中文说明：这是个完整的实现VQ方法的基于MFCC的说话人识别代码。首先对语音片段进行VAD，去除了无声段，在进行特征的提取，为每一个说话人建立各自的码本，训练完成之后当有待检测的语音时候，输入语音，并提取特征，然后给出说话人的标签，程序完整有效，可供初步学习说话人识别的人进行参考，方可了解其基本的框架和流程。
English Description：
This is a complete implementation of VQ based on MFCC in speaker recognition code. First on voice fragment for VAD, removal has silent segment, in for features of extraction, for each a talk people established respective of yards this, training completed zhihou Dang needs to be detection of voice when, entered voice, and extraction features, then to out talk people of label, program full effective, can for preliminary learning talk people recognition of people for reference, party can understanding its basic of framework and process.