Sonic Spotlight

[Sean Costello]

Hi all:

Anyone here know anything about the "Sonic Spotlight" at MIT? It seems
pretty cool - uses ultrasonic sound, which is more directional than
lower frequencies, to project sounds at spots a fair distance (up to 100
meters) away. Point it at a wall, and the sound seems to be eminating
from the spot you are pointing at. Seems pretty cool, but not much in
the way of details on how it is done. However, a few other postings on
the subject seem to point to how it is done:

Build a frequency shifter (the Electronotes design would be fine).
Instead of the quadrature VCO design used in the standard frequency
shifter, use a fixed frequency quadrature oscillator, with a frequency
of 200 kilohertz. The upshifted output of the frequency shifter is sent
to some ultrasonic...um...thing, some sort of super tweeter horn or
something. A second tweeter ultrasonic thing is used to transmit the
sine output of the VCO. Both tweeters are aimed at a specific point.
When the two signals combine, the nonlinearity of the shock waves of the
air...ok, I don't pretend to know the physics. What I do know is that
you should get sum and difference frequencies as the result. The sum
frequencies will be well above the range of human hearing, and the
difference frequencies should be the original signal. 

The tricky part in all of this would be the parabolic tweeter array, or
whatever is used to transmit the signals. Otherwise, it might work as
described. I can't really test the ultrasonic carrier frequency idea,
because my Csound frequency shifter is subject to the limitations of a
44,100 hertz sampling rate (which is why it would probably easier to
make such a device in the analog realm). However, I did a test today of
shifting a signal up by 44,100 hz, and the result was indistinguishable
from the original signal.

Later,

Sean Costello

P.S. Here is the original post, from Eric Scheirer at the MIT Media
Labs. All of the useful information, with none of the antiorp spam (who
IS Ric from the Young Ones).

> >I once read an article about some guy building a sound reproduction system
> >using arrays of tiny speakers which produce both a pure tone of 100 kHz and
> the
> >audio signal shifted over 100 kHz (such that it is located in between 100
> kHz
> >and 120 kHz, as in single sideband modulation). A listener is claimed to
> hear a
> >normal audio signal.
> 
> Yes, this technique is well-demonstrated and works fine.  We have
> a prototype running here (see AES Preprint #4853 from the 105th
> AES in San Francisco), although the basic technology has been known
> for some time.  The demodulation is not in the ear like in
> a combination tone, but in the air itself.  By focusing multiple
> ultrasonic transducers at the same spot, the air is driven into
> nonlinearity (shockwaves) that have audible distortion products.
> You preprocess by calculating the inverse distortion to get a
> desired signal from the air.
> 
> The major advantage of such a device is that it gives you audible
> sound with the directivity properties of the ultrasound carrier.
> Our prototype was built by Joe Pompei: http://sound.media.mit.edu/~pompei.
> It has about -25 dB rolloff about 3 degrees off-axis relative to
> the focal line.  You can easily hear it from over 100 meters away.
> Also, if you focus the array at a wall, the demodulation will
> be strongest at the reflection point, and thus the audible sound
> is generated at the wall, rather than from the speaker array.
> This isn't a psychoacoustic trick -- that's where the audible
> sound energy is actually coming from.
> 
> The major disadvantage right now is low-frequency response -- Joe's
> version doesn't make much sound below 300 Hz.  Previous devices
> have suffered from really bad harmonic distortion, but Joe has
> pretty much solved that.  The safety of the device is still under
> investigation -- it uses 120 KHz carrier at about 110 dB SPL.  OSHA
> guidelines claim that's safe, but there hasn't really been enough
> research to know.
> 
> Best,
> 
>  -- Eric