InverseSpectrogram

class torchaudio.transforms.InverseSpectrogram(n_fft: int = 400, win_length: ~typing.Optional[int] = None, hop_length: ~typing.Optional[int] = None, pad: int = 0, window_fn: ~typing.Callable[[...], ~torch.Tensor] = <built-in method hann_window of type object>, normalized: ~typing.Union[bool, str] = False, wkwargs: ~typing.Optional[dict] = None, center: bool = True, pad_mode: str = 'reflect', onesided: bool = True)

Create an inverse spectrogram to recover an audio signal from a spectrogram.

Parameters

• n_fft (int, optional) – Size of FFT, creates n_fft // 2 + 1 bins. (Default: 400)

• win_length (int or None, optional) – Window size. (Default: n_fft)

• hop_length (int or None, optional) – Length of hop between STFT windows. (Default: win_length // 2)

• pad (int, optional) – Two sided padding of signal. (Default: 0)

• window_fn (Callable[..., Tensor], optional) – A function to create a window tensor that is applied/multiplied to each frame/window. (Default: torch.hann_window)

• normalized (bool or str, optional) – Whether the stft output was normalized by magnitude. If input is str, choices are "window" and "frame_length", dependent on normalization mode. True maps to "window". (Default: False)

• wkwargs (dict or None, optional) – Arguments for window function. (Default: None)

• center (bool, optional) – whether the signal in spectrogram was padded on both sides so that the \(t\)-th frame is centered at time \(t \times \text{hop\_length}\). (Default: True)

• pad_mode (string, optional) – controls the padding method used when center is True. (Default: "reflect")

• onesided (bool, optional) – controls whether spectrogram was used to return half of results to avoid redundancy (Default: True)

Example

>>> batch, freq, time = 2, 257, 100
>>> length = 25344
>>> spectrogram = torch.randn(batch, freq, time, dtype=torch.cdouble)
>>> transform = transforms.InverseSpectrogram(n_fft=512)
>>> waveform = transform(spectrogram, length)

Tutorials using InverseSpectrogram:

Audio Feature Augmentation

Audio Feature Augmentation

Speech Enhancement with MVDR Beamforming

Speech Enhancement with MVDR Beamforming

forward(spectrogram: Tensor, length: Optional[int] = None) → Tensor

Parameters

• spectrogram (Tensor) – Complex tensor of audio of dimension (…, freq, time).

• length (int or None, optional) – The output length of the waveform.

Returns

Dimension (…, time), Least squares estimation of the original signal.

Return type

Tensor