mc2b#

class diffsptk.MelCepstrumToMLSADigitalFilterCoefficients(cep_order, alpha=0)[source]#

See this page for details.

Parameters:

cep_orderint >= 0: Order of cepstrum, \(M\).
alphafloat in (-1, 1): Frequency warping factor, \(\alpha\).

References

[1]

K. Tokuda et al., “Spectral estimation of speech by mel-generalized cepstral analysis,” Electronics and Communications in Japan, part 3, vol. 76, no. 2, pp. 30-43, 1993.

forward(mc)[source]#

Convert mel-cepstrum to MLSA filter coefficients.

Parameters:

mcTensor [shape=(…, M+1)]: Mel-cepstral coefficients.

Returns:

outTensor [shape=(…, M+1)]: MLSA filter coefficients.

Examples

>>> mc = diffsptk.ramp(4)
>>> mc2b = diffsptk.MelCepstrumToMLSADigitalFilterCoefficients(4, 0.3)
>>> b = mc2b(mc)
>>> b
tensor([-0.1686,  0.5620,  1.4600,  1.8000,  4.0000])

diffsptk.functional.mc2b(mc, alpha=0)[source]#

Convert mel-cepstrum to MLSA digital filter coefficients.

Parameters:

mcTensor [shape=(…, M+1)]: Mel-cepstral coefficients.
alphafloat in (-1, 1): Frequency warping factor, \(\alpha\).

Returns:

outTensor [shape=(…, M+1)]: MLSA digital filter coefficients.

mc2b#

This Page