"Arny Krueger" wrote in message ...
(...)
ie. Can you be bombarded with large SPLs of pure sinewaves at 20kHz
and 24kHz and perceive 4kHz (F2-F1) and 16kHz (2F1-F2).

You've got to pick frequencies where the ear actually responds...

I'm not sure that it's been established that the tones must be
~20KHz, i.e., I don't know whether distortion products can
conceivably occur if one (or more) of the tones are 20KHz. I've not
found any research which looks at that (though it would be pretty
simple to run a quick practical test to find this out),but I've not
searched exhaustively.

To obtain (some of) these distortion products does not require large
SPLs. In fact, these ear-based nonlinearities have been used to carry
tunes in (classical) music.

Although these tones are a function of the way the ear initially
amplifies tones, there's recent evidence that these products may be
influenced by attention, i.e. 'higher order' processes can alter the
ear's behaviour (doi: dx.doi.org/10.1017/S0048577201990109).
Interesting.

There's a nice review of this stuff he
http://physrev.physiology.org/cgi/reprint/81/3/1305 (Robles & Rugger
01, Physiol Rev 81: 1305-1352.

Steve.