Can You Believe Your Ears?Larry Armstrong
The search for realistic, concert-hall-quality sound has led audio equipment makers down some strange paths. Remember the quadraphonic craze of the early 1970s? Instead of two speakers, you needed four--and at least one manufacturer suggested a fifth, mounted on the ceiling. Or QSound, whose 1980s promoters proved more successful at selling stock than licenses for its three-dimensional sound?
Now, a new audio wave is gathering--and this time, the soundmeisters insist they've got it right. After a string of false starts, a half-dozen companies are trying to create the illusion of 3-D sound from just two speakers. What's driving them is 1990s-style multimedia, whose richly interactive games and educational programs need better sound than today's audio-equipped personal computers can deliver. "This technology fills that void," says Mal D. Ransom, marketing vice-president at Packard Bell Electronics Inc. in Westlake Village, Calif., the top seller of home PCs.
Since February, Packard Bell has been building 3-D sound into every multimedia PC it ships, using technology licensed from SRS Labs Inc. in Irvine, Calif. SRS--for sound retrieval system--first popped up in 1989 on top-of-the-line big-screen TVs from RCA and Sony. Hughes Aircraft Co., which invented the technology to enhance in-flight movies, sold it off in 1993 to the unit's managers. Last year, they sold 60% to Thomas C.K. Yuen, a founder of computer maker AST Research Inc. Packard Bell owns another 8%.
Other sound companies have jumped on the multimedia bandwagon, too. A humbled and restructured QSound Labs Inc. licensed its stereo-enhancement technology to Creative Labs Inc., the country's largest maker of PC sound cards, as well as to Capcom USA Inc., best known for its Street Fighter video games. A third company, Spatializer Audio Laboratories Inc., has deals with game maker Interplay Productions and chipmaker ESS Technology, and is planning to package its 3-D audio chip as an accessory for home PCs and stereo systems. Consumer electronics giant Matsushita Electric Industrial Co. designed Spatializer's mathematical formulas, or algorithms, into chips for its boom boxes.
All the 3-D sound companies use similar psychoacoustic techniques to spin their magic. The effect may be anything from an engulfing symphony of rain-forest sounds to the clash ef rock rhythms bouncing from the rear walls of a concert hall. Some companies, including QSound and Spatializer, can achieve that effect by encoding the sound in a studio and inserting delays in the left or right track. That lets them simulate differences in the time it takes for sound waves on one side of the listener to reach the opposite ear. Others, such as SRS, amplify only the frequencies that the brain uses to locate a sound. That makes all the difference when you're watching a movie in your home theater, says Sony Electronics Inc. Senior Vice-President James M. Palumbo. "When the telephone rings in the movie, you wheel around and reach for it."
SOUND ARC. That's a far cry from conventional stereo. While stereo arrays different instruments or voices on a flat plane between the two speakers, 3-D audio describes an expanded field of sound--a broad arc starting at the left ear and curving around to the right. In a video game, for example, that means the player could hear an enemy jet fighter approach, take a hit, and explode somewhere off his or her left shoulder.
To explain how SRS acoustics work, inventor Arnold I. Klayman suggests a simple experiment: Hold out your hand directly in front of you, and snap your fingers. Your outer ear will deflect much of the sound away from the ear canal, and both ears will hear the same sound. But as you move your hand around to the right, the snap grows louder in the right ear. The pitch also changes, because the outer ear emphasizes some frequencies and deflects others. For example, the ear may hear some treble sounds better when they come from the side.
To mimic that effect, SRS dissects the stereo signal, subtracting the left channel from the right, and boosts the resulting signal to match the frequency curves that tell the ear where sounds are coming from. It does the same on the left, and combines the processed parts back with the original stereo signal. The result: Ambient sounds on the right seem to come directly from the side, rather than from the right speaker.
Spatializer uses the same ambient signal information--the differences between the left and right channels--but processes it differently. The ear can't tell where bass sounds originate, and treble wavelengths are too short to be affected by the outer ear. So Spatializer amplifies the midrange portions of the signal. Then it adds short time delays, different for each frequency, to partially cancel the sound the opposing ear hears. That tricks the brain into thinking the sound hits one ear before the other, creating a spaciousness extending beyond the stereo speakers.
GRITTIER GAMES. QSound takes an entirely different approach. It adds effects during production, splitting monaural signals for each sound into left and right components to place the sound where it wants it. The company also adds time delays, to bring sounds forward and even behind the listener. QSound's technology produces the most striking sound effects of the three, but--like stereo--the illusion is limited to a small "sweet spot" centered between the speakers.
That limitation led to QSound's earlier audio debacle. But it's not a problem in the multimedia world, since computer users generally sit in one position: right in front of the PC. Says Perry Cook, technical director at the Center for Computer Research in Music and Acoustics at Stanford University: "Multimedia almost guarantees that the listener's head is in the right place."
A host of companies is trying to duplicate QSound's encoding techniques. Its highest-profile customer--Creative Labs, which makes SoundBlaster sound cards--has its own encoding scheme to make computer game sounds more realistic. And startup Crystal River Engineering, in Palo Alto, Calif., is trying to render 3-D sound with standard signal-processing chips.
Who will win the multimedia sweepstakes--systems that encode the sound for a 3-D effect, or those that derive it from the existing audio signal? Computer industry analysts say the intended multimedia applications are still so new it's impossible to tell. Either way, PC users will soon get sound matching the wildest new PC graphics. And stereo will never sound the same.