[sdiy] group delay (of filters) and listening

[Czech Martin]

Though it is mathematical simpler, the sine modulated sine approach
is not so educational when viewed on a scope.
The I prefer a short sine burst, 5 or 6 cycles.

This shows the real delay in a nice fashion.

Even better: a gaussian bell shaped burst.
Have to try this at home!

m.c.

-----Original Message-----
From: owner-synth-diy at dropmix.xs4all.nl
[mailto:owner-synth-diy at dropmix.xs4all.nl]On Behalf Of Magnus Danielson
Sent: 07 October 2004 11:08
To: synth-diy at dropmix.xs4all.nl
Subject: Re: [sdiy] group delay (of filters) and listening


From: Don Tillman <don at till.com>
Subject: Re: [sdiy] group delay (of filters) and listening
Date: 07 Oct 2004 01:56:13 -0700
Message-ID: <m2k6u2q53m.fsf at till.com>

>    > Date: Tue, 5 Oct 2004 13:40:05 +0200
>    > From: "Czech Martin" <Martin.Czech at Micronas.com>
>    > 
>    > After all, it is not important how the group delay was achieved,
>    > but only it's shape.
> 
> Oh?  Are you sure?
> 
> I'll claim that group delay is not an actual physical phenomenon, but
> an interpretation of phase measurements.
> 
> (If you have a device that provides a small delay without any other
> side effects or distortions, if you measure the phase of the output,
> compared to the input, you'll see the phase changing with frequency,
> linearly.  There's no "phase distortion" going on, the resulting
> waveform looks identical to the original waveform, it's just delayed a
> little bit.  Group delay is an attempt to interpret phase measurements
> in this light, as a delay that that changes with frequency.)
> 
> But a measured phase value could be due to any number of mechanisms;
> from a direct signal, from an inverted polarity signal, from a delay
> line, from a simple filter circuit, from a multiple stage filter
> circuit that shifts the phase and delays the signal, or from any
> combination.
> 
> These will all carry different sonic artifacts.

There are two distinct forms of delay here, the phase delay and the
group delay. The phase delay is the "obvious" apparent time delay due to phase
shift of a signal. It's t = Phi/(2*pi*f). The group delay is maybe best
understood by letting us see what delay a sine which is AM-modulated with a
another sine experience. The center-band will experience the normal phase-delay
where as the side-bands will effectively perform a phase-delay slope estimate
best being descripbed as the derivate of the phase. That is, the group of
signals will experience that delay rather than the phase delay.

This difference is very real and the delay can actually work in different
directions. When we modulate information onto a carrier (light, microwave
frequency or whatever) then group delay is what we want to know about.

It's a bit confusing but there it is. I'm also not sure my attempt to describe
it is very good.

Cheers,
Magnus