摘要
目的 探讨嗓音客观多参数分析与主观听感知评估的一致性，建立嗓音客观多参数评估模型，实现嗓音评估的客观化和数据化。方法受试者为嗓音障碍患者271例，其中女性124例，男性147例；嗓音正常对照组69例，女性37例，男性32例。主观听感知评估采用GRBAS系统中的总嘶哑度（grade，G），采用4级改良分级量表。听评委为来自不同医院的5名嗓音医学家。语音材料为统一的语句。全部受试者嗓音按随机方式3次排序，分别进行3次评估。嗓音客观测试采用Dr．speech for windows嗓音评估软件，受试者发长元音[α：]，取含起始段2S的嗓音样本，测试基频等7个参数。结果单参数分析结果显示，除了基频不能反映出嗓音障碍程度外，其他6个参数的测试值随嗓音障碍程度的加重而发生相应的变化，并且在两相邻嗓音组之间的差异性有统计学意义（P值均〈0．05）。借助于判别分析方法，按性别分别建立了由基频微扰、振幅微扰、基频标准差、标准化噪声能量、谐噪比和最长发声时间组成的客观多参数评估模型。客观评估与主观评估的一致性在男性达到81．6％，女性达到83．2％。正常嗓音组和重度嗓音障碍组的一致性高于轻度嗓音障碍组和中度嗓音障碍组。全部误判嗓音均被判到相邻的嗓音组。结论嗓音的客观声学参数能够反映出嗓音障碍的主观听感知特性；客观多参数模型评估结果与主观听感知结果达到较好的一致性，为嗓音障碍的评估提供了一种客观的方法。
Objective Through exploring the concordance of objective multi-parameters analysis and perceptual evaluation, to establish an objective multi-parameters evaluation protocol of voice disorder and to make the evaluation of voice objectification and quantification. Methods Voice samples from 271 patients （ 124 female and 147 male）with dysphonia and 69 control subjects with normal voice （37 female and 32 male）were recorded and assessed by a jury composed of 5 experts in phoniatrics from different hospitals. The jury was instructed to classify voice samples according to the G （grade） component of the GRBAS scale on a visual analogue scale secondarily transformed in a 4-point scale ranging from 0 for normal to 3 for severe dysphonia. The voice samples were unified sentences and ordered randomly 3 times, the mean of 3 evaluation scores were the final results. The objective parameters, including fundamental frequency （ F0）, jitter, shimmer, fundamental frequency standard deviation （FOSD）, normalized noise energy （NNE）, harmonic-to-noise ratio（HNR） and maximal phonatory time （ MPT）, were measured on a 2-second sustained vowel/a/ including its initial segment, using the software Dr. Speech for Windows. The data were analyzed using SPSSll. 0. Results All objective parameters except for F0 had high correlation with G and the variance tendency of these parameters values was coherent with the extent of voice disorder. And there were statistical differences between adjacent voice disorder groups. Male and female objective multi-parameters protocols were established respectively consisting of jitter, shimmer, FOSD, NNE, HNR and MPT usingdiscriminant analysis （P 〈 0. 05 ）. The concordance between perceptual evaluation and objective multi- parameters evaluation was 81.6% in male and 83.2% in female. The concordance of evaluation of normal voice and severe voice disorder groups were better than that of mild and moderate voice disorder groups. All mis-grading voices were judged in the adjacent voice group. Conclusions The objective parameters of voice are able to reflect the characteristic of its perceptual evaluation and the concordance between perceptual evaluation and objective multi-parameters evaluation is good. The objective multi-parameters evaluation protocol we established could provide an objective and quantitative evaluation method for voice disorders.
出处
《中华耳鼻咽喉头颈外科杂志》
CAS
CSCD
北大核心
2012年第10期817-822,共6页
Chinese Journal of Otorhinolaryngology Head and Neck Surgery
基金
国家自然科学基金（81070795）
关键词
语音障碍
语音质量
言语参数测量
判别分析
Voice disorders
Voice quality
Speech production measurement
Discriminant analysis
