lpc2lsp#

class diffsptk.LinearPredictiveCoefficientsToLineSpectralPairs(lpc_order: int, log_gain: bool = False, sample_rate: int | None = None, out_format: str | int = 'radian', device: device | None = None, dtype: dtype | None = None)[source]#

See this page for details.

Parameters:

lpc_orderint >= 0: The order of the input LPC, \(M\).
log_gainbool: If True, output the gain in logarithmic scale.
sample_rateint >= 1 or None: The sample rate in Hz.
out_format[‘radian’, ‘cycle’, ‘khz’, ‘hz’]: The output format.
devicetorch.device or None: The device of this module.
dtypetorch.dtype or None: The data type of this module.

References

[1]

P. Kabal et al., “The computation of line spectral frequencies using Chebyshev polynomials,” IEEE Transactions on Acoustics, Speech, and Signal Processing, vol. 34, no. 6, pp. 1419-1426, 1986.

forward(a: Tensor) → Tensor[source]#

Convert LPC to LSP.

Parameters:

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

Returns:

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

Examples

>>> import diffsptk
>>> lpc = diffsptk.LPC(5, 2)
>>> lpc2lsp = diffsptk.LinearPredictiveCoefficientsToLineSpectralPairs(2)
>>> x = diffsptk.ramp(1, 5) * 0.1
>>> a = lpc(x)
>>> w = lpc2lsp(a)
>>> w
tensor([0.5054, 0.5598, 1.6042])

diffsptk.functional.lpc2lsp(a: Tensor, log_gain: bool = False, sample_rate: int | None = None, out_format: str = 'radian') → Tensor[source]#

Convert LPC to LSP.

Parameters:

aTensor [shape=(…, M+1)]: The LPC coefficients.
log_gainbool: If True, output the gain in logarithmic scale.
sample_rateint >= 1 or None: The sample rate in Hz.
out_format[‘radian’, ‘cycle’, ‘khz’, ‘hz’]: The output format.

Returns:

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