irfft(input, signal_ndim, normalized=False, onesided=True, signal_sizes=None) → Tensor¶
Complex-to-real Inverse Discrete Fourier Transform.
torch.irfft()is deprecated and will be removed in a future PyTorch release. Use the new torch.fft module functions, instead, by importing torch.fft and calling
torch.fft.irfft()for one-sided input, or
torch.fft.ifft()for two-sided input.
This method computes the complex-to-real inverse discrete Fourier transform. It is mathematically equivalent with
ifft()with differences only in formats of the input and output.
The argument specifications are almost identical with
ifft(). Similar to
normalizedis set to
True, this normalizes the result by multiplying it with so that the operator is unitary, where is the size of signal dimension .
Due to the conjugate symmetry,
inputdo not need to contain the full complex frequency values. Roughly half of the values will be sufficient, as is the case when
inputis given by
rfft(signal, onesided=True). In such case, set the
onesidedargument of this method to
True. Moreover, the original signal shape information can sometimes be lost, optionally set
signal_sizesto be the size of the original signal (without the batch dimensions if in batched mode) to recover it with correct shape.
Therefore, to invert an
onesidedarguments should be set identically for
irfft(), and preferably a
signal_sizesis given to avoid size mismatch. See the example below for a case of size mismatch.
rfft()for details on conjugate symmetry.
The inverse of this function is
Generally speaking, input to this function should contain values following conjugate symmetry. Note that even if
True, often symmetry on some part is still needed. When this requirement is not satisfied, the behavior of
irfft()is undefined. Since
torch.autograd.gradcheck()estimates numerical Jacobian with point perturbations,
irfft()will almost certainly fail the check.
For CUDA tensors, an LRU cache is used for cuFFT plans to speed up repeatedly running FFT methods on tensors of same geometry with same configuration. See cuFFT plan cache for more details on how to monitor and control the cache.
Due to limited dynamic range of half datatype, performing this operation in half precision may cause the first element of result to overflow for certain inputs.
For CPU tensors, this method is currently only available with MKL. Use
torch.backends.mkl.is_available()to check if MKL is installed.
input (Tensor) – the input tensor of at least
signal_ndim (int) – the number of dimensions in each signal.
signal_ndimcan only be 1, 2 or 3
normalized (bool, optional) – controls whether to return normalized results. Default:
signal_sizes (list or
torch.Size, optional) – the size of the original signal (without batch dimension). Default:
A tensor containing the complex-to-real inverse Fourier transform result
- Return type
>>> x = torch.randn(4, 4) >>> torch.rfft(x, 2, onesided=True).shape torch.Size([4, 3, 2]) >>> >>> # notice that with onesided=True, output size does not determine the original signal size >>> x = torch.randn(4, 5) >>> torch.rfft(x, 2, onesided=True).shape torch.Size([4, 3, 2]) >>> >>> # now we use the original shape to recover x >>> x tensor([[-0.8992, 0.6117, -1.6091, -0.4155, -0.8346], [-2.1596, -0.0853, 0.7232, 0.1941, -0.0789], [-2.0329, 1.1031, 0.6869, -0.5042, 0.9895], [-0.1884, 0.2858, -1.5831, 0.9917, -0.8356]]) >>> y = torch.rfft(x, 2, onesided=True) >>> torch.irfft(y, 2, onesided=True, signal_sizes=x.shape) # recover x tensor([[-0.8992, 0.6117, -1.6091, -0.4155, -0.8346], [-2.1596, -0.0853, 0.7232, 0.1941, -0.0789], [-2.0329, 1.1031, 0.6869, -0.5042, 0.9895], [-0.1884, 0.2858, -1.5831, 0.9917, -0.8356]])