Paul Barton at the NRC
October 5, 2007
In the heart of Canada's capitol city of Ottawa, is a "campus" accommodating one of the world's most important centers of accumulated — and constantly increasing — knowledge. This is the central facility of Canada's unique National Research Council, with some 51 major buildings in which state-of-the-art facilities for research into and application of the sciences are at the disposal not only of the Center's own state-of-the-art people (doing fundamental and applied research in all of the sciences) but of Canadians from the private sector wanting to learn and do more with the NRC's help.
There are no apparent limits on scientific curiosity at the NRC, and a tremendous amount of "pure" research heading who-knows-where goes on there. But the realities of present and future day-to-day life are the center's center. And the sweep of NRC-related accomplishments is enormous — from the atomic clock used to set time around the world to the far different clock inside pacemakers; from the BCB tuberculosis vaccine and the cobalt "Bomb" used to treat cancer to the strains of wheat in today's bread; from some of the earliest experiments with VSTOL (Vertical/Short Take-Off and Landing) aircraft to important refinements in the black boxes used in air-crash investigations and the radio beacons that help locate downed planes.
The NRC's Institute for Microstructural Sciences is one of the center's most sophisticated facilities, housed in two imposing buildings on opposite ends of the campus. It goes in tremendous depth into the nature of matter and retrieves immensely useful knowledge from its various Departments, which include Advanced Processes (such as thin film display technologies and deep level transient spectroscopy), Condensed Matter, Device Physics, Epitaxy, Microfabrication, Optoelectronics, Quantum Physics, and Surfaces-and- Interfaces.
What should rightfully have been first in that alphabetical rundown is the pursuit that has preoccupied Paul Barton on his frequent visits over the past 25 years: the Department of Acoustics and Signal Processing.
To get there, you have to pass a building accommodating what used to be a football-field-long tank full of water, in which the behavior of ship hulls was modeled — and in which a working model of the Saint Lawrence Seaway was created to come up with an answer to silting problems that at one time threatened to cut off passage of ships into the Great Lakes. Passing this building tends to put you in a "this-is-SERIOUS-stuff" mode. That mood continues as you move into the two buildings where Paul Barton spends his time at the NRC. It becomes obvious very quickly that the kind of seat-of-the-pants research and design that has tended to characterize the hi-fi business is not what rules in the Acoustics and Signal Processing Department.
The three places where Paul spends almost all his time there are (A) the facility's anechoic chamber, (B) the measurements lab immediately outside the chamber, and (C) the first of the Department's two carefully designed listening rooms. Each of the facilities is, depending on your view of the world, state-of-the-art or state-of-the-science.
The anechoic (that is, completely sound-absorptive, non-reverberant) chamber is housed in its own massive substructure within the chamber/lab facility — "floated" off the building's concrete foundation and away from its inner walls, and tightly sealed off (a bit like a very large and massive walk-in refrigerator) from any possible kind of noise intrusion. The sound absorption is accomplished by an endless number of acoustically dead fiberglass wedges placed so densely and evenly on all surfaces of the chamber, including floor and ceiling, that there is literally no room anywhere for reflection of sound. Fortunately for those who have to spend time in the chamber, the color of the wedges is a fairly pleasant light and warm orange rather than the ratty pale gray of the wedges found in most of the small "quasi-anechoic" (with accent on the "quasi") chambers sometimes found at the test stations in speaker factories.
In the "pure" environment of the NRC chamber, only sound coming directly from a source is picked up by the microphones placed at various points to measure sound output strength. The loudspeakers to be tested in the chamber are placed at the end of a strongly constructed, non-reverberant (of course) filigreed catwalk suspended above the chamber's floor. (Paul Barton's description of the "balancing act" involved in designing speakers is given new meaning when you visualize him wrestling large and heavy speakers like the Stratus full-range speakers and subs into position after an extremely careful unassisted entrance on the catwalk.) The speakers being tested radiate into the longer dimension of the chamber, where calibrated microphones are strategically hung at various points to check sound output levels at the radiating angles that best sum up a speaker's radiation pattern.
The test lab in the room outside the chamber is full of every kind of electronic equipment known to have any value in the analysis of loudspeaker performance. It is equipment that engineers and audiophiles would die for, including extremely sophisticated computerized performance read-out gadgets that have mostly supplanted older, chain-driven chart recorders. (There is still a place, however, for the old and somewhat legendary Bruel & Kjaer chain-drive, with the rapidly moving stylus that yields the endless succession of electrocardiograms that traditionally testify to a speaker's well-being or lack thereof.)
What happens in the almost endless back and forth between the signals fed to the speaker being tested in the chamber and the computer or Bruel & Kjaer representations of its performance is undramatic. Everything is in the details and in constant attention to minute variations in driver materials or crossover components or grille cloths. The cross-checking goes on and on. Measure and change, change and measure. Go off on a long tangent tracking down some anomaly that makes the stylus jump upward or drop off a cliff when it should be gliding. (A tweeter or woofer that hasn't come up to room temperature after a trip along icy roads in Paul's Plymouth Voyager is easy to diagnose, but a subtly defective material sample can be hard to pinpoint.) And keep checking — from all angles, over and over and over and over. Checking also goes on in a normally reflective listening rooms, but Paul has long since established how various kinds of measured performance in the chamber will sound in a range of typically reflective home environments.
The lab facility is very much Paul's home away from home. He has spent far longer in the chamber and lab than any other speaker designer, and in the 25-year process since he began going to the NRC has gained enormous sophistication about acoustics and the sound of speakers. Much of it is from his own persistent curiosity and determination to get to the bottom of things. But a good portion is also from constant association with the scientists and researchers in the Acoustics and Signal Processing department, who are immersed in every aspect of the behavior of sound — and who range from the "purest" research to the most practical applications of the most sophisticated technology. (You walk into the test lab through a room that has housed everything from modeling of airport noise reduction to experimentation with new electronic ways to make sure everyone around a conference table is heard equally well by confreres around the table.) Paul has taken the time — again, far more than any other speaker designer — to exchange with and learn from the NRC's nonpareil acoustics staff, and the result is not only expanding knowledge but a constantly growing quiet confidence that he knows what he knows.
The other NRC location where Paul has spent a good portion of his adult life is the facility's primary listening room, where "double-blind" listening comparisons reveal which speakers sound best and how their perceived performance correlates with measurements. (We will be going into this in some depth shortly in a Web segment to be called "What People Hear." Please stay tuned.) A "double-blind(ed)" comparison of sound quality is one in which neither the person giving the test nor the one taking it knows which speakers are being compared, let alone which is on at a given time.
The conditions of comparison are completely different from those in a typical retail showroom. For one thing, listeners listen to a single speaker at a time in mono, not a pair in either stereo or mono. For another, the speakers are compared at absolutely equal volume through a switching arrangement that has been adjusted with extreme care by a third party who sets up the test, And the entire process unfolds in an objective, unemotional way in which — surprisingly to those who think of speaker design and speaker selection as black arts — listeners wind up agreeing almost all of the time about what they are hearing and what it means.
We invite you to drop back sometime in mid-winter or a bit sooner for our coverage of just what it is they hear and value. We also invite anyone who is interested in NRC activities to drop in at the Council's own Web site and take time to investigate this unique institution.