mgcep#

class diffsptk.MelCepstralAnalysis(*, fft_length: int, cep_order: int, alpha: float = 0, n_iter: int = 0)[source]#

See this page for details. Note that the current implementation does not use the efficient Toeplitz-plus-Hankel system solver.

Parameters:

fft_lengthint >= 2M: The number of FFT bins, \(L\).
cep_orderint >= 0: The order of the cepstrum, \(M\).
alphafloat in (-1, 1): The frequency warping factor, \(\alpha\).
n_iterint >= 0: The number of iterations.

forward(x: Tensor)[source]#

Perform mel-cepstral analysis.

Parameters:

xTensor [shape=(…, L/2+1)]: The power spectrum.

Returns:

outTensor [shape=(…, M+1)]: The mel-cepstrum.

Examples

>>> x = diffsptk.ramp(19)
>>> stft = diffsptk.STFT(frame_length=10, frame_period=10, fft_length=16)
>>> mcep = diffsptk.MelCepstralAnalysis(
...     fft_length=16, cep_order=3, alpha=0.1, n_iter=1
... )
>>> mc = mcep(stft(x))
>>> mc
tensor([[-0.8851,  0.7917, -0.1737,  0.0175],
        [-0.3522,  4.4222, -1.0882, -0.0511]])

diffsptk.functional.mcep(x: Tensor, cep_order: int, alpha: float = 0, n_iter: int = 0) → Tensor[source]#

Perform mel-cepstral analysis.

Parameters:

xTensor [shape=(…, L/2+1)]: The power spectrum.
cep_orderint >= 0: The order of the cepstrum, \(M\).
alphafloat in (-1, 1): The frequency warping factor, \(\alpha\).
n_iterint >= 0: The number of iterations.

Returns:

outTensor [shape=(…, M+1)]: The mel-cepstrum.

class diffsptk.MelGeneralizedCepstralAnalysis(*, fft_length: int, cep_order: int, alpha: float = 0, gamma: float = 0, c: int | None = None, n_iter: int = 0)[source]#

See this page for details. Note that the current implementation does not use the efficient Toeplitz-plus-Hankel system solver.

Parameters:

fft_lengthint >= 2M: The number of FFT bins, \(L\).
cep_orderint >= 0: The order of the mel-cepstrum, \(M\).
alphafloat in (-1, 1): The frequency warping factor, \(\alpha\).
gammafloat in [-1, 0]: The gamma parameter, \(\gamma\).
cint >= 1 or None: The number of stages.
n_iterint >= 0: THe number of iterations.

forward(x: Tensor) → Tensor[source]#

Estimate mel-generalized cepstrum from spectrum.

Parameters:

xTensor [shape=(…, L/2+1)]: The power spectrum.

Returns:

outTensor [shape=(…, M+1)]: The mel-generalized cepstrum.

Examples

>>> x = diffsptk.ramp(19)
>>> stft = diffsptk.STFT(frame_length=10, frame_period=10, fft_length=16)
>>> mgcep = diffsptk.MelGeneralizedCepstralAnalysis(
...     fft_length=16, cep_order=3, alpha=0.1, n_iter=1
... )
>>> mc = mgcep(stft(x))
>>> mc
tensor([[-0.8851,  0.7917, -0.1737,  0.0175],
        [-0.3522,  4.4222, -1.0882, -0.0511]])