摘要
本文主要通过对正常、大声两种说话状态下的普通话中三个单元音[a]、[i]、[u]的声强、时长、基频、谐波振幅差值、共振峰等声学参数的分析,综合比较了各参数的变化规律,发现大声说话时的语音并非正常语音的简单放大,二者不仅在声强上存在差别,同时在频率域上也发生了重要变化。同一人不同状态下发音的频谱特征差异性较大,同种状态下发音的相似性、可比性较强,为此,声纹鉴定中应尽量选取状态相同的语音进行比对。
This paper analyzes such acoustic parameters as the intensity, duration, fundamental frequency and harmonic amplitude of the three Mandarin monophthongs under the normal voice and loud voice. It comprehensively compares the change rule of each parameter and finds that the loud voice was not a simple amplification of the normal voice. They are different, not only in intensity, but also in the frequency domain. The spectrum of the same speaker＇s voice shows a bigger dissimilarity at different volume levels but a robust similarity and comparability at same volume level. Therefore, we should select and compare the voices at same volume level in acoustic identification.
出处
《证据科学》
2009年第6期754-764,共11页
Evidence Science
基金
中国政法大学证据科学研究院(证据科学教育部重点实验室)科研项目培育资金项目资助
关键词
正常语音
大声说话
声学分析
声纹鉴定
Normal Voice, Loud Voice, Acoustic Analysis, Forensic Voice Identification
