scipy.signal — Signal processing

The scipy.signal submodule provides a single signal-processing routine.

Functions

scipy.signal.sosfilt(sos: ndarray, x: ndarray, *, zi: ndarray | None = None) → ndarray | tuple[ndarray, ndarray]

Filter the one-dimensional, uniformly sampled data x using the cascaded second-order sections in sos.

Parameters:

• sos – an array-like of shape (n_sections, 6) giving the filter sections. Each row holds the six coefficients [b0, b1, b2, a0, a1, a2] of one biquad section.

• x – a one-dimensional input numpy.ndarray.

• zi – optional initial filter delays, a float numpy.ndarray of shape (n_sections, 2). If omitted, the initial conditions are taken to be zero.

Returns:

the filtered signal as a float numpy.ndarray. When zi is supplied, the return value is a 2-tuple (y, zf) where zf holds the final delays in the same shape as zi.