数字多媒体技术的发展使多媒体信息得到广泛使用和传播，给人类的信息交流带来极大的便利。随着语音相关技术的发展与逐渐成熟，对于语音信号处理的应用也越来越广泛。数字多媒体信息易于修改的特点，使其面临着恶意篡改带来的严重危机。近年来，手机应用软件市场上出现了大量的变声软件，例如微信变声器、超级变声器等等，类似变声器的下载量动辄上百万，这些应用软件可使说话人的声音发生巨大的改变，致使一般的听话人无法辨认发音人的身份、年龄乃至性别，即使是对话者非常熟悉的人也很难识别出说话者的身份。提出了一种鉴定电子伪装语音的方法，通过GMM模型建模，将其均值矢量构成组合特征，然后基于SVM分类器进行训练和鉴别。通过对比语音伪装前后的梅尔倒谱特征参数的统计特性变化，对特征参数的变化规律进行了分析研究。实验结果表明，提出的方法对电子伪装语音的鉴定正确率达到90%。%With the development of digital multimedia technology,digital information has been widely used and spread,which brings great convenience to human communication. Speech related technology gradually becomes mature,and its application is more and more exten-sive. This kind of information is easy to be modified,so that it is facing a serious crisis of malicious tampering. In recent years,a large number of software appear in mobile phone application store,such as Wechat Voice Changer,Super Voice Changer and so on,which can change the speaker’ s voice a lot. As a result,the listener cannot identify the speaker’ s age and sex,even they are familiar. A novel algo-rithm for identification of electronic disguised voice is put forward based on supervector combined by mean vectors of Gaussian mixture model and SVM classifier for training and identification. By comparing the statistical change of MFCC between nature and disguised voice,the variation of voice parameters is studied. Experimental results show that the identification rate can reach 90%.
展开▼
