Hi-Fi+ Publisher Chris Martens recently had an opportunity to catch up with Paul Barton, founder of and chief designer for PSB Speakers of Ontario, Canada. In particular, we took the opportunity to ask Mr. Barton about how PSB—a firm specifically known for its high performance/high value loudspeakers—decided to become a manufacturer of serious, high-end headphones.
Background: Paul Barton founded PSB Speakers in 1972 with the aim of producing loudspeakers that would offer accurate, natural sound. In 1974, Mr. Barton met and began to work with the legendary Dr. Floyd Toole who was conducting research on loudspeakers and listeners’ perception of loudspeaker sound quality at Canada’s National Research Council facilities. In due course, PSB became the world’s first manufacturer to develop new loudspeaker designs from start to finish at the NRC—a practice that continues to this day. Over the years, PSB loudspeaker designs have consistently received critical acclaim and have won numerous awards both for sound quality and value for money.
Hi-Fi+: PSB is of course best known for its loudspeakers, but lately you have done a lot of work to develop high performance headphones. What led you in that direction?
PB: Well, here I am with a company that is about forty years old, developed around traditional hi-fi, but the landscape of musical listening habits has—over the past five to seven years—changed dramatically. So, people’s need or desire to have stationery music systems—only to be listened to when they are at home—has either stayed the same or is declining. The new audience is the person on the go.
In response to this, I couldn’t help but think, “Gee, I have been witness to some pretty incredible research that was taking place in parallel with the loudspeaker research conducted at Canada’s National Research Council back around the time when I first met Dr. Floyd Toole.” That was the research that Dr. Edgar Shaw, Toole’s boss at the time, was doing on mapping of the pinnae of the human ear and research on the human ear canal. Much of that research was focused, at the urging of the medical community, on a pursuit of better ways to make hearing aids, but it also had implications for music reproduction.
So Edgar (Dr. Shaw) took it upon himself to do some fundamental research on the anatomy of the outer ear and ear canal. Given all the resources available to us and given the insights I had picked up by ‘osmosis’ through being around the NRC and Edgar and Dr. Floyd Toole, it was natural for me to become interested in transducers meant to be worn on (or in) the ear.
HI-Fi+: What kinds of insights did the experience afford you?
PB: I got to participate in Edgar’s research to an extent and to learn from it. I can remember Edgar coming to me when he was doing his research, trying to create a perfect point source. He asked me, “Do you have a 1-inch tweeter I could put on the end of a pipe,” where the idea was to use a 1-inch pipe that was about 25 feet long with a tweeter at one end to create am ideal point source. He needed a perfect point source to do experiments on the directivity pickup of the human pinnae. I was able to get a really fundamental understanding and appreciation for the complexity of human hearing—not just for the anatomy of the outer ear and ear canal, but also for the ear as a transducer that turns acoustic energy into electrical energy that the brain can process, enabling us to perceive three-dimensional sound with just two ears. Wow! That’s phenomenal.
So, that’s kind of the background for me getting into headphones. Of course headphones represent a vast, growing market that serves all of these new listening habits that younger people have. It seemed to be a natural for us to share our understanding of what sounds good in a pair of headphones. That’s how the PSB headphone came to be done.
Hi-Fi+: As I understand it, your studies on the in-room response characteristics of loudspeakers have influenced your headphone designs. Can you tell us more about that?
PB: With all of the NRC research being done, and in parallel to our work on loudspeakers, their interfaces with rooms, and on what sounds good (to most listeners), due diligence was done back in the late 1980s and early 1990’s to study optimal in-room response curves for loudspeakers. We did something called the Athena Project at the NRC in which PSB, Paradigm, API (representing Energy and Mirage), and a studio monitor maker called SOTA worked together for three years to figure out how to apply automated room EQ for speakers, or to make a “smart” loudspeaker. This was long before Audyssey or other room EQ algorithms were developed and before powerful DSP chips that would eventually fit in A/V receivers became available.
Back in those days you needed four Motorola DSP 56001 processors under one hood in order generate a proper room EQ curve, and even then it took a very long time because of limitations in processor speeds. Because we could do room EQ, we could make a speaker perfectly flat in a room. But when we did double blind screened listening tests, the speakers that were equalized to be perfectly flat in the room got negative comments from listeners. Most people would say, “Hey, it doesn’t have enough bass.”
Now you have heard me and other people talk about this before, but there is a phenomenon you expect to hear when you put a speaker, a speaker whose anechoic response is flat, into a room. The room adds some gain because the directivity of the bass driver at its lowest frequencies is much more omnidirectional than it is as the frequencies get higher. What this means is that the sound power response of the loudspeaker is not flat. On-axis it can be flat, but if you take all of the energy the speaker puts out, it’s not flat. When you are in a room, you hear some combination of the direct sound and the sound power. And if you look at an average number of rooms and loudspeakers using average types of music, you’ll find that the response that’s derived from a good loudspeaker in a good room is somewhere between the on-axis response and the sound power.
So that’s something we call “room gain” and every headphone—if you play music that’s normally meant to be heard through loudspeakers in a room—should have that same room gain in order for you to appreciate how the music would sound if it were played through a pair of good speakers in a good room. And isn’t that the sound we know we like?
Hi-Fi+: And some would say that since recordings typically are mastered using monitor-quality loudspeakers in rooms, allowing for room gain in headphones is necessary in order for listeners to hear what the record producer meant for them to hear, correct?
PB: Yes, the record producers and mastering engineers were probably making decisions based on hearing speakers that were flat (on axis) in a room. Remember that everything about the recording process, include the playback process where engineers and producers listened to their work and then made decisions about what they were hearing and changing things to make it sound the way they wanted—all of that activity is still part of the art form. And only beyond that point must we be loyal to the way the finished recording sounds.
In a perfect world, we would want to be in an environment, preferably with the exact same loudspeakers that the engineers and producers were using, in order to hear exactly what they heard. Over the years, the sound of good loudspeakers in a good room has evolved as a de facto standard for how engineers and producers intended recordings to sound. So when you listen to a pair of headphones that add in the room gain that you would get if listening to a normal pair of speakers in your room, then you’re not putting together the playback chain in the way the producers expected you would do.
Hi-Fi+: Do you think the high-end audio community will, over time, embrace headphones more fully than it perhaps does now?
PB: I can’t speak for the high-end manufacturers, but I can speak on the headphone world and how it may interface with the high-end community (or even with loudspeaker companies other than my own).
I can only do a comparison here. In an earlier interview we talked about how I started PSB in 1972 and then got involved with the NRC in 1974. Back in those days, when I walked into the NRC listening room and auditioned and compared loudspeakers under evaluation, I can remember different brands from different parts of the country and of differing sizes and shapes all sounded dramatically different. And my reaction at the time was to ask, “What’s going on here? Who’s right? Because they all can’t be right.”
Well, nowadays speakers are much more consistent; they aren’t “black and white”, “day or night” different and if they are then they are rejected. So, I think headphones—in 2013—are in about the same state right now that loudspeakers were in back in the 1970’s. Because there is no agreement on what the ideal target response profile for a headphone should be.
Have you been reading the papers that Sean Olive (of Harman International) has been writing on the subject?
Hi-Fi+: I’ve not yet read the formal papers, but I have seen a fairly detailed PowerPoint presentation that Olive prepared on his recent headphone research, with findings I gathered were extracted from his papers.
PB: Yes. The first paper he did was delivered in San Francisco last October and it confirmed his suspicions that there is no agreed upon consensus as to what the target function needs to be for a headphone. In short, he saw that there was no apparent agreement among manufacturers; they were all different.
Then, the next paper he did described an experiment where he tried two different headphones—one was about $1000 and the other around $300—and applied a number of different EQ curves (I think there were about six of them) to both headphones. Finally, he used double blind screened listening tests to evaluate the results.
One of the EQ curves used was taking as its model the characteristics of the Harman listening room, which has been calibrated to have the Athena Project EQ. By mimicking the Harman/Athena transfer function that they get when playing loudspeakers in the room and then applying that EQ curve to both headphones—that was the transfer function that everyone (in the listening group) preferred.
Hi-Fi+: Is your thought, then, that as headphones march down the same path that speakers have already gone down, becoming more consistent and sharing an agreed upon target function, they will be more widely accepted within the high end?
PB: I think so, because the high-end will then say, “Oh, they figured it out. Now they’re worth listening to.”