python语音处理库librosa的java实现。
–> 加载音频文件,读取幅值(magnitude)
–> 梅尔频率倒谱系数
–> 从wav提取mel(MelSpectrogram)特征值
–> 短时傅里叶变换
–> 短时傅立叶逆变换(ISTFT)
运行成功后,命令行应该看到下面的信息:
... Audio Feature Values: 0.000040 0.000040 0.000000 0.000000 0.000100 0.000070 0.000070 0.000000 0.000000 0.000000 Sample Rate: 16000 ....... Size of MFCC Feature Values: (40 , 97 ) -643.894348 -644.722900 -645.119995 -629.035706 -542.736816 -473.609894 -457.379211 -463.240082 -407.562378 -255.510406 ....... Size of STFT Feature Values: (1025 , 97 ) Real and Imag values of STFT are 0.031431286060502925,0.0 Real and Imag values of STFT are 0.02153403593129419,0.0 Real and Imag values of STFT are 0.00658287100191068,0.0 Real and Imag values of STFT are 0.02978720482718357,0.0 Real and Imag values of STFT are 0.15048249086374446,0.0 Real and Imag values of STFT are 0.2464715605426407,0.0 Real and Imag values of STFT are 0.2570755996529303,0.0 Real and Imag values of STFT are 0.15607790080422798,0.0 Real and Imag values of STFT are -0.047981990330747716,0.0 Real and Imag values of STFT are -0.27254267235230273,0.0
http://www.aias.top/
https://github.com/mymagicpower/AIAS
https://gitee.com/mymagicpower/AIAS