Delta特征用于说话人识别的研究  

Research on Application of Delta Feature to Speaker Recognition

作　　者：张凯[1] 朱立新[1] 金家宝[1] 

机构地区：[1]解放军电子工程学院,安徽合肥230037

出　　处：《电声技术》2009年第4期52-55,共4页Audio Engineering

摘　　要：多数说话人识别方法采用的都是基于对语音信号的静态特征进行分析,忽略了语音信号动态特征对识别性能的影响。而Delta特征是反映语音信号帧间动态特征的重要特征,对LPC和它的Delta特征进行了具体实例求解,并对计算结果进行了分析,探讨了将它用于说话人识别系统的有效性和可行性。In most adopted methods of speaker recognition, static features of speech signal are analyzed, but the influence of speech signal＇s dynamic features on discernment performance is neglected. Delta features is an important feature which could reflect the inter-frame dynamic feature of speech signal. The LPC（Linear Prediction Coefficient） and Delta feature of one example are educed and analyzed, then the validity and feasibility of their application to speaker recognition are educed.

关 键 词：Delta特征 说话人识别 线性预测系数 

分 类 号：TN912[电子电信—通信与信息系统]