sliding_window()

audinterface.utils.sliding_window(signal, sampling_rate, win_dur, hop_dur)

Reshape signal by applying a sliding window.

Windows that do not match the specified duration at the end of the signals will be dropped.

Parameters:

• signal (ndarray) – input signal in shape (samples,) or (channels, samples)

• sampling_rate (int) – sampling rate in Hz

• win_dur (Union[float, int, str, Timedelta]) – window duration, if value is as a float or integer it is treated as seconds. See audinterface.utils.to_timedelta() for further options

• hop_dur (Union[float, int, str, Timedelta]) – hop duration, if value is as a float or integer it is treated as seconds. See audinterface.utils.to_timedelta() for further options

Return type:

ndarray

Returns:

view of signal with shape (channels, samples, frames)

Raises:

ValueError – if win_dur or hop_dur is smaller than 1/sampling_rate

Examples

>>> signal = np.array(
...     [
...         [0, 1, 2, 3, 4, 5],
...         [0, 10, 20, 30, 40, 50],
...     ],
... )
>>> signal
array([[ 0,  1,  2,  3,  4,  5],
       [ 0, 10, 20, 30, 40, 50]])
>>> frames = sliding_window(
...     signal,
...     sampling_rate=1,
...     win_dur=3,
...     hop_dur=2,
... )
>>> # First frame
>>> frames[..., 0]
array([[ 0,  1,  2],
       [ 0, 10, 20]])
>>> # Last frame
>>> frames[..., -1]
array([[ 2,  3,  4],
       [20, 30, 40]])
>>> # Mean per frame
>>> frames.mean(axis=1)
array([[ 1.,  3.],
       [10., 30.]])