fftr

class diffsptk.RealValuedFastFourierTransform(fft_length: int, out_format: str | int = 'complex', learnable: bool = False)

See this page for details.

Parameters:

fft_lengthint >= 2

The FFT length, \(L\).

out_format[‘complex’, ‘real’, ‘imaginary’, ‘amplitude’, ‘power’]

The output format.

learnablebool

Whether to make the DFT basis learnable. If True, the module performs DFT rather than FFT.

forward(x: Tensor) → Tensor

Compute FFT of a real signal.

Parameters:

xTensor [shape=(…, N)]

The real input signal.

Returns:

outTensor [shape=(…, L/2+1)]

The output spectrum.

Examples

>>> x = diffsptk.ramp(1, 3)
>>> x
tensor([1., 2., 3.])
>>> fftr = diffsptk.RealValuedFastFourierTransform(8, out_format="real")
>>> y = fftr(x)
>>> y
tensor([ 6.0000,  2.4142, -2.0000, -0.4142,  2.0000])

diffsptk.functional.fftr(x: Tensor, fft_length: int | None = None, out_format: str = 'complex') → Tensor

Compute FFT of a real signal.

Parameters:

xTensor [shape=(…, N)]

The real input signal.

fft_lengthint >= 2 or None

The FFT length, \(L\).

out_format[‘complex’, ‘real’, ‘imaginary’, ‘amplitude’, ‘power’]

The output format.

Returns:

outTensor [shape=(…, L/2+1)]

The output spectrum.

See also

ifftr stft spec