pysptk.sptk.amcep

pysptk.sptk.amcep(x, b, alpha=0.35, lambda_coef=0.98, step=0.1, tau=0.9, pd=4, eps=1e-06)

Adaptive mel-cepstral analysis

Parameters:

x : float

A input sample

b : array, shape(order + 1), optional

MLSA filter coefficients. The result is stored in-place.

alpha : float, optional

All-pass constant. Default is 0.35.

lambda_coef : float, optional

Leakage factor. Default is 0.98.

step : float, optional

Step size. Default is 0.1.

tau : float, optional

Momentum constant. Default is 0.9.

pd : int, optional

Order of pade approximation. Default is 4.

eps : float, optional

Minimum value for epsilon. Default is 1.0e-6.

Returns:

prederr : float

Prediction error

Raises:

ValueError

if invalid order of pade approximation is specified