The Room Vanishes

November 21, 2000

by Edward Jay Epstein

In my modest home theater, Sasha Lazard, a bel canto singer, is listening to the theme song she sang in the Japanimated fantasy movie "Princess Mononoke," while toggling back and forth between two settings on the remote. One, "Bypass," sends the digital information from the CD directly to the speakers; the other, marked "A," reroutes it through an intermediary stop -- a state-of-the-art sound processor called a Sigtech 2000, developed by Cambridge Signals Technology and now used in ultra-high-tech broadcasting studios, such as Walt Disney's in Los Angeles.

Each time she presses button A, the music changes. She looks up agog. "It sounds like I'm someplace else. It sounds like I'm back in the recording studio." In fact, by pressing button A she had made my room disappear -- at least psychoacoustically -- and what remained was the sound of another place in space and time, the room in which the recording was made.

Up until the 20th century, music was heard only in the venue in which it was performed. Whatever its quality, everyone experienced it in much the same way. With the commercial development of the record player in the early 1900s, performers were no longer inseparable from their performance. What they sang and played in a concert hall or studio could now be transported to another time and space. Any room in a home, no matter how small, now became a concert hall. For example, an entire symphony orchestra performing Tchaikovsky's "Overture of 1812," as well as the sound of the auditorium in which it was played, could now be jammed into a 100-square-foot bedroom. The echoing vibrations from so many instruments, not to mention the exploding cannons, will obviously no longer sound the same. What one would hear is a musical cocktail. One part would be the originally recorded music, another part would be the reverberations of the original auditorium. And a third part would be the sound of the bedroom and its furnishings. Since the sonic waves from the vibrating room parts, ricocheting off walls and ceiling, will crash on the listener's ear only about 1/5,000 of a second after the sonic waves arrive directly from the speakers, the brain, which had no evolutionary needs for parsing such minute distinctions in incoming sounds, will blend the two sounds together in a new and idiosyncratic aural brew. Unintended instruments, such as walls, cabinets, table tops, lamps and doors, will join the clarinets, french horns, drums, violins and cellos in the original orchestra, at times blurring, if not totally obscuring, them. To hear the performance without these interfering obstacles, the only practical option is to get rid of the offending room, which is the purpose of the Sigtech 2000.

When Ronald Genereux at Cambridge Signals Technology in Cambridge, Mass., developed the Sigtech 2000, he took advantage of a curious phenomenon in acoustics: sound cancellation. If two identical sound waves encounter each other, and one is inverted -- that is, 180 degrees out of phase with the other -- they cancel each other out. A noise and its upside-down clone thus produce no noise. Such noise cancellation has been used for years in places where offensive noise occurs uniformly and constantly. A microphone simply passes it to a processing chip that inverts it and retransmits it to knockout the unwanted sound. Airplane manufacturers, for example, have eliminated much of the engine noise in cockpits in this way.

But canceling unwanted room reflections that occur over a wide range of frequencies in reaction to highly variable recorded sound isn't nearly as easy. Mr. Genereux had worked out the mathematical concepts for the machine almost a decade before advances in computer technology made it commercially feasible to build one. What made it practical was Motorola's mass production of the dedicated digital-sound-processing chip, or DSP, for the defense and telecommunications industries in the late 1980s. Initially, Mr. Genereux used 25 of these DSPs to analyze, replicate and invert room sounds, but later found that the job could be done with only five DSPs. Still, even with the falling price of computing power, the Sigtech 2000 for CDs and DVDs costs a minimum of $6,300. I couldn't resist.

Cambridge Signals Technology dispatched Mark Donahue, a field engineer with considerable experience in room vanishings for recording studios and concert halls, to my New York apartment, fully equipped, with a cart full of professional-looking equipment. First, he efficiently repositioned my tower speakers so they were optimally located in the room. Next, he connected the Sigtech 2000, a black box about 18 inches square and 4 inches high, to a portable computer and to a microphone fixed on a tripod placed where I usually listened to music. Then, for the next hour, the Sigtech emitted a preprogrammed set of pings, hisses, thumps and other test signals, which registered on the computer screen in jagged lines that looked like a mountain chain drawn by someone with double vision. One set represented the sounds coming directly from the speaker; the other, those reflecting off various room surfaces. From this sort of sonogram of the room, he identified the most prominent haunts of the sonic trolls. It took another hour or so to construct computer programs, or "filters," through which the Sigtech would add digital data to that already encoded on my CDs, laser disks and DVDs. This army of inverted clones generated by the machine's DSPs would zap at least 10 decibels of unwanted room sounds and, through more digital wizardry, also restore some of the music that out-of-phase room sounds had nullified. The filters were then put into the Sigtech's memory. Mr. Donahue left behind (temporarily) the Sigtech 2000 so I could listen to music, operas and movies at leisure in a disappeared room.

Some two-dozen volunteers turned up, including the diva-in-training Ms. Lazard with her recording. Aside from "Princess Mononoke," the visitors tested a wide range of DVDs, including Janacek's opera "The Cunning Little Vixen," Donizetti's "Lucia di Lammermoor" and Woody Allen's film "Bananas." When the room vanished, they discovered previously hidden woodwinds, french horns, cellos and other instuments, as well as auxiliary voices previously camouflaged by room noise. During Joan Sutherland's rendition of the mad scene in "Lucia," one guest temporarily managed to break the illusion of being somewhere else -- the Metropolitan Opera House in 1982 -- by simply clapping his hands together once. Sonically, room abnegation proved to be a vast improvement, but it posed a nagging philosophic problem. What approaches reality more closely: real music in an artificial "space" or inauthentic music in a real room?