speaker-recognition

  mfcc用于求取训练样本的特征值，其需要melbankm和enframe两个程序。 enframe用于将语音分帧，melbankm用于产生mel三角滤波器组，vqlbg用于获取说话人码书，disteu程序用于计算两个矩阵间的距离，(mfcc strike eigenvalues ​ ​ for the training sample, which requires melbankm and enframe two procedures. enframe for the voice of the frame, melbankm used to produce the triangular mel filter bank, vqlbg codebook is used to get people talking, disteu procedures used to calculate the distance between two matrices,)