cross-discipline job descriptions

ot

What?!

You blame *me* as an excuse to get sloppy with the language? See you at dawn (Central European Time)! Swords or pistols?

Rune

ot

Huelsman and Allan, Introduction to the theory of and design of active filters, page 92:

D(w) =3D - d/dw arg n(jw)

"Considering the above we see that the maximally linear (at the origin) phase characteristic of Thompson filters produces a delay characteristic which is maximally flat (at the origin). Thus these filters are frequently referred to as maximally flat delay (MFD) filters."

You can't have maximally flat phase. In these all pole filters with zeroes at infinity, the phase is constantly accumulating. Flat phase means no delay!!!!

Flamethrowers! ;-)

Neutron emitters.

he

.Not

I didn't blame you for anything. I accused you of excusing sloppiness.

Jerry

...

I don't read anything there about "maximally flat linear phase"

An ideal Hilbert transformer exhibits quadrature phase shift at all frequencies, along with a very long delay. "Linear phase" implies phase shift proportional to frequency i.e., delay independent of frequency.

Jerry

How about a symmetrical FIR filter?

Cheers

Phil Hobbs

(centred on t=0)

Cheers

Phil Hobbs

You don't see "Maximally flat phase: phase their either, right?

No idea why you brought up the Hilbert transformer, but an ideal Hilbert transformer has no delay.

...

because you said "You can't have maximally flat phase." and a zero- delay Hilbert has a maximally *flat* (as in "level") delay.

depends on what you mean by "ideal". (remember semantics? like what the meaning of "ringing" is?)

some people might consider a filter of some class "ideal" if follows the ideal characteristics of that filter, but with a finite delay. trouble is, you can get a nice, phase-linear Hilbert transformer with a finite delay, and it can be perfectly 90 degrees (plus the linear- phase component) but the magnitude won't be perfectly flat. nonetheless, the phase is maximally flat (w.r.t. the constant delay) because it is perfectly flat.

r b-j

On Dec 17, 10:26=A0pm, robert bristow-johnson wrote: ...

meant to say a zero-delay Hilbert transformer has a maximally-flat

r b-j

An ideal buffer has maximally flat phase, but it isn't much of a filter. Seriously, you never admit you are wrong, even when, shall we say, you are technically bitch slapped with the truth.

The group delay of a Bessel is maximally flat. Let's leave it at that.