The Boston Section of the Audio Engineering Society met on 14 January 2003 in MIT's Bartos Theater to hear a presentation by Dr. Joseph Pompei of Holosonic Research Labs and demonstration of his audio spotlight technology.

Dr. Pompei discussed his interest in '3D Audio' and the research he pursued into using ultrasound as a loudspeaker, first as a graduate student at Northeastern University, then at MIT in their Media Lab facility. Since the directivity of a sound source depends on the size of the source, to fabricate a highly directional source requires that the source be much larger than the wavelengths of sound that it is reproducing. A gigantic conventional loudspeaker could be highly directional, but also impractical; one solution is to use a reasonably sized loudspeaker that generates very small wavelengths of sound, or ultrasound.

Borrowing from research into underwater sonar techniques dating back to the 1960s, Dr. Pompei applied these principles to audible propagation through air. Both researchers and potential manufacturers, such as Matsushita and Ricoh, had also tried to develop ultrasonic arrays to propagate sound through the air in the 1970s and 1980s, but were thwarted by the high cost of large arrays, the excessive distortion (~50%) produced, as well as inconsistencies in and limited bandwidth of the transducers themselves. Others have also tried "Tartini tone" or beat frequency methods, where audible sounds are generated through the interaction of two inaudible sounds; a 200 kHz tone interacting with a 201 kHz tone could generate an audible 1 kHz tone. These techniques have proved to be of limited practicality, however.

Through mathematical analysis and engineering insight, Dr. Pompei overcame these barriers to develop his audio beam system. He realized the air modifies and distorts the ultrasonic waves as they propagate, but in a predictable way. Through characterizing this distortion mechanism, ultrasonic signals can be calculated and generated such that as they propagate through the air, their audible distortion product is the desired sound. Since this transformation of the ultrasonic signal into the audible byproduct takes place continuously over a distance (of about ten meters, for his current system), the planar array of ultrasonic transducers generates a sound source that behaves as a columnar audible array, and is highly directional.

Dr. Pompei's demonstration system consisted of a portable CD player, a powered box about half the size of a shoebox that processes and amplifies the source signal, and a panel about fourteen inches across and an inch thick that contained an array of about forty ultrasonic transducers. A variety of sound samples (birdcalls, breaking glass, whispers) were played for the audience through the system; the perceived sound was indeed highly directional, as one could hear the sound clearly when the array was pointed directly at the listener, but not so when it was pointed only a few degrees away. When aimed at the ceiling, the perceived sound source seemed to be the ceiling itself, as the audible sound would reflect off of any hard surface. Dr. Pompei also noted that since no reverberant energy accompanied the sound perceived by the listener, the perceived sound source was quite close.

The audible signal generated by the audio spotlight has a 12dB/decade upward slope, such that the preprocessing available includes an adjustable lowpass to reduce the audible high frequency content. Increasing the bandwidth to include more low frequency content simply limits overall maximum output of a given system. The system demonstrated to us seemed to have audible content down to a couple of hundred cycles per second; Dr. Pompei noted that some applications can benefit from the use of a conventional subwoofer to augment the signal.

In discussing various applications for the audio spotlight, Dr. Pompei noted that a variety of companies had shown interest in the technology. Initial installations were in amusement parks and museums, and now the field of applications has broadened to include automobiles, office environments, and wherever localized sound is desired. The meeting concluded with a lively question-and-answer session, and a demonstration of the system installed at the Media Lab, which consisted of three audio spotlights and a directional video screen. A video of a singer with accompanying musicians varied as you walked across the room; walk toward one side and you see and hear the trumpet's part, then walk to the other side and you both see and hear the violin's part.

The Boston AES is grateful to Dr. Pompei for his illuminating presentation, and we trust that Holosonic Research Labs has an upwardly directional future!

-Tom Wethern