SB AWE32 Developer's Information Pack
PART V 3D Positional Audio API
••
95
Copyright
Creative Technology Ltd., 1994-1996
Version 3.00
The next best cue is making the determination whether a sound is on the median plane or not. Substantial
research shows that we are also good at this distinction.
Beyond these two factors, people are actually quite weak at determining spatial location, without head
movement. This may seem quite suprising, but research subjects who have their head immobilized in
research test conditions consistently identify sounds which are actually coming from in front of them as
coming from behind them. Some speculate that this so called front-back confusion may have an
evolutionary advantage, since unseen sounds might be predators! Elevation cues are also quite fragile.
Finally, another of the most important cues for a sound's location is vision! If we see an object that we
believe to be making a sound, then we think the sound is coming from that object. This phenomenon has
been exploited for years by ventriloquists. This phenomenon can be exploited just as well by games
developers in that on-screen visual cues can also help establish the apparent location of a sound source.
In contrast, an off-screen sound source can help to direct the viewer's attention to the (unseen) apparent
source of that sound. The most recent generation of games adopts a first person perspective which is
ideally suited for spatial audio, in that the user can direct the view of the game in different directions. If
robust cues are employed to attract the viewer's attention, the viewer can be cued to direct the view
towards the unseen sound. The strategy of directing the view from audio cues will only be as strong as
the audio cues, however.
Head Tracking
If a 3D audio spatialization system is producing audio for headphones and the listener is using a head
tracking system, so that the position and orientation of the listener's head is known in read time, the 3D
audio system can provide the correct left-right cues to simulate the actual behavior that physical sounds
have in physical space. Head tracking is widely known to be able to significantly enhance spatial audio
cues dramatically.
People are not likely to start wearing head trackers and headphones in large numbers any time soon. For
one thing, head tracking hardware is quite expensive. Secondly, it is cumbersome and isolates a person
from the room environment.
Instead, people mostly listen to multimedia audio over speakers. Speaker playback of spatial audio is
still quite practical, but there will always be limitations. The developer should not expect robust front-
back and elevation cues without head tracking headphone audio displays.
However, distance cues and lateral cues should be very robust. Interestingly, distance cues are quite
practical even with only one speaker!
AWE-32 Implementation of 3D Audio Cues
The AWE-32 implementation of Parametric 3D Audio provides strong lateralization and distance cues.
The lateralization cues have been found to be quite robust either on speakers or on headphones. They do
not seem to suffer a lot from a small "sweet spot" either. The sweet spot is the region midway between
two stereo loudspeakers where the stereo effect is strongest. While the optimum AWE-32 spatial effect
is still in the center between loudspeakers, listeners can be almost anywhere between a pair of
loudspeakers and the lateralization effect will still be quite clear. Many existing 3D audio techniques
have a very sensitive sweet spot, and are severely degraded by listening off-center.
Distance cues will work even on a single speaker.
Dynamic Versus Static Cues
Dynamic spatial cues are far more robust in both the real world and in synthetic spatialization. If a
sound emitter (or a receiver) is moving, then the change in spatial location will create a much stronger
impression than a sound that is fixed, relative to a receiver.