gammatone#

class diffsptk.GammatoneFilterBankAnalysis(sample_rate: int, *, f_min: float = 70, f_base: float = 1000, f_max: float = 6700, filter_order: int = 4, bandwidth_factor: float = 1, density: float = 1, exact: bool = False)[source]#

Gammatone filter bank analysis module.

Parameters:

sample_rateint >= 1: The sample rate in Hz.
f_minfloat >= 0: The minimum frequency in Hz.
f_basefloat >= 0: The base frequency in Hz.
f_maxfloat <= sample_rate // 2: The maximum frequency in Hz.
filter_orderint >= 1: The order of the Gammatone filter.
bandwidth_factorfloat > 0: The bandwidth of the Gammatone filter.
densityfloat > 0: The density of frequencies on the ERB scale.
exactbool: If False, use all-pole approximation.

References

[1]

V. Hohmann, “Frequency analysis and synthesis using a Gammatone filterbank,” Acta Acustica united with Acustica, vol. 88, no. 3, pp. 433-442, 2002.

forward(x: Tensor) → Tensor[source]#

Apply Gammatone filter banks to the input signal.

Parameters:

xTensor [shape=(B, 1, T) or (B, T) or (T,)]: The input signal.

Returns:

outTensor [shape=(B, K, T)]: The analyzed signal.

Examples

>>> x = diffsptk.impulse(15999)
>>> gammatone = diffsptk.GammatoneFilterBankAnalysis(16000)
>>> y = gammatone(x)
>>> y.shape
torch.Size([1, 30, 16000])