Arny Krueger wrote:
"Scott Dorsey" wrote in message

Well, surprisingly enough, Phil actually made the good point that the
woofer position does not directly follow the input signal, but that
the excursion at lower frequencies is exaggerated. This is indeed
the reason that we get Doppler distortion.

I would say that the exaggerated excursion at low frequencies is a
contributing cause for Doppler distortion, but not the only cause.

On a typical full-range speaker _not_ breaking up, what other good
causes are there? With coaxials and with speakers in breakup, there
are all kinds of wacky things going on.

The only other cause I can think of has to do with compressibility of
air and it would seem to be a comparatively small issue.

But, how do we compensate for this? And can we, even?

Of course, reducing the bandwidth to each driver and reducing the
driver excursion as much as possible are crude ways around the problem.

Crude but effective! ;-)

Doppler is exactly proportional to the upper frequency being modulated. Drop
the upper crossover frequency on that woofer by a factor of two, and you
drop the Doppler distortion by 2.

That's reducing the bandwidth.

Double the diaphragm area, and you get the
same benefit.

That's reducing the driver excursion.

Subwoofers make even more sense!

Yes, but they bring another whole set of issues along with them.

A more exaggerated example of the distortion, though, is found in
coaxial speakers where the moving woofer cone is used as the horn for
the tweeter.

Yes, and I even have a KEF Q-15 to test that with.

A better one would be one of the Radian drivers, which are really bad about
it.

It would be interesting to see if the Urei horn assemblies on the Altec
coaxial drivers really do minimize doppler modulation compared with the
original Altec horn assemblies. That was one of the arguments the Urei
guys used for the extended horns they employed.

However, our preliminary results show that even with a reasonable worst case
(small woofer, relatively high upper frequency) the Doppler tends to get
lost in the AM distortion. Claiming it isn't there is wrong, but getting
worked up about it seems a little foolish.

Yes, but how audible is it? You can treat the doppler modulation sort of
like spurious sidebands, BUT they are sidebands that are modulated by
the signal. Does it make it mode or less audible than a fixed sideband?

Here, though, I am not sure the math model is quite so
easy, and it would be interesting to see if anyone can model the
boundary effects near the moving woofer cone.

It's tough enough to work with the case we're working with, which seems to
be far simpler.

Right, because you have pretty much one dominant distortion source, and it
is an easy one to model. You should be able to plug and chug and get a
simple value for doppler distortion due to increased excursion at low
frequencies, knowing little more than the tone frequencies and the driver
excursion for a given cabinet. How does that compare with the measured
doppler modulation on that cabinet? That will tell you if there are any
other hidden effects to worry about.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."