lpc#

diffsptk.LPC#: alias of LinearPredictiveCodingAnalysis

class diffsptk.LinearPredictiveCodingAnalysis(frame_length: int, lpc_order: int, eps: float = 1e-06)[source]#

See this page for details. Double precision is recommended.

Parameters:

frame_lengthint > M: The frame length, \(L\).
lpc_orderint >= 0: The order of the LPC coefficients, \(M\).
epsfloat >= 0: A small value to improve numerical stability.

forward(x: Tensor) → Tensor[source]#

Perform LPC analysis.

Parameters:

xTensor [shape=(…, L)]: The framed waveform.

Returns:

outTensor [shape=(…, M+1)]: The gain and LPC coefficients.

Examples

>>> x = diffsptk.nrand(4)
tensor([ 0.8226, -0.0284, -0.5715,  0.2127,  0.1217])
>>> lpc = diffsptk.LPC(5, 2)
>>> a = lpc(x)
>>> a
tensor([0.8726, 0.1475, 0.5270])

diffsptk.functional.lpc(x: Tensor, lpc_order: int, eps: float = 1e-06) → Tensor[source]#

Perform LPC analysis.

Parameters:

xTensor [shape=(…, L)]: The famed waveform.
lpc_orderint >= 0: The order of the LPC coefficients, \(M\).
epsfloat >= 0: A small value to improve numerical stability.

Returns:

outTensor [shape=(…, M+1)]: The gain and LPC coefficients.