Birthplace of Good Sound
November 17, 2005
Robb Report - HOME ENTERTAINMENT - The Finest in Audio/Video Technology and Design
December 2005 - By Brent Butterworth
It is so quiet I can hear the blood rushing through my ears. Deprived of sound reflections that would tell my brain where the floor is, I start to lose my balance. I reach out for the only handhold I can find; a slim wedge of fiberglass shrouded in wire mesh. I don’t want to fall from the spindly metal platform that supports me, because I know that even though there are people outside, just 15 feet away, they won’t hear me no matter how loud I scream. Yet all I can think is, “I sure wish I had one of these.”
Paul Barton In the anechoic chamber at the NRC with the PSB CW800E In-wall Loudspeaker
Although it may sound like I’m trapped in some sci-fi movie set, I’m actually exploring one of the most important places in audio history: the anechoic chamber at the National Research Council campus in Ottawa, Ontario. In this room and another I am soon to visit, NRC researchers discovered much of what we know about building a good loudspeaker.
As a lifelong loudspeaker fetishist, I’ve always dreamed of making a pilgrimage to the NRC. Much of what I know about speakers comes straight out of the numerous Audio Engineering Society (AES) papers generated at this facility. Of course, not just anyone can walk into the NRC and start asking what all those buttons do. Luckily for me, I have an invite and an immensely qualified tour guide; Paul Barton, founder and chief designer of PSB Loudspeakers. Barton was the first manufacturer to use the NRC’s research and facilities to improve his speaker designs; not only is he intimately familiar with the place, he played an active role in shaping the lab into what it is today.
The chamber in which I’m so blissfully trapped is termed anechoic because sound cannot echo inside. Nor can it escape, nor can it enter from outside. The entire space is lined with sound-absorbing fiberglass wedges several feet thick. A speaker sits inside, opposite array of microphones. Because the chamber is completely soundproof and free of echoes, the mics pick up only the sound coming directly from the speaker; there are no reflections of sound from floors and walls to smear the measurement.
“When I first brought my speakers to the NRC for evaluation back in 1974, there wasn’t much science in the design of speakers,” Barton says. “Very few companies had the facility to do lab measurements, and there wasn’t even much test equipment available then. The first speakers I measured here definitely needed some work. But during just that first visit I was able to massage the crossover [the circuit that sends bass to the woofer and treble to the tweeter] on my Avanté speaker to the point where it became one of the best the NRC had tested at the time. It made us realize how bad most of the speakers on the market then were.”
Dr. Floyd Toole – now vice president of acoustical engineering for Harman International and one of the world’s pre-eminent authorities on speakers – was Barton’s first contact with the NRC. “Back then, we didn’t know what the relationship was between measurements in the anechoic chamber and what listeners liked and didn’t like,” Barton says.
“Floyd took an interest in the challenges that speakers offered, and he decided to start looking for answers.” And the room where he found those answers is our next stop.
We step into what is surely the world’s most famous listening room, in a building adjacent to the one holding the chamber. It is here that Toole conducted the research that led to the publication of the Green book, a revered series of AES papers that set industry
Brent Butterworth starts his blind loudspeaker listening test
standards for loudspeaker testing. Indeed, this very listening room has become a worldwide benchmark for audio evaluations.
Which is why I am a bit disappointed to find that this hallowed space looks much like an ordinary suburban living room. Goldenrod-colored drapes, decades out of style, line the wall behind the speakers. Standard residential cabinetry covers the opposite wall. In the center, four chairs surround a little table that holds a switch panel with five buttons. The space measures 22 feet long by 13.5 feet wide by 9.2 feet high – no larger than a den in an average tract home. But that’s the point. This room represents not some technical ideal, but what Barton calls a “typically good” space characteristic of the environment into which most North Americans would plop their speakers.
The focus of the room is a line of three gray fabric screens. Behind the screens, I see an illuminated numerical indicator, and I know immediately what Barton has in store: He’s going to put me through the same demanding blind listening test that PSB goes through to evaluate its speakers.
“By conducting carefully controlled blind speaker tests, then measuring all the speakers to see what sonic characteristics appealed to the listeners in the tests, we were able to figure out what matters and what doesn’t,” Barton explains. “For example, we now know that a peak in a certain band of frequencies is more objectionable than a dip, and that a broad, gentle peak is more objectionable than a sharp, narrow peak. It saves a lot of time in speaker development, because I know what measurement results to shoot for. I can now do most of the design work in the lab, and go into listening tests only after the design is quite well refined.”
I take a seat and Barton hands me score sheets for four speakers. The sheets include 0-10 ranking scales for such characteristics as clarity/definition, fullness, and brightness; more scales for overall pleasantness and fidelity; and a frequency response chart on which I’m supposed to estimate the frequency ranges at which a speaker’s imperfections lie. The categories on this sheet were determined as a result of decades of research at the NRC and elsewhere. Still, filling in all this info for each of four speakers is an intimidating task; I feel a despair I haven’t felt since the Saturday morning in 1979 when I stared down a blank SAT test form. What’s worse, because I’d bragged to Barton about how many blind tests I’ve participated in, he launches right into it, assuming I’ll understand the ranking categories.
As the NRC’s René St. Denis switches randomly among the four speakers behind the screens, I become frustrated. After all, I have 40 ranking scales to complete. On top of that, I strongly suspect that Barton has included at least one really crummy speaker in the mix, and I sure don’t want to pick that one as my favorite. But soon I settle down and concentrate on one characteristic at a time; it’s pretty easy to decide, for example, which speaker sounds brightest, and which sounds dullest. In 10 minutes, my score sheet is full.
I share my results with Barton, praying that I didn’t choose a $29 plastic
After listening to the blind test Brent picks the CHS60 In-Cabinet Loudspeaker as his favorite
computer speaker as my favorite. In my rankings, one speaker sounded terrible, one sounded competent but rather thin, one sounded good although a little too full, and one sounded outstanding. He slides the screen to the side to reveal a primitive-looking 40-year-old Tannoy recording monitor, a PSB Image B25 Monitor (a $449-per-pair bookshelf speaker), a PSB CHS60 In-Cabinet Loudspeaker (a $1,749-each speaker designed for custom installation), and a PSB Avanté II (a 1975 model that was the first speaker designed from start to finish at the NRC). My reputation is safe – I picked the CHS60 as my favorite, followed by the Image B25 and the Avanté II, with the old recording monitor far behind.
“This is a nerve-wracking process,” I point out. “Don’t you worry that the test subjects will be so concerned about picking the ‘wrong’ speaker as their favorite that it’ll throw off their concentration?”
“No, because we normally do a dry run so the test subjects can get comfortable with the procedure,” Barton replies. “We start collecting data in the second run. What you’re getting is just a taste of the procedure. These tests aren’t beer and pizza with your feet propped up – they’re a serious and exhausting activity.
“Some audio critics don’t believe in blind testing,” he continues.“They think it places too much stress on the listener, and of course they’re worried that their blind test results won’t match what they’ve said in print. But I am absolutely convinced that if those critics got more experience as blind test subjects, they’d become more comfortable and they’d start to prefer blind testing. We all have our preconceptions about how a certain type or brand of speaker will sound, and that bias affects your perception of a speaker unless you don’t know what speaker you’re listening to. It’s even worse if you design speakers – you think I’m not going to have an emotional attachment to a speaker design I’ve sweated over for a month?”
Has all this research actually improved PSB’s products? Absolutely, according to Barton. “Take the Stratus Gold,” he says, “which has been one of our most successful speakers.
Paul Barton going over the measurements from a recent Loudspeaker test
We designed it for a flat frequency response [in which no particular frequency range – bass, midrange, or treble – is emphasized] in the anechoic chamber. But thanks to blind testing, we realized that a perfectly flat response sounds a little too bright in an actual room. So we toned down the response a little in the 5 kilohertz [treble] range, and right away it became a winner. If you don’t do this kind of testing, you’ll probably never uncover those kinds of flaws.”
Whether or not the standards developed at the NRC become accepted by even the most obstinate audio enthusiasts remains to be seen, but one thing is certain – the discoveries made in these rooms have had a major effect on what we hear in our own living rooms.
Be the first to comment below!