On Tue, 10 Mar 2015 18:51:46 GMT, (Don Pearce) wrote:

On Tue, 10 Mar 2015 18:29:28 +0000, Sumatriptan
wrote:

On 10/03/2015 14:38, Don Pearce wrote:

I think a likely explanation is the simple non-linearity of the ear
causing audible intermod products when the ultrasonics were on.


There's a simple test that could confirm that idea.

Two (or more) high level ultrasonic tones a few hundred Hz apart should
result in audible tones. Surely this has been tried.

I know that this sort of two-tone intermod test is used in audio systems:

http://www.whathifi.com/forum/comput...ms-performance

There is a download file for this test at the above location.

For testing human ear linearity, the levels would have to be high and
the tones generated separately in independent systems to eliminate audio
system intermod from the test results. I would think the required
equipment would be available in many audiophile premises. Any takers? ;-)


Actually, 30 and 33kHz are going to generate products at 27 and 36kHZ
- not much help.

I made a file with 20kHz and 38kHz. The 2kHz product was clearly
audible, and very level dependent. There was actually some kind of
threshold effect. I prevented distortion in the amplifier being a
confounding factor by recording them on separate channels.

d

And attempting to move it all down as far as possible to make life
easier on the tweeters, I find I can still hear 16kHz! I thought those
days were long gone - one happy bunny here.

d