Can You Hear Me Now?Most loudspeakers are designed to throw sound as far and wide as possible without compromising audio quality. Parametric speakers are one of the exceptions. 10/25/2006 10:59 PM Eastern
Can You Hear Me Now?
Most loudspeakers are designed to throw sound as far and wide as possible without compromising audio quality. Parametric speakers are one of the exceptions.
You're standing in front of a mall kiosk that features a display of sunglasses. Suddenly, you're enveloped by sounds of waves crashing and seagulls crying, along with a gentle voice inviting you to imagine wearing them at the beach.
“Did you hear that?” you ask your spouse, who's just a few steps away. “Doesn't that sound like the real thing?” Instead, you get only a puzzled look. Why?
James Croft (left), chief technology officer, and Dr. Wensen Liu, senior research scientist at American Technology Corp., San Diego, prepare to test one of the company’s parametric loudspeakers in an anechoic chamber.
Most loudspeakers are designed to throw sound as far and wide as possible without compromising audio quality. Parametric speakers are one of the exceptions. With an effect akin to a ventriloquist throwing his voice, parametric speakers create a narrow beam of sound — so narrow, in fact, that a person standing out of the coverage pattern can't hear it.
Parametric speakers are a good fit for a variety of applications. One example is a museum exhibit, where a patron standing in front of a sculpture — and only in front of that sculpture —can listen to a description. Another application is a supermarket check-out line, where parametric speakers are paired with digital signage over each lane to deliver multimedia ads. Although conventional loudspeakers can be used instead, they can be annoying to others who may not wish to hear the message.
“They put in normal semi-directive loudspeakers, and the checkers get sick of hearing the messages all day long,” says James Croft, chief technology officer at American Technology Corp. (ATC), a San Diego-based manufacturer. “So they unplug them.”
The supermarket example highlights another advantage of parametric speakers. Unlike a museum, where audio needs to be discreet so it doesn't disturb the mood, a supermarket is a cacophony. In that environment, a focused beam of sound has a better chance of knifing through the din and capturing a consumer's attention. But even in a noisy environment, parametric speakers can be an attractive choice because they don't add to the din.
“A marketplace that values quiet in a very noisy gadget-filled world has helped create a solid market for Audio Spotlight products,” says Dr. F. Joseph Pompei, founder of Holosonics, a Watertown, MA-based vendor. “We have thousands installed in the world's top museums, retailers, corporations, and trade show exhibits.”
One way to understand how parametric speakers work is to think of them as the reverse of shotgun mics. Definitions vary, but most include the use of an ultrasonic carrier wave.
“A parametric loudspeaker is a transducer system that emits modulated or multiple ultrasonic frequencies that are driven to an intensity level such that the air medium becomes nonlinear and therefore produces audible difference tones,” Croft says. “By using ultrasonic frequencies that have very short wavelengths (on the order of 6 mm) delivered from an emitter that's many wavelengths in diameter, audio is created in the air along a directional column of ultrasonic energy.”
Some vendors offer highly directional speakers that provide an experience comparable to parametric speakers, but they don't consider them to be parametric because they don't use ultrasonic techniques. One example is Panphonics. “Ours are based on an electrostatic operating principle, making it possible to generate a highly directional audio beam without using ultrasonic carrier waves,” says Teemu Sivén, an audio engineer at the Espoo, Finland-based company. “A very thin diaphragm is tensioned between two stators. The diaphragm is charged with a bias voltage. When a high-voltage audio signal is fed to the stators, the diaphragm starts to vibrate as the electric forces between the bias-charged diaphragm and the stators push and pull the diaphragm, creating sound.”
One major difference between parametric speakers and their conventional counterparts is that the audio isn't created at the transducer. “Instead, it's created in the air, over and over again along the ultrasonic column until the ultrasonic energy dissipates,” Croft says. “Then the audible sound propagates onward in a highly directional form.”
The ultrasonic carrier is projected into the listening area, where it's demodulated using a non-linear environment, which can be the listener's ear, the air, or a special material such as a fibrous fabric.
If those explanations sound slightly puzzling, that's because the technology is. Even many veteran audio engineers scratch their heads when pondering exactly how parametric speakers pull off the feats that they do.
“There's considerable dispute in the technical community about how parametric loudspeakers really work,” says Ken Kantor, chief technology officer at Tymphany, a Cupertino, CA-based transducer manufacturer that doesn't offer parametric products. “For example, the amount of non-linearity in the air isn't mathematically sufficient to explain what's happening. If air was this non-linear, think about what a stereo pair of tweeters would sound like.”
Another puzzler is why the sound is created at its target rather than at another point along the beam. “The air is just as non-linear at the point the ultrasound first leaves the parametric speaker as it is 10 feet away,” Kantor says. “Also, why doesn't the audible sound spread out once it's finally created in the air?”
Some believe that the answers — and the beauty — could lie not in the speaker or the air, but rather in the ear of the beholder. “Some audio engineers, myself included, conclude that the non-linear demodulation doesn't happen in the air but, in fact, occurs due to high sound pressure level non-linearities in the listener's ear,” Kantor says. “This assumption answers all the questions above.”
A challenge made and met
The history of parametric speakers varies, depending on who's telling the story, but most generally trace the technology's roots back to the early 1960s. The initial research was for underwater applications rather than AV because the technology's pioneer — Peter Westervelt, a Brown University physics professor — argued that parametric speakers wouldn't work in the air.
In fact, Westervelt was adamant to the point that other researchers started to see it as a challenge. Some believe that the debate reached a pivotal point at a 1971 meeting in Washington, D.C., where Westervelt noted that no one had reported a successful parametric experiment using air rather than water. He then provided a litany of theoretical reasons why there would never be one. One attendee, Dr. David Blackstock of the University of Texas, had heard enough.
“When Peter sat down, very close to where we were sitting, I couldn't contain myself,” Blackstock recounted in an email to Joe Norris, son of ATC founder Elwood Norris, five years ago. “I had the audacity to say, ‘BS, Peter.' I've often wondered since then how I had the guts to say that to the grand old man of parametric arrays. Peter retorted, ‘All right, you prove it!'”
Three years later, Blackstock coauthored a paper summarizing experiments that proved parametric speakers worked in air. Research then more or less languished for nearly a decade until a series of papers by Japanese researchers tackled other fundamental questions, including the optimum carrier frequency, distortion correction, and insulating the listener from the ultrasound.
“They state that the carrier frequency needs to be as low as possible, but not so low as to cause beam spreading,” Croft says. “They recommend a frequency between 30 kHz and 70 kHz. In order to correct for distortion, they introduce the concept of square-rooting the modulation signal, which makes a lot of sense based on the Berktay far field solution, stating that the envelope is squared by the air. They quickly discover that this will require additional bandwidth.” (In 1965, British acoustician H.O. Berktay theorized that a demodulated secondary waveform exists along the axis of the beam and that it's proportional to the second time derivative of the square of the modulation envelope. As a result, the secondary wave is subject to intense harmonic distortion. For more information on this concept, see our web-exclusive sidebar on www.proavmagazine.com.)
Twelve years later, Elwood Norris began experimenting in his garage with ways to demodulate sound directly in the air, while his son pondered the problem at ATC. “Joe figured out that sending both frequencies through one transducer would ensure that the airborne signals would align and mix properly,” Croft says. “The effect would be maximized because both signals (actually, a single composite signal) travel down a common axis.”
Meanwhile, Pompei was also refining parametric technology at the Massachusetts Institute of Technology (MIT) before commercializing it under the Audio Spotlight brand. All of these projects by Pompei, ATC, and others eventually led to claims, counter-claims, and a few lawsuits.
“Very high distortion and cost caused the technology to be abandoned, and it wasn't revived until my own breakthroughs at MIT in the field in the mid 1990s that solved these problems,” Pompei says. “A few other companies have attempted to copy the Audio Spotlight technology from my publications and claim its invention. Some have even been sued for fraud regarding these false claims.”
Parametric speakers are designed to focus sound in a limited area, but some say that they're also limited in terms of fidelity. “In my experience, parametric loudspeakers have rather poor sound and a limited frequency range,” says Panphonics' Sivén.
But vendors such as Holosonics claim that their products are the exception to those perceptions. “The sound quality of the Audio Spotlight is quite comparable to a traditional loudspeaker, especially when combined with a subwoofer,” says Pompei, whose product has a range of around 200 Hz to about 18 kHz.
Good low-frequency response can be a must-have in some applications, such as those featuring music rather than just speech. One catch is that adding a subwoofer isn't always an option, such as in a supermarket check-out lane, where there often isn't room to hide the extra box. Another catch is that when looking for a parametric speaker with good low-frequency response, the manufacturer's specs can be misleading.
“Accurate measurement of parametric loudspeakers is rather difficult, as the intense ultrasonics impact the microphone output at audio frequencies, often fooling the test technician into thinking that they're getting more extended low-frequency response than what's actually heard by the listener,” says Croft, whose products have a range of 400 Hz to 16 kHz. “We've seen a number of AES papers delivered on parametrics showing false extended bass response.”
Indeed, in the late 1990s, ATC had its speakers tested by Lucent Technologies, whose tools showed that they went all the way down to 20 Hz. Those results prompted ATC to look for a more accurate way to measure performance.
“We use specially developed, proprietary measurement schemes to allow us to properly measure the frequency response accurately,” Croft says. “Beware of claims of extended response from parametric loudspeaker developers.”
Finding a home
Four decades after the technology was originally proposed, parametric speakers are still a niche play. One reason is that they require specialized expertise and components, so leveraging existing audio products wasn't really an option when the technology was first developed — nor is it today. “There aren't many active practitioners of this technology in the world,” Croft says. “You can make one work using off-the-shelf stuff, just as a proof of concept, but making it function properly is quite difficult. The concept was ahead of the hardware that was available.”
That's changing, which suggests that parametric speakers could become more common. “Back when the Japanese were doing it, it wasn't easy even to invent from what was available,” Croft says. “We at least have the advantage that switching power transistors and some of the signal-processing chips have increased in capability and come down in price enough to configure these in new, novel ways. But there's still nothing off the shelf that allows you to plug this together.”
On the plus side, installing parametric speakers is generally straightforward. However, AV pros still have to account for the loudspeakers' unique capabilities.
“Think of installations more like lighting design than sound design,” Pompei says. “Sound will shine, reflect, scatter, shadow, and beam just like light. You can combine beams and make rows of ‘audio spots,' much like track lighting. But in most cases, you simply mount the speaker disc, and aim it to where you want the sound to go.”
On the sales side, one potential hurdle is convincing the client that parametric speakers are a better solution than other types of directive speakers or even headphones, especially if conventional alternatives are cheaper. (Holosonics says that its systems are “priced in line with other professional quality loudspeaker systems,” but other vendors say that parametric speakers are significantly more expensive than conventional loudspeakers.)
“In many situations, the real competition for the parametric loudspeaker isn't a normal speaker, but a pair of headphones,” says Tymphany's Kantor. “Headphones require user activity and are subject to vandalism and weather. On the plus side, headphones are comparatively inexpensive, easy to wire up, and reduce ambient sounds, which a parametric loudspeaker does not.”
Even so, vendors say that parametric speakers are starting to build awareness among AV pros and their clients. But often the sale comes down to hearing-is-believing. “We're seeing a lot of demand worldwide for a lot of different things,” Croft says. “Digital signage is the big play at the moment. I've shown this technology to a lot of seasoned audio engineers who were very skeptical. But once you play it for them, their jaws drop over how directive it is.”
Tim Kridel is a freelance writer and analyst who covers telecom and technology. He's based in Kansas City and can be reached at firstname.lastname@example.org.