grpdelay#

class diffsptk.GroupDelay(fft_length, alpha=1, gamma=1)[source]#

See this page for details.

Parameters:

fft_lengthint >= 2: Number of FFT bins, $L$ .
alphafloat > 0: Tuning parameter, $α$ .
gammafloat > 0: Tuning parameter, $γ$ .

forward(b=None, a=None)[source]#

Compute group delay.

Parameters:

bTensor [shape=(…, M+1)] or None: Numerator coefficients.
aTensor [shape=(…, N+1)] or None: Denominator coefficients.

Returns:

outTensor [shape=(…, L/2+1)]: Group delay or modified group delay function.

Examples

>>> x = diffsptk.ramp(3)
>>> grpdelay = diffsptk.GroupDelay(8)
>>> g = grpdelay(x)
>>> g
tensor([2.3333, 2.4278, 3.0000, 3.9252, 3.0000])

diffsptk.functional.grpdelay(b=None, a=None, *, fft_length=512, alpha=1, gamma=1, **kwargs)[source]#

Compute group delay.

Parameters:

bTensor [shape=(…, M+1)] or None: Numerator coefficients.
aTensor [shape=(…, N+1)] or None: Denominator coefficients.
fft_lengthint >= 2: Number of FFT bins, $L$ .
alphafloat > 0: Tuning parameter, $α$ .
gammafloat > 0: Tuning parameter, $γ$ .

Returns:

outTensor [shape=(…, L/2+1)]: Group delay or modified group delay function.

See also