Here are some binaural recordings I made, as well as links to other webpages
having 3D sound so you can judge the various technologies yourself. The
true test is in the hearing! Normally, we can spatially hear sounds, be
they left, right, up, down, front, back, near or far. Binaural recordings
attempt to capture all the information necessary for this effect. When done
successfully, one experiences outstanding sonic clarity with the auditory
image perceived externally out in three dimensional space, just as if you
were there in real life.
One way to make binaural recordings, is to install a pair of microphones
into a dummy head modeled after a real human head, one located in each ear.
Notable examples include: Fritz,
KEMAR
, Aachen -
Head, B&K
4128, and Cortex.
The more exact the replica, the more realistic is the binaural recording.
Or you can use your own head and insert probe
or miniature microphones
into your own ear canals. Different pinnae sound different, and indeed certain
ear shapes are better in capturing 3D sound than others. So, who has
the golden ears? Just like imperfections in the lens of the eye can lead
to distortions in spatial vision, similarly, imperfections in the shape
of the external ear can lead to ambiguities in spatial hearing. Of course,
your retina and inner ear must be in good working order too, for accurate
spatial perception. Test your high
frequency limit here.
Synthesizing 3D sound digitally is of great interest to the Virtual Reality
industry. Basically, this is how it's done. A computer model representing
the acoustics of the human head and pinnae can be measured and is called
the head related
impulse response. Digitally convolving a sound with this response synthetically
creates binaural sound. Again, the closer the computer model is to your
own head acoustics, the more realistic is the binaural sound. Actually,
full realism and clarity requires the modelling of the complete environment,
such as room reflections or the acoustics of the concert hall, and even
modelling the directional characteristics of each sound source. Computationally,
very intensive! You can create your own binaural sounds from monaural sounds
using Tom Erbe'sSoundHack
program for the Mac, which uses Durand Begault's HRTF measurements for NASA's
ASAD. I just made
this helicopter.mp2/0.2M fly around using Phil
Brown's monaural helicopter
sound.
To experience the 3D effects, you must replay binaural recordings over headphones/earphones.
This is to insure the left ear gets only the left ear signal, and the right
ear gets only the right ear signal. When binaural recordings are played
over loudspeakers, there is the problem of cross talk. When the left loudspeaker
plays a sound, the listener hears this in both ears, and similarly for the
right loudspeaker, resulting in garbled 3D. Afterall, if a sound source
is moving front to back, or up and down, you are not going to hear these
effects via a conventional stereo sound system, which is one dimensional
(left-right). One way to make binaural recordings playable over loudspeakers
is to preprocess the signals via a transaural cross-talk
cancellation scheme: One of the most amazing of these systems is the
new stereo-dipole
being developed at the Institute
of Sound and Vibration Research and Tokyo
Denki University! I am currently writing a transaural
3D holographic sound page to complement this one.
If your browser has Apple QuickTime
(3.0 preferred, but 2.1 or greater will work), it can be configured to automatically
play the .mp2 formatted sound files. Or you can download them and use one
of the listed players. For two comprehensive listings of web pages, see
either the 3D
Audio Page or the Ultimate
Spatial Audio Index.
*This is not binaural, but blind persons use ripple pitch to echolocate
distance. Ripple pitch can also be clearly heard on the sparkler
recording. You can hear a rising pitch as the sparkler gets closer to the
ground. The effect is otherwise known as comb-filtering, or phasing/flanging,
as when an jet plane takes off.
