Subject: Re: conjugate poles/zeros for the laplace and zi
From: Martin O'Leary (oleary@cadence.com)
Date: Wed Feb 28 2001 - 09:51:23 PST
I think that it is reasonably that they don't have to match as closely
as IEEE floating point precision.
If they match within a specific percentage like 0.1% or 1ppm that should
be close enough,
--Martin
On Feb 27, 5:10pm, Ian Wilson wrote:
> Subject: RE: conjugate poles/zeros for the laplace and zi
> >
> > By the way, I don't think having conjugates automagically completed is
> > a good idea. If two roots are almost conjugates, we get in the game of
> > guessing whether they really are conjugates or are each half of a pair
> > to be completed. It is much clearer to error off and let the user be
> > explicit.
> >
>
> I think we would all agree with this. The problem with the current scheme
> (and what I believe Martin is looking for resolution on) is how to tell when:
>
> (a0 + jb0), (a1 + jb1) ...... (1)
>
> are intended to be conjugates. (Under the current scheme it's an error
> if the conjugate isn't found; under an automagical completion scheme this
> would specify 2 distinct roots).
>
> My opinion for what it's worth is that the answer to (1) above is that
> the the two form a conjugate pair when a0-a1 and b0+b1 are both zero to
> IEEE floating point precision (i.e. the presence of a root and its
> conjugate is a synbolic indication, not a numerical one).
>
> If we are looking for a better syntax in the future, I'd like one that:
>
> - specifies whether a root is real or complex;
> - specifies whether a root is degenerate and if so, its degree
>
> --ian
>
>
> --ian
>
>-- End of excerpt from Ian Wilson
This archive was generated by hypermail 2b28 : Fri Mar 02 2001 - 15:43:41 PST