COLA

class COLA(process, store, n_total, n_samples, n_overlap, sfreq, window='hann', tol=1e-10, *, verbose=None)[source][source]

Constant overlap-add processing helper.

Parameters:

  • process (callable) – A function that takes a chunk of input data with shape (n_channels, n_samples) and processes it.

  • store (callable | ndarray) – A function that takes a completed chunk of output data. Can also be an ndarray, in which case it is treated as the output data in which to store the results.

  • n_total (int) – The total number of samples.

  • n_samples (int) – The number of samples per window.

  • n_overlap (int) – The overlap between windows.

  • window (str) – The window to use. Default is “hann”.

  • tol (float) – The tolerance for COLA checking.

  • n_jobs (int) – The number of jobs to run in parallel.

  • verbose (bool) – If True, print a message when the COLA condition is not met.

Notes

This will process data using overlapping windows to achieve a constant output value. For example, for n_total=27, n_samples=10, n_overlap=5 and window='triang'

1 _____               _______
  |    \   /\   /\   /
  |     \ /  \ /  \ /
  |      x    x    x
  |     / \  / \  / \
  |    /   \/   \/   \
0 +----|----|----|----|----|-
  0    5   10   15   20   25

This produces four windows: the first three are the requested length (10 samples) and the last one is longer (12 samples). The first and last window are asymmetric.

feed(*datas, verbose=None, **kwargs)[source][source]

Pass in a chunk of data.