"Don Pearce" wrote in message

On Fri, 6 Aug 2004 07:30:49 -0400, "Arny Krueger"
wrote:

"Don Pearce" wrote in message


Well, speakers generally are nonlinear, so what you are seeing here
is intermod.

Really?

Doppler distortion in speakers is supposedly a "built-in"
effect - nothing to do with non-linearity - that is caused by the
same cone reproducing two frequencies simultaneously.

Agreed.

The argument goes
that if a speaker is reproducing a 1kHz tone, but is simultaneously
moving back and forth at 50Hz, the 1kHz tone must be modulated by
the Doppler effect. Of course, if you do the maths of
superposition, this doesn't happen - the tones coexist perfectly
without any doppler.

Tell that to the AES! ;-)

A great deal of BS has emanated from that organ!

So this is simple, straight-forward intermodulation between the two
tones.

Two reasons I think this really is predominantly FM:

(1) The sideband structure looks a lot more like FM than AM, per the
simulations I added to http://www.pcavtech.com/techtalk/doppler/ .

(2) I redid the experiment using high frequency tones at 1 KHz and 4
KHz.

All other things being equal, AM is frequency-independent. FM is
frequency-dependent. Since the stimulus for the IM is the 50 Hz
tone, the stimulus for 50 Hz sidebands for both the 1 KHz tone and
the 4 KHz tone is the same.

I did a simulation in Audition of pure FM, and the sidebands on the
1 KHz tone were about 12 dB lower than the ones on the 4 KHz tone,
which exactly follows this theory.

However, speakers don't have just one kind of distortion.

I have added the results of triple tone test results to
http://www.pcavtech.com/techtalk/doppler/ .

Comparing the amplitudes of the first two sidebands around 1 KHz and
4 KHz, I find that there is an approximate difference of 6 dB. The
sidebands around the 4 KHz average about 6 dB higher than those
around the 1 KHz tone. If this was pure FM distortion, I would
expect a 12 dB difference. I conclude that there is a mixture of AM
and FM.

The only way to verify this is to look at the phase as well as the
amplitude of the sidebands.

That's one way, but it's a very hard row for me to hoe.

So, you decline to believe that the relative amplitudes of the sidebands are
different and relevant, as the frequency has increased?