Function stft

espnet2/legacy/transform/spectrogram.py:7–41

Process STFT from signal.

(
    x, n_fft, n_shift, win_length=None, window="hann", center=True, pad_mode="reflect"
)

def stft(
    x, n_fft, n_shift, win_length=None, window="hann", center=True, pad_mode="reflect"
):
    """Process STFT from signal."""
    # x: [Time, Channel]
    if x.ndim == 1:
        single_channel = True
        # x: [Time] -> [Time, Channel]
        x = x[:, None]
    else:
        single_channel = False
    x = x.astype(np.float32)

    # FIXME(kamo): librosa.stft can't use multi-channel?
    # x: [Time, Channel, Freq]
    x = np.stack(
        [
            librosa.stft(
                x[:, ch],
                n_fft=n_fft,
                hop_length=n_shift,
                win_length=win_length,
                window=window,
                center=center,
                pad_mode=pad_mode,
            ).T
            for ch in range(x.shape[1])
        ],
        axis=1,
    )

    if single_channel:
        # x: [Time, Channel, Freq] -> [Time, Freq]
        x = x[:, 0]
    return x


def istft(x, n_shift, win_length=None, window="hann", center=True):