stft#

diffsptk.STFT#: alias of ShortTimeFourierTransform

class diffsptk.ShortTimeFourierTransform(frame_length, frame_period, fft_length, *, center=True, zmean=False, window='blackman', norm='power', eps=1e-09, relative_floor=None, out_format='power')[source]#

This module is a simple cascade of framing, windowing, and spectrum calculation.

Parameters:

frame_lengthint >= 1: Frame length, \(L\).
frame_peirodint >= 1: Frame period, \(P\).
fft_lengthint >= L: Number of FFT bins, \(N\).
centerbool: If True, assume that the center of data is the center of frame, otherwise assume that the center of data is the left edge of frame.
zmeanbool: If True, perform mean subtraction on each frame.
window[‘blackman’, ‘hamming’, ‘hanning’, ‘bartlett’, ‘trapezoidal’, ‘rectangular’]: Window type.
norm[‘none’, ‘power’, ‘magnitude’]: Normalization type of window.
epsfloat >= 0: A small value added to power spectrum.
relative_floorfloat < 0 or None: Relative floor in decibels.
out_format[‘db’, ‘log-magnitude’, ‘magnitude’, ‘power’, ‘complex’]: Output format.

forward(x)[source]#

Compute short-time Fourier transform.

Parameters:

xTensor [shape=(…, T)]: Waveform.

Returns:

outTensor [shape=(…, T/P, N/2+1)]: Spectrum.

Examples

>>> x = diffsptk.ramp(1, 3)
>>> x
tensor([1., 2., 3.])
>>> stft = diffsptk.STFT(frame_length=3, frame_period=1, fft_length=8)
>>> y = stft(x)
>>> y
tensor([[1.0000, 1.0000, 1.0000, 1.0000, 1.0000],
        [4.0000, 4.0000, 4.0000, 4.0000, 4.0000],
        [9.0000, 9.0000, 9.0000, 9.0000, 9.0000]])

diffsptk.functional.stft(x, *, frame_length=400, frame_period=80, fft_length=512, center=True, zmean=False, window='blackman', norm='power', eps=1e-09, relative_floor=None, out_format='power')[source]#

Compute short-time Fourier transform.

Parameters:

xTensor [shape=(…, T)]: Waveform.
frame_lengthint >= 1: Frame length, \(L\).
frame_peirodint >= 1: Frame period, \(P\).
fft_lengthint >= L: Number of FFT bins, \(N\).
centerbool: If True, assume that the center of data is the center of frame, otherwise assume that the center of data is the left edge of frame.
zmeanbool: If True, perform mean subtraction on each frame.
window[‘blackman’, ‘hamming’, ‘hanning’, ‘bartlett’, ‘trapezoidal’, ‘rectangular’]: Window type.
norm[‘none’, ‘power’, ‘magnitude’]: Normalization type of window.
epsfloat >= 0: A small value added to power spectrum.
relative_floorfloat < 0 or None: Relative floor in decibels.
out_format[‘db’, ‘log-magnitude’, ‘magnitude’, ‘power’, ‘complex’]: Output format.

Returns:

outTensor [shape=(…, T/P, N/2+1)]: Spectrum.