c2mpir#

class diffsptk.CepstrumToMinimumPhaseImpulseResponse(cep_order, ir_length, n_fft=512)[source]#

See this page for details. The conversion uses FFT instead of recursive formula.

Parameters:

cep_orderint >= 0 [scalar]: Order of cepstrum, \(M\).
ir_lengthint >= 1 [scalar]: Length of impulse response, \(N\).
n_fftint >> \(N\) [scalar]: Number of FFT bins. Accurate conversion requires the large value.

forward(c)[source]#

Convert cepstrum to minimum phase impulse response.

Parameters:

cTensor [shape=(…, M+1)]: Cepstral coefficients.

Returns:

hTensor [shape=(…, N)]: Truncated minimum phase impulse response.

Examples

>>> c = diffsptk.ramp(3)
>>> c2mpir = diffsptk.CepstrumToMinimumPhaseImpulseResponse(3, 5)
>>> h = c2mpir(c)
>>> h
tensor([1.0000, 1.0000, 2.5000, 5.1667, 6.0417])