modules and then classified female voice samples according to the three features under investigation (ie, modal register, vocal fry, and uptalk). The listeners then appraised a variety of speaker attributes, including physical attractiveness, trustworthiness, competence, and level of education, based on the audio samples of connected speech. The participants labeled voice samples of vocal fry with 85% accuracy, uptalk at 79% accuracy, and modal register only 51% of the time. Cohen's Kappa showed substantial agreement between the repeated ratings of modal register and a moderate agreement for repeated ratings of samples of vocal fry and uptalk. More raters identified speakers using the modal register positively across the attributes and occupations under investigation than voices produced in vocal fry and uptalk. Generalized estimating equation models (GEE) revealed statistically significant differences between modal register and vocal fry and modal register and uptalk for all perceptual attributes and occupations. However, there were no statistically significant differences found between vocal fry and uptalk for the attributes and occupations except for the occupation of actress, where the difference between vocal fry and uptalk was statistically significant. Training notwithstanding, the listeners had difficulty accurately identifying modal register despite their frequent exposure to that manner of phonation. They did, however, with a high level of accuracy, identify vocal fry and uptalk and were similar in their judgment of attributes to those vocal features. The listeners' responses revealed similar negative assumptions associated with vocal fry and uptalk and positive assessments of modal register. Our results raise questions rather than conclusions. If, as has been shown in previous studies, female speakers often speak in vocal fry and uptalk, why are their responses to those vocal features so negative? Are the cultural, sociolinguistic, and peer pressures so powerful that they override the individual’s judgment?

中文翻译：

---

模态音域、声乐 Fry 和 Uptalk：缺乏经验的听众的识别和感知判断


本研究的目的是确定传播科学与疾病系的研究生识别女性名人演讲者音频样本中的模态音域、发声和上行语的准确性，并报告这些听众对各种属性（例如，值得信赖、有能力、受过教育）。这项调查是一项匿名在线调查研究。作为匿名在线调查的一部分，交流科学与疾病系的研究生听了培训模块，然后根据所调查的三个特征（即模态语域、声带和上行语）对女性声音样本进行分类。然后，听众根据相连语音的音频样本评估说话者的各种属性，包括外表吸引力、可信度、能力和教育水平。参与者标记发声鱼苗语音样本的准确度为 85%，上行语音的准确度为 79%，而模态注册的准确度仅为 51%。 Cohen 的 Kappa 表明，模态音域的重复评级与声带和上行声样本的重复评级之间存在适度的一致性。更多的评估者在所调查的属性和职业中积极地使用模态语域识别说话者，而不是通过发声和上行说话产生的声音。广义估计方程模型（GEE）揭示了所有感知属性和职业的模态音域和发声状态以及模态音域和上行音之间存在统计学上的显着差异。 然而，除女演员职业外，vocal fry 和 uptalk 之间的属性和职业没有统计学上的显着差异，其中女演员的职业之间的差异具有统计学意义。尽管接受过培训，听众仍然难以准确识别情态语域，尽管他们经常接触这种发声方式。然而，他们确实以很高的准确度识别了发声和上行声，并且对这些发声特征的属性的判断是相似的。听众的反应显示出与发声和上调相关的类似负面假设以及对模态音域的正面评估。我们的结果提出了问题而不是结论。正如之前的研究所示，如果女性说话者经常用低声说话和上声说话，为什么她们对这些声音特征的反应如此消极？文化、社会语言和同伴压力是否如此强大以至于超越了个人的判断？