lsp2lpc#

class diffsptk.LineSpectralPairsToLinearPredictiveCoefficients(lpc_order: int, log_gain: bool = False, sample_rate: int | None = None, in_format: str | int = 'radian')[source]#

See this page for details.

Parameters:

lpc_orderint >= 0: The order of the LPC, $M$ .
log_gainbool: If True, assume the input gain is in logarithmic scale.
sample_rateint >= 1 or None: The sample rate in Hz.
in_format[‘radian’, ‘cycle’, ‘khz’, ‘hz’]: The input format.

forward(w: Tensor) → Tensor[source]#

Convert LSP to LPC.

Parameters:

wTensor [shape=(…, M+1)]: The LSP frequencies.

Returns:

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

Examples

>>> w = diffsptk.ramp(3)
>>> lsp2lpc = diffsptk.LineSpectralPairsToLinearPredictiveCoefficients(3)
>>> a = lsp2lpc(w)
>>> a
tensor([ 0.0000,  0.8658, -0.0698,  0.0335])

diffsptk.functional.lsp2lpc(w: Tensor, log_gain: bool = False, sample_rate: int | None = None, in_format: str = 'radian') → Tensor[source]#

Convert LSP to LPC.

Parameters:

wTensor [shape=(…, M+1)]: The LSP frequencies.
log_gainbool: If True, assume the input gain is in logarithmic scale.
sample_rateint >= 1 or None: The sample rate in Hz.
in_format[‘radian’, ‘cycle’, ‘khz’, ‘hz’]: The input format.

Returns:

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