13
$\begingroup$

Please tell me. Also I dont know why phase is linear with FIR filters. I would like quantitative analysis. And why linear phase is not achieved by IIR filters ?

$\endgroup$
  • $\begingroup$ Causal IIR filters cannot have linear phase, non-causal IIR filters can. As in $H(e^{j\omega})=e^{-j\omega\alpha}$ $\endgroup$ – Parsa Oct 29 '14 at 2:46
9
$\begingroup$

For digital filters, linear phase places the following requirement on the transfer function:

$$ H(z) = H(z^{-1}).$$

That restriction implies a linear phase IIR filter would need to have poles both inside and outside the unit circle, making it unstable. Similar arguments apply for analog filters.

That being said, there are any number of approximations which may be "close enough" to linear phase, depending on the application--especially if the causality of filter is sacrificed. For a review of techniques, see the introduction to this paper:

S.R. Powell, P.M.Chau, A Technique for Realizing Linear-Phase IIR Filters, IEEE Trans. Signal Processing, Vol 39, No 11, Nov 1991, pp 2425-2435.

The algorithm in that paper achieves linear phase with acausal block processing, rather than the usual offline "Forward-Backward" zero-phase approach.

$\endgroup$
  • $\begingroup$ Shouldn't your first sentence say "zero phase", not "linear phase"? when "the causality of filter is sacrificed", you are converting zero-phase to delayed linear phase, no? $\endgroup$ – endolith Apr 22 '15 at 22:51
  • $\begingroup$ @endolith Isn’t a flat line still a line? $\endgroup$ – Stanley Pawlukiewicz Nov 10 '17 at 23:06
  • $\begingroup$ @StanleyPawlukiewicz The requirement specified is zero phase. Filters can be linear phase and not meet this requirement. $\endgroup$ – endolith Nov 11 '17 at 20:43
7
$\begingroup$

The impulse response of a linear phase filter must be symmetric. If the impulse response is infinitely long, then the center of the impulse is an infinite distance away from the beginning, giving the symmetric IIR filter infinite delay.

$\endgroup$
  • 1
    $\begingroup$ "The impulse response of a linear phase filter must be symmetric" ... though not necessarily symmetric about zero? $\endgroup$ – endolith Apr 22 '15 at 22:53
  • 1
    $\begingroup$ Your question is more about if the signal needs to be even, since that is the sharpest form of symmetry. Regarding hotpaw2's comment I wonder by what logic it needs be symmetric? $\endgroup$ – Starhowl Apr 29 '17 at 7:43
2
$\begingroup$

Clements and Pease have shown that causal infinite-duration impulse responses can also have Fourier transforms with generalized linear phase. The corresponding system functions, however, are not rational, and thus, the systems cannot be implemented with difference equations.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.