![[SoundStage!]](../sslogo3.gif) Factory
TourFeature Article |
| March 2000
The Canadian loudspeaker industry rose from humble beginnings to worldwide prominence in the 1980s. Today it is renowned for producing some of the finest products in the world. Paradigm, Mirage, Energy and PSB are the largest of the many manufacturers in Canada and have become the cornerstone of the industry. Their products are sold worldwide, have gained acclaim from reviewers and consumers, and stand second to none in their respective markets. However, the industry’s status did not rise by accident. The history of each of these companies holds a common thread -- Canada’s National Research Council (NRC) located in Ottawa, Ontario, Canada. Humble beginnings for PSB. . .and for me If asked to pick name one designer who is synonymous with the Canadian industry, most would likely choose Paul Barton. His initials, combined with his wife Sue's, compose the company name, PSB Speakers, whose products have sold now for more than a quarter century. Barton has personally overseen all of the loudspeaker development since PSB's creation in 1972. PSB is a name with which I am very familiar. The
first pair of loudspeakers that I purchased in 1981 were PSB Avantés, which opened my
eyes to the audio world. They were modestly priced bookshelf speakers that sold for about
$400 at the time. My initiation to buying speakers was obviously a good one. PSB has transformed into a speaker company with worldwide clout. The product line is vast and is still targeted at the affordable market. With speakers like the Alpha Mite, priced at $169 per pair, to the $2899 glossy-finished Stratus Gold i, there are plenty to choose from. PSB speakers are equally suitable for two-channel audio or multichannel home-theater use. PSB and the NRC Since 1974, Barton has been using the world-class measurement facilities of the NRC. According to Barton, up until about 1980 he could basically come and go as he pleased since very few, if any, other companies used the NRC like PSB did. Times changed. As companies learned of this incredible resource, they, too, began developing there. The 1980s witnessed the highest level of loudspeaker activity at the NRC. Almost every major Canadian company tested their loudspeakers in the NRC’s anechoic chamber, and plenty of research was taking place. Today, Paradigm, Mirage and Energy have their own facilities, but Barton still loves using the NRC. There are a number of reasons, some obvious, some not. The NRC's facilities are still top-notch and readily accessible. He knows the place, the people, and all of the equipment like the back of his hand. Therefore, there is a familiarity and consequent efficiency that is evident when Barton works at the NRC. However, there are the not-so-tangible aspects too. Barton loves the "collegial atmosphere" of the NRC and the wealth of knowledge within its walls. It’s a place of research, development and inspiration. The NRC employs some 1000 Ph.D.s who conduct research in a multitude of areas. A tour through the main facility reveals a vast array of buildings, each dedicated to different aspects of research. The audio-based research is just a small portion of what goes on. Still, audio research far exceeds what all but very few, very large companies have access to. Therefore, the ability for even the smallest companies to use these facilities offers a competitive edge. In fact, SoundStage! has begun using the NRC's facilities for our own loudspeaker measurements. If it weren’t for the NRC, we wouldn't likely be capable of producing measurements of this quality. Anechoic chambers The NRC's buildings pertinent for audio are known as M-36 and M-37, and they sit toward the rear of the main campus. M-36 is chiefly office space that houses the research scientists, but it does have some nifty devices like a small anechoic chamber used primarily for microphone testing. It also houses a cesium clock, not something one sees every day. The clock has a variety of purposes, but it gives the scientists great delight as they point to one of the many clock displays on the wall and proclaim it the most accurate timepiece in Canada. M-36 also includes the famous IEC standardized listening room, which was developed at the NRC. While regular listening can be done here, it is also where blind listening tests are performed. Blind listening is, of course, a hot topic in audio. However, like almost all scientific research, it is the type endorsed by the NRC and is the basis for Floyd E. Toole’s groundbreaking work on subjective listening and loudspeaker measurements, which will be described below.
The easiest way to describe an anechoic chamber is to say that it is a room that is reflection-, or echo-, free. If you yell into a vast, open space and there are no objects to reflect the sound (mountains, buildings, etc.), the sound will travel from the mouth and never return. It’s a hollow, eerie, vacuous sound. An anechoic chamber is pretty much the same thing, and it's a strange environment for a human. The metal walls of the NRC’s anechoic chamber are located about a foot and a half from the internal walls that surround it. The whole chamber is suspended on springs. This makes it a building within the M-37 building. The purpose of all this is to provide a completely isolated environment that, according to Barton, registers a noise level that is less than 0dB. (0dB is a statistical average of the lowest level of human hearing.) Wedges made from fiberglass are inside the chamber, and they help create the reflection-free environment. No sound gets in, none gets out, and what occurs within gets completely absorbed with nary a bounce.
People have actually tried to listen to music inside an anechoic chamber, but the results are not that pleasing. According to Barton, the imaging is "impeccable" since all room reflections are absent. However, other aspects of performance are dismal. There is "no bass," he cites, because without room reinforcement, bass frequencies are weak from 300Hz down. A couple more reasons why the NRC is so famous with audiophiles The rise and recognition of Canadian manufacturers was not mere luck, nor was it due solely to the expense that went into building such facilities. A lot of personal effort went into the tests conducted at the NRC, and they were done with such precision that the results are still considered groundbreaking. At the center of it all is Floyd E. Toole, a former NRC employee. Toole now works for Harman International, parent company of Infinity, JBL, Madrigal, Revel, among others. Toole holds a B.S., Ph.D. and D.I.C., each in electrical engineering, and his name still comes up when speaking with Barton, or Rene St. Denis, who works for the NRC and performs the testing in the chamber.
These papers yielded such substantial results that designers used the criteria and could actually design their speakers accordingly. Some of these criteria for loudspeaker designs included: wide bandwidth, low distortion, and wide and even dispersion characteristics. Besides Barton, Ian Paisely of Mirage and Kevin Voecks now of Revel worked closely with the NRC at that time. Barton explained that at the time there were no criteria for manufacturers, sales people, and most importantly, consumers, to make informed buying decisions. Now they had some through Toole's rigorously executed research. Today, almost all major players in the Canadian industry adhere to Toole’s findings. PSB’s Image From Pickering, Ontario, the home of PSB Speakers (PSB Speakers is owned by Lenbrook who also owns NAD), Barton makes his journeys to Ottawa religiously. I mentioned to him that I had thought that he did in fact live in Ottawa. He laughed and said that people think that "because I’m up here so much."
On the day of my visit Barton and Mason were working on the design for their new Image 6T loudspeaker -- a large floorstanding loudspeaker priced at about $1000 per pair. The Image loudspeakers replace the company’s Century series. There will also be a similar-looking speaker with a powered woofer section called the 7PT. The Image series is PSB’s newest affordable line of speakers that first debuted at CEDIA '99. The lowest-priced Image speaker is the 1B, priced at $269 per pair. Barton’s goal with the Image series is to produce a series of loudspeakers that are "visually and acoustically matched." Visually means that all the speakers look similar in terms of their outer construction and driver complement? This is particularly important for people who value aesthetics in their multichannel systems. Acoustically refers to matching the sonic characteristics of the speakers so that consumers can mix and match speakers in the line when assembling multichannel systems. For example, a person may use the new 6T for the front left and right speakers, the 9C center-channel speaker, and a pair of 2B speakers for surrounds. Or perhaps he may wish to use 2B speakers for the front left and right channels and augment the bass with the Image SubSonic 6 sub. Whatever the case, the Image series allows such mixing and matching. What’s consistently remarkable is how Barton develops such cost-effective speakers that perform so well -- he’s been doing it for decades now. This can be attributed to his design expertise and his focus on manufacturing technique. To keep the prices low for the Image series, PSB has used innovative engineering, such as the plastic front baffles and the use of uniform drivers for all models. What’s in a PSB design? I was able to quiz Barton about his approach to loudspeaker design. Obviously, the main criterion is flat frequency response. This involves, among many variables, correct driver matching and crossover design. When John Atkinson reviewed and measured the Stratus Gold i loudspeaker in the October, 1997 issue of Stereophile, he wrote that, "Other than a very slight presence-region depression between 2kHz and 4kHz and a narrow ripple an octave lower, the response is astonishingly flat." This is something that Barton is exceptionally proud of, given the price of the loudspeaker and its lofty competitors. However, flat frequency response on-axis is only part of the equation and not necessarily the recipe for a great-sounding speaker.
One of Barton's main goals with each speaker is to match the dispersion characteristics of each of the drivers exactly at the crossover point. Using a two-way speaker as an example, a woofer, working optimally, generally has wide and even dispersion up to a certain frequency point. At that point, the woofer starts to "beam" (i.e., it becomes directional, and dispersion is limited). Since the goal is to maintain wide dispersion, the driver must not be allowed to play into the range where it beams. To do that, the crossover must "roll off" its output before it reaches that frequency. It must also be matched with a tweeter that has its own dispersion pattern not only match the woofer at the crossover frequency so the transition from one driver to the other is seamless, both on- and off-axis. How steeply each driver must be rolled off will determine the crossover topology that is used. Barton describes a higher-order crossover, fourth order for example, as a more forgiving one. First order and the like are less forgiving because there is far more overlap as the drivers' output diminishes.
Digitally erasing your room A few years back, again in cooperation with the NRC, a group of Canadian speaker companies, including PSB, joined together for a project called Athena (not to be confused with API’s new Athena brand of loudspeakers). The project evaluated digital signal processing (DSP) for loudspeaker design. DSP has been, and is being, considered as a way to dynamically alter a loudspeaker’s performance so that a speaker can effectively adapt to different rooms' acoustics. In effect, Athena is attempting to use digital processing to eliminate the effects of a room. Many other companies worldwide are researching this, and some products have come to market with some promising results. Although no products came out of the Athena project, Barton described an interesting experiment they conducted that had astounding results. The goal was to figure out just how much the room contributed to the sound of a speaker. To conduct this experiment they chose two respected loudspeaker designs. Both speakers used dynamic drivers, but were very different in terms of their execution and price. In other words, the speakers weren’t close in design or measured performance. Microphones were placed in a dummy head and, one at a time, they played and recorded the signal from each loudspeaker from the exact same position (one loudspeaker was recorded, removed from the room, and then replaced with the other). They compared the recorded signals and found they were almost identical. This meant that in that room, from that listening location, the room itself had more effect than the loudspeakers! While some may take this information and think that speakers are irrelevant, that’s hardly the case. It does, though, prove that the room is vitally important, and care must be taken to ensure that a speaker gets to perform to its optimum ability. Barton also said that although they found that DSP works wonders on poorly designed speakers in bad rooms, it was found not to offer the same type of benefit to well-designed speakers played in good rooms. Therefore, the Canadian companies are still attempting to produce speakers that sound good in real rooms. Tips and techniques for your room
Regarding subwoofers, PSB has championed the separate subwoofer and satellite system for more than two decades. (One of his older Subsonic II subwoofer designs still rests in the NRC's facilities). Although home theater has made this combination commonplace, the sub/sat approach has not always been a popular one. Through the ‘70s and ’80s, many companies and critics said that attempting to match satellite and subwoofer systems was nearly impossible if you wanted good sound. The integration of the two was deemed too tough. Barton doesn’t necessarily agree. He says that sometimes a sub/sat system can be superior, even though he likes his all-in-one designs like the Stratus Gold i and Image 6T because they allow him to control the end-product’s sound as a whole, in some ways. Barton points to the fact that a separate subwoofer can offer much more placement flexibility. For example, the subwoofer(s) can be placed for best bass response, and the satellites can be situated for best tonal balance and imaging. While this may seem easy, it’s not. The toughest thing is knowing exactly where to place that subwoofer, and Barton has a special technique for just that. Barton says that to determine the right spot, first place the subwoofer in the listening chair (if possible, otherwise on the floor exactly where the chair is). Connect the subwoofer while it is in that position and play bass-heavy sounds. While the sound is playing, you should crawl along the floor with your ears at a low level and listen to find the spot where the bass sounds best. Best sound, in this case, is purely subjective and up to you. The spot where you think the bass sound is best is where the subwoofer should be placed. Simple! For center-channel speakers, Barton talks about the effect of the television itself on the sound of the speakers. When a speaker is placed on a television, the screen acts as an extended baffle for the loudspeaker, which alters the frequency response of the speaker to the tune of a 2-3dB increase over the range of 150-500Hz. Barton designs this into his own center-channel speakers. Therefore, in the free field they won’t measure flat (they will be recessed in that region), but on a television they will. The important thing is to know how the loudspeaker has been designed. If it is intended to be used on a television, use it there. If not, care must be taken to reduce the baffle-effect of the TV, perhaps by moving it away.
Barton has examined his own speakers to test this. He has taken a Stratus Gold loudspeaker, built and measured some ten years ago, and re-measured it today. The deviation is slight, perhaps 1/4dB at most. Although that deviation can possibly be heard, it is certainly not a huge difference that one may attest to hearing. Instead, Barton surmises that the difference in sound that people are hearing over time is conditioning of the brain. He cites experiments done with sight that indicate the brain can accommodate for enormous changes fairly quickly and certainly within the hundreds of hours that audiophiles claim changes occur in. Could this apply to hearing, too? Barton thinks that more often than not, what happens is that the changes in perceived sound that are attributed to component break-in are simply the brain becoming accustomed to the sound. He warns listeners not to fool themselves. In closing Having the opportunity to conduct this interview/tour with Paul Barton was special to me because I can attest to the fact that he indirectly helped introduce me to this industry. When I purchased his Avanté loudspeakers in 1981, I never would have guessed that some 18 years later I would be interviewing him for a magazine I publish. What’s particularly rewarding about all of this, though, is the enthusiasm that Barton maintains for this industry. You may think that someone in this industry for so long, having been interviewed so many times, would become complacent, perhaps even bored. I found the opposite to be true. Barton's enthusiasm is almost overwhelming. While I only expected him to spare the time for a one-hour interview, it spanned most of the day, and he was more than happy to go on longer. In the end, it was all what I had hoped for and then some. I wish to thank the NRC for allowing us to visit their facilities and take photographs. A personal thanks goes out to Paul Barton for the products he’s made and continues to make and for taking the time to help us bring his story to our readers. To find out more about PSB Speakers To find out more about the NRC's loudspeaker testing facilities
|
| All Contents Copyright © 2000 SoundStage! All Rights Reserved |
A Day at the NRC With Paul Barton of PSB Speakers
These
speakers' longevity in my system is a testament to Barton's ability to design musically
satisfying loudspeakers at affordable prices. When I traded my Avantes in some six years
later, the dealer resold them within a day! PSB's reputation carries weight. In fact, one
of the original Avante models is still in use at the NRC for monitoring what goes on
inside the anechoic chamber. It is shown in the photo, right, perched atop the cabinets. 
Next door to M-36 is building M-37 (shown
left), which houses the big anechoic chamber and all of the test equipment for
loudspeakers. The first chamber was built in the 1950s, but this latest version was built
in the last couple of years by Eckoustic Noise Control out of Morrisburg, Ontario. The
chamber is massive, and when its construction is examined closely, you find it is actually
a building within the M-36 building (M-36.1 perhaps?).
What’s it like
inside? Stepping into the chamber results in a disoriented, almost nauseous feeling due to
lack of spatial clues that you get accustomed to in the real world. Speaking inside the
anechoic chamber results in no echoes, so it is easy to pinpoint the source of a sound.
Closing one’s eyes and screaming brings the realization that you cannot even judge
the size of space in the room. Simply standing and doing nothing results in ringing sound
in the ears, likely a bit of tinnitus due to the constant bombardment of noise to which
we’re all subjected in modern life. Such is the world inside the chamber, the perfect
place for accurately measuring a speaker’s performance. 
Toole took on the arduous task
of attempting to correlate loudspeaker measurements with listening impressions -- a topic
that still sparks hot debate today. Using the NRC's vast resources, including the anechoic
chamber, precision measurement devices and standardized listening room (listening chairs
with calibration microphone shown in photo above right), Toole performed measurements not
only on loudspeakers but on the listeners themselves and documented his findings in three
important papers. Today those three papers are bound in one booklet titled Subjective
and Objective Measurements of Loudspeaker Performance, which is available from the
NRC. The booklet is far too complex to go into any great detail here, but it is highly
recommended reading for anyone interested in this topic. Barton points out three of
Toole's main findings that he believes are critical:
Almost all of the development for PSB
Speakers happens at the NRC, and Barton oversees it all. But he has a helper now. Mark
Mason is an electronics designer with PSB Speakers, and he not only assists Barton in the
loudspeaker design, but he also works on the amplifier portions of the new subwoofer and
loudspeaker designs. 
Consistent with the NRC
findings, wide and even dispersion (measured by the off-axis response) is crucial for good
loudspeaker performance in a real room. A listener not only hears the direct radiation
from the loudspeaker, but the reflections off the room boundaries are heard as well.
According to Barton, "65 to 70 degrees off-axis is the ‘second sound’ that
is heard. Together with the direct sound, the two combine to form the timbre of the sound
for the listener." Taking that further, if a speaker is highly directional throughout
the frequency spectrum, the boundaries have less impact -- this, according to Barton, is
why recording engineers usually listen to monitors in the nearfield. Nearfield listening
"eliminates the room." However, most listeners don’t have that luxury, and
eliminating the room through nearfield listening is not possible or practical. The room is
a very big consideration. Speakers with uneven off-axis performance may result in uneven
frequency response at the listening position.
This two-way speaker example
is very simplified. Obviously, there are other factors that affect the design. However, it
does help to explain why Barton chose to use a low crossover point of 2100Hz on the
Stratus Gold i for the transition between the midrange and tweeter and 250Hz for the
midrange and woofer. When viewing a whole set of measurements, you can see flat on-axis
response (see photo right), as well as excellent off-axis response. One interesting
observation I made is that Barton and the NRC take off-axis-response measurements of
loudspeakers over a complete 360-degree sphere (i.e., vertically and horizontally
from the front to the back). This is called Total Radiated Power.
Barton’s knowledge of
loudspeaker design and placement is so vast that he’s really a walking library of
helpful tips and techniques. While insight into design certainly helps, you can appreciate
what goes into developing a good loudspeaker -- as well as tips and techniques for getting
the most out of your current setup in your own room has rewards right now. Barton was more
than happy to share some of his knowledge with our readers.
Finally, and perhaps most
controversially, Barton talks about the supposed break-in effect of components that has
become so popular in audio today. Break-in refers to running components for a long time
(sometimes hundreds of hours) to the point where their components "settle" into
their proper operating mode. Barton doesn’t doubt that some components do change
subtly, but he thinks that the major improvements people think they’re hearing
aren’t in the components at all. Barton doesn’t doubt that people are hearing
these changes, but thinks that what they’re hearing is actually brain break-in.