中文摘要：

本文旨在尝试运用同步采集的sP和EGG信号对普通话情感语句中的发声进行探讨，以期为连续语流中的情感语音发声研究探路。我们从喜、怒、中、惧、哀五种情感语句中提取了基频、开商、速度商、频率抖动百分比、振幅抖动百分比、谐波噪声比、PIC、PDC、声门碰撞时间和声门松开时间共十个发声参数，并对它们进行了方差分析和因子分析。结果发现，用音强、音色及时长等参数不好区别的喜怒二情感可以用发声力度等予以很好的区分。总之，从声带的打开接触特征、声带振动的速度惯性特征、声带振动规则度和声带发声力度这四个生理维度人手再参考各发声参数的值，可以很好地区分和界定本文中的五种情感类型。

英文摘要：

Based on simultaneously recorded SP and EGG signals, this paper attempts to find out whether there are significantly different pho- nation features between Putonghua sentences uttered in five different emotional states （joyful, irritated, neutral, frightened and sad ） . Ten phonation parameters （FOEGG, OQEGG, SQEGG, JitterEGG, ShimmerEGG, HNREGG, PIC, PDC, VAT, and VRT ） are extracted from these sentences. ANOVA and factor analysis show, （ 1 ） that joy and irritation, which have not been well distinguished by intensity, duration and articulation quality, are now easily told apart by excitation strength; （ 2 ） and that the four physiological aspects of vibrating vocal folds namely regularity, rate/inertia, excitation strength and openness/contact prove to be efficiently distinctive among the five emotions under study.