to calculate time delay between two signals

robert bristow-johnson <rbj@audioimagination.com> wrote:

(snip)
> i guess not.  i understand the multipath problem, i was just assuming
> that the direct path (from the single signal source to the two
> microphones) is shorter and louder than the reflected paths.

I was just remembering, how we learn in 3rd grade science class
(or maybe 4th grade) that light goes in straight lines, and
sound doesn't.  They can't teach diffraction at that point, though.

You can easily be in a place where the direct path involves lots
of diffraction and absorption, where another path reflects off
a large object, such as a building.  That could be true for audio
and, for people in big cities, for radio signals.  

-- glen

(snip, someone wrote)
>> and their correlation function is a mess. That happens due 
>> to multipath and reverberations,

Reminds me of our library book sale, which is in a really large
building (previously owned by the Navy, but now the city).

They have a PA system and sometimes make announcements, talking
really slow.  There are a few speakers around, but also large brick
walls which make good reflections, and you really can't hear well
at all.

I was wondering at the last sale, if you have the appropriate DSP
equipment, and some number of amplifiers and speakers, what would
be the best way to make a system that is most audible in the room?

I was wondering about enough speakers and delay modules to try
to cancel out some of the reflections.  It might be that sound
absorbing pads are cheaper, but maybe not.  

-- glen

On Oct 28, 1:17&#4294967295;pm, robert bristow-johnson <r...@audioimagination.com>
wrote:
> On Oct 28, 1:16&#4294967295;pm, maury <maury...@core.com> wrote:
>
> > On Oct 28, 10:43&#4294967295;am, robert bristow-johnson
>
> > > i don't see how AMDF will work at all. &#4294967295;AMDF is more related to *auto*-
> > > correlation (operating on a single input signal), not the cross-
> > > correlation between two inputs.
>
> > Robert, it does work for estimating delay between a speech signal and
> > its delayed form, not two independent signals.
>
> so what if the delayed signal has a significant amount of attenuation
> besides the delay? &#4294967295;how well does AMDF work for that?
>
> r b-j

It worked fine. You still get a minimum at the point of (envelope)
correlation.

The AMDF isn't usually used this way, but when you're trying to "cram"
a bunch of stuff in a 44-pin quad ASIC, the reduced complexity of the
AMDF was needed.

On Oct 28, 10:52�am, robert bristow-johnson
<r...@audioimagination.com> wrote:
> On Oct 28, 1:06&#4294967295;pm, dbd <d...@ieee.org> wrote:
> > So Hardy is right that cross correlation gives you the delay (lots of
> > them, from many effects) and Maurice is right that further processing
> > is required to produce useful results.
>
> well, i didn't see what maury was saying about further processing.

I took "further processing" than cross correlation to be additional
processing not necessarily just later processing. Use of the start of
the envelope is one I have used. It requires that you have some method
of identifying signal present as different from background level so
that you can select the "front end" of the envelope to avoid multipath
and reverberation. Towards this end, I've used time domain processing
to identify  wideband signals in parallel with frequency  domain
processing to identify narrowband signals. There can be a lot of
"further processing" so the system is far more than a simple cross
correlator. Even the conversion of time delay to direction of arrival
can involve further processing to avoid needing to measure propagation
velocity directly.

Dale B. Dalrymple

On Fri, 28 Oct 2011 18:34:49 +0000 (UTC), glen herrmannsfeldt
<gah@ugcs.caltech.edu> wrote:

>robert bristow-johnson <rbj@audioimagination.com> wrote:
>
>(snip)
>> i guess not.  i understand the multipath problem, i was just assuming
>> that the direct path (from the single signal source to the two
>> microphones) is shorter and louder than the reflected paths.
>
>I was just remembering, how we learn in 3rd grade science class
>(or maybe 4th grade) that light goes in straight lines, and
>sound doesn't.  They can't teach diffraction at that point, though.
>
>You can easily be in a place where the direct path involves lots
>of diffraction and absorption, where another path reflects off
>a large object, such as a building.  That could be true for audio
>and, for people in big cities, for radio signals.  
>
>-- glen

In EM propagation we call that the "urban canyon" environment.  It's
one of the most aggressive multipath cases, especially when you're
mobile.


Eric Jacobsen
Anchor Hill Communications
www.anchorhill.com