Machine Listening Group Digital Life Consortium A beam of light can be controlled in many ways - it can be aimed at one person in a crowd, spread to fill a room, or projected to create rich, distant imagery. We can now do these very same things with sound.

 
The Audio Spotlight^TM, invented and developed at the MIT Media Lab, is a device which uses subtle nonlinear properties of the air to create an extremely narrow Sound Beam^TM.  This beam of sound behaves just like a beam of light - 'shining' it at a specific listener allows only that person to hear it, and projecting it against a surface creates an acoustic 'image' at the point of reflection.  It is the first device that provides total control over both the location and distribution of high quality sound, something impossible to achieve with traditional loudspeakers.

The circular transducer is very thin, and can be constructed in a variety of sizes and configurations as needed.  A typical Audio Spotlight transducer has an active area of approximately 1 foot diameter, and, depending on size and frequency content, projects an approximately three-degree wide beam of sound audible to well over 100 meters.  Harmonic distortion has been reduced to close to that of a traditional loudspeaker, sound level is quite appreciable (on the order of 80-90dBA) at several meters, and frequency response, depending on size, extends down to a few hundred Hertz, and upwards beyond the range of hearing. Continued research is being conducted on all facets of the technology.

While still under development, we are testing applications of the device in collaboration with several of our Media Lab Sponsors in preparation for eventual commercial release.
 

Put sound wherever you want it.^TM

Photo (c) 1998 Web Chappell

F. Joseph Pompei  pompei@media.mit.edu

 

The Audio Spotlight can be used in two major ways:  As directed audio, sound is directed at a specific listener or area, to provide a private or area specific listening space.  As projected audio, sound is projected against a distant object, creating an audio image.  This audio image is literally a projected loudspeaker - sound appears to come directly from the projection, just like light.

The Audio Spotlight consists of a thin, circular transducer array and a specially designed signal processor and amplifier.  The transducer is about half an inch thick, nonmagnetic, and lightweight.  The signal processor and amplifier are integrated into a unit about the same size as a traditional audio amplifier, and has similar power requirements.

Because it is impossible to generate extremely narrow beams of audible sound without extremely large loudspeaker arrays, we instead generate the sound indirectly, using the nonlinearity of the air to convert a narrow beam of ultrasound into a highly directive, audible beam of sound.

The device transmits a narrow beam of ultrasound (blue), which, due to the inherent nonlinearity of the air itself, distorts (changes shape) very slightly as it travels. This distortion creates, along with new ultrasonic frequencies, audible artifacts (green) which can be mathematically predicted, and therefore controlled. By constructing the proper ultrasonic beam, this nonlinearity can be used to create, within the beam itself, an audible sound beam containing any sound desired. This is presently done in real-time using low cost circuitry, a specially designed amplifier, and transducers developed at MIT specifically for this project.

Hyperdirectivity

The directivity, or narrowness, of an acoustic wave generated by a circular transducer is proportional to the ratio of the diameter of the transducer to the wavelength of the sound. So a transducer much larger than the wavelength of the sound creates a very narrow beam.

Audible sound contains wavelengths reaching lengths of several feet, so a reasonably sized loudspeaker will always produce a very wide, non-directional source at lower frequencies. The Audio Spotlight, in contrast, outputs short, millimeter sized ultrasonic waves, which form a very narrow beam even in a small transducer, which in turn generates audible sound. The nature of the nonlinear transformation also essentially eliminates sidelobes in the resulting beam, and maintains relatively uniform directivity across the entire audible frequency range.

The figure to the right compares the directivity of the Audio Spotlight (yellow) to that of an ordinary loudspeaker (purple).at 400 Hz. Note that the directivity of the Audio Spotlight is only three degrees, compared to the essentially omnidirectional directivity of the loudspeaker.

In order to obtain such narrow directivity from a traditional loudspeaker system, one would need a loudspeaker array fifty meters across!

A loudspeaker is like a light bulb, but the Audio Spotlight is like a laser.

History

The use of nonlinear interaction of high frequency sound to generate directive low frequency sound sources has been a well researched subject in the field of underwater acoustics since the early 1960's. Often misattributed to so-called "Tartini Tones", the effect is more accurately described as a parametric array, a term introduced by Westervelt [1].  In the past several decades, many underwater sonar researchers have used the effect to both generate directive low frequency sonar beams, detect underwater sound (parametric receiving array), and extend the bandwidth of underwater transducers.

The first published demonstration of an airborne parametric array was in 1975 by Bennett and Blackstock [2]. Rather than using inaudible ultrasound, they instead used very intense, high frequency audible sound to produce simple difference tones. While their goal was not a practical audio reproduction device, they nonetheless effectively demonstrated that the parametric array would work in air in addition to underwater.

In the early 1980's, several Japanese companies, such as Nippon Columbia, Ricoh, and Matsushita, attempted to develop the parametric array for the reproduction of broadband audible sound. They typically deployed large arrays containing hundreds of piezoelectric transducers, such as the one to the right [3], to transmit simple AM modulated audible signals. While successful in reproducing sound, tremendous problems with cost, robustness, and extremely poor sound quality (up to 50% total harmonic distortion) caused them to abandon the technology as unfeasible.

More recently in mid 1996, an American company produced their own version of this device and proclaimed it 'a revolution' in audio. In fact, this device, contrary to their claims and unbeknownst to the popular press, was very similar to those described in audio journals a decade earlier (shown to the left), and of course suffered from the very same problems of poor sound quality and lack of robustness that plagued the earlier researchers [4].  Since then, there has been no published evidence of progress towards a practical device.

Background

Since his days as a part-time musician and young acoustics engineer at Bose in the early 1990's, Mr. Pompei recognized that a key ingredient missing from audio reproduction was the ability to reliably spatialize sound. While in a natural environment, sound occurs all around us, giving us a tremendously strong impression of our environment, the reproduction of sound over loudspeakers, at best, provides a very vague and limited spatial impression.  Similarly, what was missing from music, he decided, was the ability to choreograph musical instruments in space, just as you would dancers.

While pursuing as a Master's student techniques related to '3D Audio' technologies, he realized that this method would simply not work in an uncontrolled acoustic environment - if the listener moved out of the small 'sweet spot', the illusion would vanish, and there were no practical remedies to this problem, so long as traditional loudspeakers were used.  The solution, then, was to not rely on psychoacoustic illusions, but instead to create sound independently of the loudspeaker.  One of several ideas he had at the time was the use of interacting ultrasound beams to produce audible sound.

After briefly researching the idea, he discovered the numerous papers describing the underwater parametric array and the earlier attempts of its application as an audible sound source.  From these papers, he saw that there were two key concepts which were overlooked in the previous attempts, mitigating their success:

• Preprocessing


Earlier attempts used simple AM modulation to generate the ultrasound signal, which does create audible byproducts, but also substantial distortion.  The nonlinear transformation from ultrasound to audible sound is much more complex than AM demodulation. Therefore, in order to reduce distortion, this specific transformation needed to be mathematically modeled, inverted, and then applied as a preprocessing algorithm.   The lowest-order preprocessing method, used in the earliest MIT prototypes, was derived from a simple model [5] proposed in 1965.
 

• Transducer Design


The transducers used in previous attempts were common piezoelectric transducers used for ultrasonic ranging.  These transducers are highly resonant, and do not have sufficient bandwidth to reliably reproduce the preprocessed ultrasonic signal.  Thus, even with a preprocessing algorithm, substantial distortion would continue to result until we developed transducers capable of reliably reproducting the broadband preprocessed signal.

As a side project during his Master's work, he continued his development of these ideas, studying nonlinear wave interactions and ultrasonic transducer design, eventually deciding to pursue the area as the focus of a doctoral dissertation.  Of all the universities that he applied to, he decided that the free-wheeling nature of the MIT Media Lab was the ideal environment for developing the idea.

The first full size prototype was demonstrated in April 1998 to our Media Lab Sponsors, and performed beyond all expectations. The first demonstration was a John Coltrane solo, whose saxophone was heard loud and clear, projected like a spotlight all around a movie theater, and flying right over the audience. Power consumption was nominal (<30W), construction was straightforward, and distortion had been reduced by several orders of magnitude compared to all earlier attempts.

A paper [6] describing the results of the first prototype, as well as a live demonstration, were presented at the 105th Convention of the Audio Engineering Society in September, 1998, and received a standing ovation. While the parametric array itself is not patentable, MIT has applied for patents on key aspects of the technology which make it a practical device.
 
 

This directivity plot of a prototype clearly illustrates the extreme narrowness of the beam. (Published in [6]).	During the summer of 1998, we compared distortion of prior devices with our  prototype.  Note that distortion has been reduced nearly to that of a traditional loudspeaker.  (Published in [6]).

Since then, development has been remarkably productive, with engineering and mathematical advances resulting in more sound output, better sound quality, and reliable performance.
 
 

"Everything you do with light, you can now do with sound."^TM

References:

[1]    Westervelt, P. J., J. Acoust. Soc. America, v35 535-537 (1963)
[2]    Bennett, M. B., and Blackstock, D. T., J. Acoust. Soc. America, v57, 562-568 (1975)
[3]    Yoneyama, M., et al., J. Acoust. Soc. America, v73, 1532-1536 (1983)
[4]    Blackstock, D. T., J. Acoust. Soc. America, v102 3106(A) (1997) link
[5]    Berktay, H. O., J. Sound Vib., v2, 435-461 (1965)
[6]    Pompei, F. J., J. Audio Eng. Soc., v47, 726-731 (1999)
                (originally in Proc. 105th AES Conv., Preprint 4853 (1998) )
 
 

About the Inventor:

Beginning his career in acoustics at 16 while in high school, starting as the first high school co-op and becoming the youngest engineer at Bose Corporation, Frank Joseph Pompei continued working part-time and summers for Bose while earning a degree in Electrical Engineering with an Electronic Arts Minor from Rensselaer Polytechnic Institute.  Recognizing the importance and underutilization of spatialized sound, he decided to pursue research in psychoacoustics and application of auditory localization at Northwestern University, earning a Master's degree.  Acutely aware of the limitations of traditional loudspeakers, he had the idea of using ultrasound as an acoustic projector, and is now developing such a device at the MIT Media Lab, continuing his education in pursuit of a Ph.D.

Mr. Pompei is honored to have been chosen as a British Telecom fellow for his third year in a row.
 
 

For More Information:

A technical paper [6] describing the basic device (along with a live demo) was presented at the Audio Engineering Society's 105th Convention (September, 1998).  Please contact them directly with preprint requests.  The same paper was just published in the September 1999 issue of the Journal of the Audio Engineering Society.

"Official" press/public inquiries:  Contact our Communications and Sponsor Relations team.

Or, you can email me.
 

All content (c) 1999 F. Joseph Pompei, MIT Media Lab, except where noted.
B&W photo of early parametric array (c) 1983 Acoustical Society of America.
Reproduction, archiving, and/or redistribution of any part of this document prohibited without written permission from Mr. Pompei or the MIT Media Lab.
Patents Pending.