摘要
本文以目前社会上流行的两款不同类别的电子伪装语音器材为对象,对多名实验对象变声前、后的声学特征(基频、共振峰、声调、能量和过零率等)进行了深入的实验分析。实验结果和分析表明,变声相对于原声,其声学特征变化是有规律的,根据这一声学特征变化规律,对变声语音逆变声就能得到与原声符合很好的恢复语音,这为深入开展对经过电子伪装的语音进行同一人鉴定奠定了基础。
In this paper,two different voice changer applications were analyzed and used to create disguised voices.The acoustic features,such as pitch,formant,tone,energy,and zero-crossing rate,of both the disguised voices and the original voices were analyzed.Results showed that,compared to the original voices,the acoustic features of the disguised voices had their particular changing rules.The experiment also found that,according to the rules,the disguised voices could be easily recovered very well,which laid the foundation for forensic electronic disguised voice identification in the future.
出处
《证据科学》
2010年第4期503-509,共7页
Evidence Science
关键词
电子伪装语音
变声规律
声学特征
同一人鉴定
Eelectronic disguised voice
Voice conversion
Acoustic features
Speaker identification
