Re: delay added by matlab butterworth filter

Dirk,

I think it depends on how one construes "delay". A _real_ Butterworth filte=
r (as distinguished from a digital approximation) exhibits a response as so=
on as an input is applied. It may be small enough to go unnoticed, but it's=
 there. Was Rune splitting hairs?

Jerry

On Jan 24, 1:39&#4294967295;pm, Jerry Avins <j...@ieee.org> wrote:
> Dirk,
>
> I think it depends on how one construes "delay". A _real_ Butterworth filter (as distinguished from a digital approximation) exhibits a response as soon as an input is applied. It may be small enough to go unnoticed, but it's there. Was Rune splitting hairs?
>
> Jerry

Jerry,

If you put an impulse into the filter and look at the response you see
the most impulsive part of the output delayed. Based on the OPs
comments about a delay of 1/Wn it appears that is what he is referring
to. That is an effective 'delay' (granted, a function of freq), and
the OP should be getting it. When we talk about typical linear-phase
filters we don't talk about them having no delay because the first
output point is non-zero.

Dirk