网络语音同一性鉴定中的共振峰差异分析  

DIFFERENCE ANALYSIS OF FORMANT IN NETWORK VOICE IDENTIFICATION

作　　者：张晓[1] 孔华锋 王海燕 刘文超 Zhang Xiao;Kong Huafeng;Wang Haiyan;Liu Wenchao(Key Laboratory of Information Network Security, Third Research Institute of the Ministry of Public Security, Shanghai 201204, China;Wuhan Business University, Wuhan 430056, Hubei, China;College of Criminal Justice, East China University of Political Science and Law, Shanghai 201620, China)

机构地区：[1]公安部第三研究所信息网络安全公安部重点实验室,上海201204 [2]武汉商学院,湖北武汉430056 [3]华东政法大学刑事司法学院,上海201620

出　　处：《计算机应用与软件》2019年第3期187-191,共5页Computer Applications and Software

基　　金：公安部科技强警基础工作专项项目(2017GABJC33)

摘　　要：语音同一性鉴定中共振峰的特征比对是认定同一的重要方法。司法鉴定中的语音检材大多来自于录音笔等电子设备,而网络语音的出现相对于通常的语音同一性鉴定是一种挑战。即时通信软件在传输网络语音的过程中为了保持网络传输的低带宽占用,对语音进行高比例的压缩,因此即时通信软件产生的网络语音的共振峰特征相对于原始语音会有变化。选取当前互联网社交领域常用的5款即时通信软件,通过实验分析出即时通信软件产生的网络语音与原始语音在共振峰特征上的差异。研究结果表明,不同即时通信软件在不同元音的共振峰特征上会呈现出变化差异,通过总结变化差异可以提高即时通信软件的网络语音同一性鉴定的准确率。In the identification of voice identity, the feature comparison of formants is an important method to identify. Most of the voice samples in forensic identification come from electronic devices such as recording pens, and the appearance of network voice is a challenge to the usual voice identification. In the process of transmitting network voice, IM software compresses the voice in a high proportion in order to maintain the low bandwidth occupied by network transmission. Therefore, the formant characteristics of network voice generated by IM software changes compared with the original voice. In this paper, we selected five IM software commonly used in the field of Internet social networking and analyzed the differences of formant characteristics between the network voice generated by IM software and the original voice through experiments. The results show that the formant characteristics of different vowels vary with different IM software. And we can improve the accuracy of network voice identification by summarizing the differences.

关 键 词：网络语音 共振峰 差异 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]