preprocessing.py 文件源码

def _call(self, signal):
        """Compute MFCC features from an audio signal.

        Args:
            signal: the audio signal from which to compute features. Should be
            an N*1 array

        Returns:
            A numpy array of size (NUMFRAMES by numcep) containing features.
            Each row holds 1 feature vector.
        """
        feat, energy = super(MFCC, self)._call(signal)

        feat = np.log(feat)
        feat = dct(feat, type=2, axis=1, norm='ortho')[:, :self.num_cep]
        feat = self._lifter(feat, self.cep_lifter)

        if self.append_energy:
            # replace first cepstral coefficient with log of frame energy
            feat[:, 0] = np.log(energy + self.eps)

        if self.d:
            d = delta(feat, 2)
            feat = np.hstack([feat, d])

            if self.dd:
                feat = np.hstack([feat, delta(d, 2)])

        return feat