cupyx.scipy.signal.group_delay(system, w=512, whole=False, fs=6.283185307179586)[source]#

Compute the group delay of a digital filter.

The group delay measures by how many samples amplitude envelopes of various spectral components of a signal are delayed by a filter. It is formally defined as the derivative of continuous (unwrapped) phase:

          d        jw
D(w) = - -- arg H(e)
  • system (tuple of array_like (b, a)) – Numerator and denominator coefficients of a filter transfer function.

  • w ({None, int, array_like}, optional) –

    If a single integer, then compute at that many frequencies (default is N=512).

    If an array_like, compute the delay at the frequencies given. These are in the same units as fs.

  • whole (bool, optional) – Normally, frequencies are computed from 0 to the Nyquist frequency, fs/2 (upper-half of unit-circle). If whole is True, compute frequencies from 0 to fs. Ignored if w is array_like.

  • fs (float, optional) – The sampling frequency of the digital system. Defaults to 2*pi radians/sample (so w is from 0 to pi).


  • w (ndarray) – The frequencies at which group delay was computed, in the same units as fs. By default, w is normalized to the range [0, pi) (radians/sample).

  • gd (ndarray) – The group delay.

See also


Frequency response of a digital filter


The similar function in MATLAB is called grpdelay.

If the transfer function \(H(z)\) has zeros or poles on the unit circle, the group delay at corresponding frequencies is undefined. When such a case arises the warning is raised and the group delay is set to 0 at those frequencies.

For the details of numerical computation of the group delay refer to [1].