Faculty of Medicine and Health Sciences, University of Antwerp, Antwerp, Belgium.
Department of Phoniatrics and Pedaudiology, Clinics of Otolaryngology, Head and Neck Surgery, Ernst-Moritz-Arndt-University Greifswald, Greifswald, Germany.
J Voice. 2020 Jan;34(1):157.e17-157.e25. doi: 10.1016/j.jvoice.2018.07.026. Epub 2018 Sep 12.
The aim of this study was to validate the acoustic voice quality index (AVQI) version 03.01 and the acoustic breathiness index (ABI) in the German language.
Concatenated voice samples of continuous speech (cs) and sustained vowel (sv) [a:] from 175 subjects with dysphonia and 43 vocally healthy subjects were perceptually judged for overall voice quality and breathiness severity. First, to reach a higher level of ecological validity, the proportions of cs and sv were equalized regarding the time length. Second, concurrent validity and diagnostic accuracy were statistically verified using the Spearman rank-order correlation coefficient (r) and several estimates of the receiver operating characteristics, likelihood ratio, and Youden index.
A moderate reliability of overall voice quality and breathiness severity from three experts were found. It was found that 27 syllables as standardization of the cs part, which represents 3 seconds of voiced cs, allows the equalization of both speech tasks. A strong correlation was revealed between AVQI and overall voice quality (r = 0.86, P < 0.01), and ABI and perceived breathiness severity (r = 0.85, P < 0.01). Additionally, the best diagnostic outcome for AVQI and ABI was identified at a threshold of 1.85 (sensitivity of 72 % and specificity of 90 %) and 3.42 (sensitivity of 72 % and specificity of 95 %), respectively.
AVQI and ABI showed in the German language valid and robust results to quantify abnormal voice qualities regarding overall voice quality and breathiness severity.
本研究旨在验证声学语音质量指数（AVQI）版本 03.01 和声学粗糙音指数（ABI）在德语中的有效性。
将 175 名嗓音障碍患者和 43 名嗓音正常者的连续语音（cs）和持续元音（sv）[a:]的拼接语音样本进行感知判断，评估整体语音质量和粗糙音严重程度。首先，为了提高生态有效性，将 cs 和 sv 的比例根据时长进行均衡。其次，使用 Spearman 等级相关系数（r）和几种接收者操作特征、似然比和 Youden 指数的估计值，对并发有效性和诊断准确性进行统计学验证。
发现三位专家对整体语音质量和粗糙音严重程度的可靠性中等。发现 27 个音节作为 cs 部分的标准化，代表 3 秒的浊音 cs，可实现两种语音任务的均衡。AVQI 与整体语音质量（r=0.86，P<0.01）和 ABI 与感知粗糙音严重程度（r=0.85，P<0.01）之间存在较强相关性。此外，AVQI 和 ABI 的最佳诊断结果分别为阈值 1.85（敏感性为 72%，特异性为 90%）和 3.42（敏感性为 72%，特异性为 95%）。
AVQI 和 ABI 在德语中显示出有效的、稳健的结果，可定量评估整体语音质量和粗糙音严重程度的异常嗓音质量。
