一个非常好的说话人识别特征提取程序，直接对wav文件操作，提取mfcc语音特征值，通过调试-A very good speaker recognition feature extraction process, directly on the wav file, the extraction mfcc voice eigenvalue, through the debugger

带双步的QR分解方法求所有特征值，以及反幂法求解对应特征向量-two-steps-qr-descomposure to solve all eigenvalues of a matrix. and antipower method to solve a engenvector correspongding to the eigenvalue