嗓音声学参数与嗓音音质的相关研究  

Relationship between voice acoustic parameters and vocal quality

作　　者：黄昭鸣[1] 万萍[1] 

机构地区：[1]华东师范大学学前教育与特殊教育学院言语听觉康复科学研究院,上海200062

出　　处：《临床耳鼻咽喉头颈外科杂志》2008年第6期251-255,共5页Journal of Clinical Otorhinolaryngology Head And Neck Surgery

基　　金：教育部2004年哲学社会科学研究重大课题攻关项目<人工耳蜗术后汉语言康复教育机理和方法研究>资助(No:04JZD00033)

摘　　要：目的:通过病理嗓音合成器控制声学参数的变化,考察听觉感知判断的变化,寻找能够反映嗓音音质的有效声学参数。方法:采用6个声学参数(嗓音基频、基频微扰、振幅微扰、声门噪声能量、声门谱斜率、共振峰频率微扰),合成36个声学样本/e/,并在每个参数维度上随机复制25%,总共45个样本,秩序随机化之后,再将样本编号;在室内将声学样本通过计算机播出,声音强度控制在70dB,由8名耳鼻咽喉科专家参与听觉感知评估,针对GRBAS中的G、R、B进行等级评分,评分结果显示听感判断存在较高的内在一致性(P<0.01)。然后分别将声学参数的变化值与听觉感知得分进行等级相关分析。结果:6个声学参数中基频微扰、振幅微扰以及声门噪声能量与感知判断指标(嘶哑声、粗糙声、气息声)显著相关(P<0.01)。结论:基频微扰、振幅微扰以及标准化声门噪声能量是衡量嗓音音质的有效参数,并证明嗓音的多维性质。嗓音的听觉感知判断与音质声学分析相结合,是评估嗓音音质的最好方式。Objective： This research is aimed at looking for the effect acoustic parameters of voice quality through the Spearman correlation analysis between auditory-perceptual judgment of voice quality and synthesized acoustic parameters. Method：Thirty-six vowel samples of /ae/ synthesized with six acoustic parameters （F0, Jitter, Shimmer, NNE, Spectral tilt, Formant Flutter）, each parameter dimension included 25% coped ones, altogether 45 samples. Then give the serial number to each sample after their order is randomized. The sound samples were given by computer, intensity was fixed at 70dB. Eight famous ENT doctors participated in the task of auditory-perceptual judgment. Give the grades to G, R, B, resulting that there are high inherent consistence （P〈 0.01）. Then do the Spearman correlation analysis with each acoustic parameter respectively. Result：There are significant correlation between the auditory-perception （G, R, B） and Jitter.Shimmer and NNE （P〈0.01）. Conclusion：Jitter, Shimmer and NNE are effect acoustic parameters reflecting the voice quality. Human voice is multidimensional. The combination the auditory-perceptual （G, R,B） of assessing the voice quality.

关 键 词：语音声学参数 语音质量 

分 类 号：R767.92[医药卫生—耳鼻咽喉科]