摘要
目的 :探讨嗓音客观多参数分析与主观听感知评估的相关性 ,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 :声音样本采自 83例嗓音障碍患者和 4 0例嗓音正常者 ,全部受试者均为女性。客观检测采用Dr.SpeechScienceforWindows嗓音评估软件 ,在长元音 /α :/上测试下列参数 :基频、基频微扰(jitter)、振幅微扰 (shimmer)、基频标准差、基频震颤、振幅震颤、标准化噪声能量 (NNE)、谐噪比 (HNR)、信噪比和最大发声时间 (MPT)。主观听感知评估参数采用日本言语音声学会声音嘶哑评估GRBAS系统中的总嘶哑度 ,4级评估标准。结果 :应用逐步判别分析方法 ,建立了 5个参数 (MPT、jitter、NNE、HNR和shimmer)的嗓音客观评估模型。客观评估结果与主观评估结果的一致性达到 79.8%。结论 :嗓音的客观评估是多参数的 ;嗓音的客观多参数评估模型与主观听感知结果有较好的一致性。
Objective:The purposes of this paper are to develop a multi-parametric protocol, which can evaluate the voice objectively and quantitatively,and to investigate its correlation with the perceptual analysis.Method:The voice samples were collected from 83 patients with dysphonia and 40 subjects with normal voices. All the subjects were women.The objective parameters, including fundamental frequency (F 0),jitter,Shimmer, fundamental frequency standard deviation (F 0SD), fundamental frequency tremor, amplitude tremor, normalized noise energy (NNE), harmonic-to-noise ratio(HNR), signal-to-noise ratio(SNR) and maximal phonatory time (MPT),which were measured mainly on a sustained vowel /a/, were recorded on a software named Dr. Speech for Windows. According to the G component of the GRBAS, the voice was graded from 0 for normal to 3 for severe dysphonia.Result:Using the discriminate analysis, a five parameters protocol (MPT, jitter, NNE, HNR and shimmer) was developed and it showed that this protocol allowed 79.8% concordance with jury classification.Conclusion:The voice objective analysis should be multi-parameters.Our multi-parameters evaluation protocol is able to reflect the result of perceptual analysis.
出处
《临床耳鼻咽喉科杂志》
CSCD
北大核心
2004年第12期751-754,共4页
Journal of Clinical Otorhinolaryngology
关键词
嗓音障碍
言语声学
Voice disorders
Speech acoustics
