A Matter Of Diffusion

Like many Stereophile readers, I have often sped home from a concert to fire up the audio system and then, to the sore vexation of my wife and guests, spent the rest of the evening plunged in the morbid contemplation of what, exactly, was missing.

Every lover of music has his pet theories and every theory, it seems, its season. Speaker efficiency and wide frequency response were popular in the 1950s, during the heyday of those hulking Klipschorns and James B. Lansings. The Quad ESLs caught on a few years later, when midrange transparency seemed to be what separated hi-fi from the real McCoy. In the late '60s Amar Bose of MIT concluded that the ratio of direct to reflected sound was lopsided, and redressed the ostensible imbalance with his model 901s. Jon Dahlquist, Richard Vandersteen, and others converged on loudspeaker time and phase discontinuities in the '70s. Up in Canada, Mike Wright reworked the textbook electrostatic speaker for maximum bass slam and dynamic scaling; he had to seal the resulting apparatus in a fragile shroud of sulphur hexafluoride gas to keep it from arcing itself to death. Near Los Alamos, physicist Alan Hill blamed excessive moving mass; he weighed in with a near-massless tweeter comprised of a helium-bathed plasma that expanded and contracted in mid-air at audio frequencies. While Pass (Threshold), Iverson (Electro Research), Curl (Mark Levinson), Otala, and others tackled amplifier linearity problems, over in Scotland Ivor Tiefenbrun impishly insisted the front end was to blame, eventually elevating the turntable in general and the Linn Sondek in particular to new import.

Without wishing to deny the aforementioned trailblazers their due, it is my lot to report that a peculiar-looking device called a "quadratic residue diffusor" leaves little doubt that what largely separates the listening room from the concert hall is the acoustic "fingerprint" of the listening room itself. Most anybody could have told you so (my grandma for one: "Can't put a gallon in a quart jar, son"), but I'm here to say these diffusors represent a powerful new way to address the fact. They're already turning up in recording studios and some concert venues, and once you've heard them, as you will, in the soundrooms of dealers and the living rooms of fellow audiophiles, you're likely to accuse your own listening chambers of having heretofore obstructed your very pursuit of the high end---of having bludgeoned sonic detail, mocked correct instrumental timbre, and thwarted the replication of concert-hall soundfields.

The concept of diffusion will be new to most audiophiles. By and large, we're accustomed to thinking that sound striking a surface can be either reflected (as when a billiard ball bounces off a side cushion) or absorbed (as when it drops into a pocket).

Think of diffusion this way: The billiard ball strikes the cushion and is instantly shattered into sawdust-sized microballs that scatter across the whole table. You need reflection and (especially) absorption to win at billiards; to get your living room to sound like a symphony orchestra playing in a good concert hall, you'll need diffusion, too. To grasp why, consider that, at your ear, the arrival schedule of acoustic reflections supplies the brain the information it needs to get a "fix" on the listening room: its approximate dimensions and geometry, and to some extent the treatment of wall surfaces and other large objects. Let's assume that, with software of appropriate quality, your audio system successfully resolves the nuances of timbre, imaging, and acoustic space that mark the recording site as Carnegie Hall. What happens in your listening room? The ear-brain is presented with two conflicting acoustical field reports: Carnegie Hall on the one hand and your living room on the other. Maybe if you trained yourself long enough, you could become deaf to the living-room layer and experience only the Carnegie Hall layer. (If so, I promise you'll be the subject of great curiosity among psychoacoustics researchers.) At best you'll learn to live with the dichotomy you've got.

Theoretically, diffusion chips away at the listening-room layer in several important ways. First, it tampers with the arrival schedule of normal room reflections in ways that make it hard for your brain to get a clear fix on the listening-room layer. In an untreated room there will be clearly defined energy spikes over time as the sound ricochets off hard surfaces and arrives at the listening position. The position of these spikes on the time track is what tips off the ear-brain as to the location of side walls, ceiling, floor, and other large planar surfaces. Increase the distance between your speakers and room boundaries and you move the spikes further down the time track, away from the direct sound. A strong energy spike arriving 15ms after the direct sound suggests a reflecting surface, say a side wall, nearby. Because sound travels about a foot every millisecond, the reflected path length is only about 15' longer than the direct path length, putting the speaker rather near the reflecting surface. (If the time gap between the direct sound and the first reflected spike is less than 10ms, the ear-brain processes the information as a disruption, or shift, in image.) Place a diffusor at the reflecting site and you scatter that energy across a very broad angle (functionally 180 degrees), meaning only a very tiny fraction of the reflected energy will be steered directly to the listening position. Deprived of the clear 15ms spike, the ear-brain should not be able to accurately "fix" the location of the side wall.

Note about the author: Keith Yates was a founder-owner of both Keith Yates Audio, a high-end retail store in Sacramento, California, and Audio Vaeritae Recordings, a small classical record label, from March 1981 through April 1991. He now designs Home Concert Halls and advanced Home Theater systems (website,) and works as a consultant to high-end audio/video retailers and manufacturers nationwide. His writing has appeared in Stereophile, Audio, and many other magazines.