Yokonishi Hisayuki, Imagawa Hiroshi, Sakakibara Ken-Ichi, Yamauchi Akihito, Nito Takaharu, Yamasoba Tatsuya, Tayama Niro
The Department of Otolaryngology, The University of Tokyo Hospital, Bunkyo-Ku, Tokyo 113-8655, Japan.
The Department of Otolaryngology, The University of Tokyo Hospital, Bunkyo-Ku, Tokyo 113-8655, Japan.
J Voice. 2016 Mar;30(2):145-57. doi: 10.1016/j.jvoice.2015.01.009. Epub 2015 May 4.
In the present study, we examined the relationship between various open quotients (Oqs) and phonation types, fundamental frequency (F0), and intensity by multivariate linear regression analysis (MVA) to determine which Oq best reflects vocal fold vibratory characteristics.
Using high-speed digital imaging (HSDI), a sustained vowel /e/ at different phonation types, F0s, and intensities was recorded from six vocally healthy male volunteers: the types of phonation included modal, falsetto, modal breathy, and modal pressed phonations; and each phonation was performed at different F0s and intensities. Electroglottography (EGG) and sound signals were simultaneously recorded with HSDI. From the obtained data, 10 conventional Oqs (four Oqs from the glottal area function, four kymographic Oqs, and two EGG-derived Oqs) and two newly introduced Oqs (Oq(edge)+ and Oq(edge)) were evaluated. And, relationships between various Oqs and phonation types, F0, and intensity were evaluated by MVA.
Among the various Oqs, Oq(edge)+ and Oq(edge) revealed the strongest correlations with an acoustic property and could best describe changes in phonation types: Oq(edge) was found to be better than Oq(edge)¯. Oq(MLK), the average of five Oqs from five-line multiline kymography was a very good alternative to Oq(edge)¯. EGG-derived Oqs were able to differentiate between modal phonation and falsetto phonation, but it was necessary to consider the change of F0 simultaneously. MVA showed the changes in Oq values between modal and other phonation types, the degree of involvement of intensity, and no relationship between F0 and Oqs.
Among Oqs evaluated in this study, Oq(edge)+ and Oq(edge) were considered to best reflect the vocal fold vibratory characteristics.
在本研究中，我们通过多元线性回归分析（MVA）研究了各种开放商数（Oqs）与发声类型、基频（F0）和强度之间的关系，以确定哪种Oq最能反映声带振动特征。
使用高速数字成像（HSDI），从六名嗓音健康的男性志愿者中记录了不同发声类型、F0和强度下的持续元音/e/：发声类型包括正常发声、假声、正常呼吸声和正常挤压声；每种发声在不同的F0和强度下进行。同时用HSDI记录了电声门图（EGG）和声音信号。从获得的数据中，评估了10个传统的Oqs（4个来自声门面积函数的Oqs、4个波形记录Oqs和2个EGG衍生的Oqs）以及2个新引入的Oqs（Oq(edge)+和Oq(edge)）。并且，通过MVA评估了各种Oqs与发声类型、F0和强度之间的关系。
在各种Oqs中，Oq(edge)+和Oq(edge)与声学特性的相关性最强，并且最能描述发声类型的变化：发现Oq(edge)比Oq(edge)¯更好。来自五线多线波形记录的五个Oqs的平均值Oq(MLK)是Oq(edge)¯的一个很好的替代。EGG衍生的Oqs能够区分正常发声和假声，但需要同时考虑F0的变化。MVA显示了正常发声和其他发声类型之间Oq值的变化、强度的参与程度，以及F0和Oqs之间没有关系。
在本研究评估的Oqs中，Oq(edge)+和Oq(edge)被认为最能反映声带振动特征。
