专业歌手计算机嗓音分析测试元音的选择研究  

The study of vowels with computer voice analysis in trained singer

作　　者：葛平江[1] 李艳[2] 彭莉佳[2] 李静远[1] 卢川[1] 张思毅[1] 陈少华[1] 

机构地区：[1]广东省人民医院(广东省医学科学院)耳鼻咽喉科,广州510080 [2]星海音乐学院

出　　处：《临床耳鼻咽喉头颈外科杂志》2014年第1期32-34,共3页Journal of Clinical Otorhinolaryngology Head And Neck Surgery

基　　金：广州市科技支撑项目(No:11A56040659)资助

摘　　要：目的:了解专业歌手同普通人嗓音计算机分析中持续/a/音和/i/音的差别。方法:30例经过声乐专业训练的专业歌手(专业组),男15例,年龄(22.5±1.6)岁;女15例,年龄(22.7±2.5)岁。正常普通人30例(普通组),男16例,年龄(30.9±4.7)岁;女14例,年龄(29.0±3.8)岁。在隔声室中使用德国Xion公司的头戴式麦克风录音,研究对象发持续/a/音和/i/音各3遍,选取中间平稳段3s作为分析段。使用Praat嗓音分析软件进行分析。取基频均数(F0)、频率微扰Jitter、振幅微扰Shimmer、噪谐比、谐噪比强度作为测试值,使用独立样本T检验统计专业组和普通组以及持续/a/音和/i/音之间的差别。Pearson相关检测持续/a/音和/i/音之间的相关性。结果:持续/a/音和/i/音的Shimmer值,普通组大于专业组(P<0.01)。专业组男性发持续/a/音和/i/音的基频均高于普通组(均P<0.05)。专业组F0值、Shimmer值及噪谐比/i/音明显大于/a/音(均P<0.01),Jitter值及谐噪比强度/i/音小于/a/音(均P<0.05)。普通组Jitter值2个元音无差别,其余指标同专业组。专业组的/i/-/a/音谐噪比强度差值较普通组大(P<0.05)。专业组持续/a/音和/i/音的F0、Jitter值、噪谐比和谐噪比强度显著相关(均P<0.05)。普通组持续/a/音和/i/音之间F0、谐噪比强度显著相关(均P<0.01)。结论:专业组男性发/a/音和/i/音的基频均高于普通组,前后元音的杂音均少于普通组。专业组选择/a/音或/i/音均可作为测试音,两者区别不大。Objective：To investigate the differences between the trained singer and normal people by computer voice analysis with sustained vowel/i/ and /a/. Method： Fifteen male（22. 5±1. 6 years） and 15 female（22.7±2.5 years） trained singers and 16 male（30.9 ± 4.7 years） and 14 female（29.0±3.8 years） normal controls were involved in this study. Sustained vowel/a/and/i/audio recordings were made with the recording head-mount mi- crophone（Xion, Germany） in a sound-attenuated room. The middle three second voice of samples was analyzed. The subject repeated vowels for 3 times. The Praat voice analysis software were used to analysis the F0, Jitter, Shimmer, NHR（Noise-harmonic ratio）, and HNR（Harmonic-noise ratio） of vowels. The independent sample t- test was applied to get the differences of sustained vowels/a/ and/i/between the trained singers and normal con- trois The correlation also was analyzed between the sustained vowel /a/ and/i/. Result： The Shimmer of trained singers were greater than normal controls in vowel/a/and/i/. The F0 of vowel/a/ and/i/of trained male singers were more than those of normal controls. The F0 （P〈0.01）, Shimmer（P〈0.01）, and NHR（P〈0.01） of vowel /i/ were greater than vowel/a/ and the Jitter and HNR（Harmonic-noise ratio）of vowel/i/ were less than vowel /a/ in trained singers. There was no difference between the vowel/a/ and /i/ in Jitter in normal controls. The F0, Shimmer, NHR, HNR were same as trained singers in difference between the vowel/a/and/i/. The trained singers/i/-/a/ difference-value of HNR was greater than normal controls＇（P〈0.05）. There were significant cor- relation between the vowel/a/and/i/ in F0 （P〈0.01）, Jitter（P〈0.01）, NHR（P〈0.05）, HNR（P〈0.01）of trained singers. The F0 （P〈0.01）, HNR（P〈0.01） of normal controls had significant correlations between the vowel/a/and/i/. Conclusion ： The training may increase the F0 of male singer. Singer training can reduce the voice perturbation o

关 键 词：专业歌手 嗓音分析 元音发声 

分 类 号：R767.64[医药卫生—耳鼻咽喉科]