ipnorm#

class diffsptk.MelCepstrumInversePowerNormalization(cep_order: int)[source]#

See this page for details.

Parameters:

cep_orderint >= 0: The order of the cepstrum, \(M\).

forward(y: Tensor) → Tensor[source]#

Perform mel-cepstrum inverse power normalization.

Parameters:

yTensor [shape=(…, M+2)]: The log power and power-normalized cepstrum.

Returns:

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

Examples

>>> x = diffsptk.ramp(1, 4)
>>> pnorm = diffsptk.MelCepstrumPowerNormalization(3, alpha=0.1)
>>> ipnorm = diffsptk.MelCepstrumInversePowerNormalization(3)
>>> y = ipnorm(pnorm(x))
>>> y
tensor([1., 2., 3., 4.])

diffsptk.functional.ipnorm(y: Tensor) → Tensor[source]#

Perform cepstrum inverse power normalization.

Parameters:

yTensor [shape=(…, M+2)]: The log power and power-normalized cepstrum.

Returns:

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