摘要： Ｗｉｎｄｏｗｓ的ＷＡＶ文件提供的语音信息可以满足语音特征识别的要求．根据讲话者在讲话时鼻腔相对固定以及汉语鼻音声母少的特点，使用极零模型，并以所有汉语鼻声母音节的零点谱为参数，实现汉语讲话音辨别．所选择的参数有效，系统模板存储量少，系统训练和识别的运算量小．

关键词: WAV文件, 讲话者识别, 鼻音声母, 极零模型

Abstract: Phonic information which WAV file of Windows provides can satisfy request of Phonic recognititon.According to that nasal cavity is relatively fixed and Chinese nasal sound is little,we use pole-zero model and take zero spectrum of all Chinese nasal sound syllable as parameter to realize phonic recognition.Parameter is valid and storage of systematic model is little etc.