About half an hour north-west of Los Angeles, in a nondescript industrial park a few blocks from a Six Flags theme park with its roller coasters leaping above the horizon, lies a Samsung Research America office dedicated to audio gear. Home to 21 employees with combined experience in the industry totalling more than 300 years, and housing some of the most advanced audio testing equipment on earth, this facility has a crucial role to play in helping its parent company achieve a new goal: global dominance in the consumer audio market.
“Samsung wants to be number one in the audio business,” research facility head Allan Devantier tells us on an overcast and wet Monday in January. Devantier is middle-aged, slim and bald. He wears a blue-and-white-striped collared shirt without a tie, a simple gold wedding band and rimless glasses. Nothing in his manner suggests he’s prone to hyperbole.
“In white goods, phones and TVs, Samsung’s already at the top,” he says. “We want to achieve the same for the audio business by making great sound. Samsung TVs are dominant because of their great picture quality — there’s great industrial design, too — but the picture is really, really good.”
Sound and vision
Devantier knows good sound when he hears it. He designed a now legendary speaker called the Camber 3.5 and has worked for a number of prestigious audio hardware companies, most recently Harman (a company Samsung looks set to acquire) where he held various positions over a greater than 20-year stint. He left Harman three and a half years ago to join Samsung. In those early days with the Korean electronics giant his team was five strong, now it’s grown to more than 20.
But the facility’s biggest asset has to be its staff. Between them, they’ve worked at most of the biggest names in audio. From Bang & Olufsen and Beats to Gibson, Roland and JBL. And a full third of them are active musicians. Which can’t hurt, as even after meticulous measuring and testing, the primary means of testing audio gear is still by playing music over it.
Correlation and causation
Devantier explains that the reason so much attention is giving to testing is the clear link between results on specialist microphones and regular listeners’ ears. “Good results on the measurements tend to mean good real-world experience.” And it’s why the methods used at the facility have now become industry standards for accurately measuring sound quality.
To be clear, when Samsung says it wants to dominate the audio business it means speakers and, in particular, the sound bars many consumers buy to accompany their TVs, especially high-end, wafer-thin panels that cost tens of thousands of rand. “This lab is 100% funded by [Samsung’s] video division,” Devantier says. “Headphones are part of the mobile business. That will change in the future, but I don’t know when that’ll happen.”
Allan Devantier, VP audio R&D for Samsung Research America’s digital media solutions lab.
Consolidation of Samsung’s various audio-related divisions is the least of the lab’s concerns. More pressing ones include the fact that TVs get thinner every year, especially at the premium end of the market, but his team is expected to make the built-in speakers ever more impressive to match the improved picture quality. “But to get good bass you need to be able to move a certain amount of air,” Devantier says. “It’s about physics. There are certain limits.”
Instead of defying the laws of physics, Devantier’s team is using computers to model speakers and monitor them in real time. “Like car suspension systems, speaker cones tend to be designed with more excursion than necessary. Our new range includes a computer that knows the cone position at all times, so we can run the cone at its limits all the time.”
This computer awareness can also preempt things like massive bass notes and adjust the woofer accordingly, reducing distortion. This “anti-distortion” technology — Devantier stresses it’s only a working title — is just one of the patents the team is trying to secure. While academics get measured on papers published, teams like the one in Samsung’s Sound Lab get measured on the number of patents they secure.
Fortunately for Devantier and his colleagues, Samsung is willing to spare no expense in the pursuit of its goal. He won’t be drawn on how much the lab has cost but does say the only request he’s had turned down since he started was for an administrative assistant… but even that’s about to change now that the team is growing, and the facility with it.
The team in California is also responsible for the computer modelling that’s made Samsung’s 360-degree speaker range possible. “Those used to have a tonne of drivers,” Devantier says. “Now there’s one tweeter and one woofer. We used computer modelling to figure out the shape that would allow for that, from where to locate the drivers to the shape and materials of the enclosure.”
For Devantier, the only difference between a $500 piece of Samsung sound gear and a $1,000 piece should be volume and bass. “Everything else should sound as good across the range. Like TVs, which all keep the same tech and offer the same picture quality… some are just bigger. For sound, that means the midrange should be uncoloured, and the sound should be neutral and natural. It’s only the volume and bass that make it ‘bigger’. And that requires bigger drivers and enclosures.”
Proof’s in the listening
So how does Samsung actually test audio gear, whether its own or rivals’? For starters, it uses the two anechoic rooms on site. The first has four walls and a floor and ceiling of sound-absorbing material. The second is almost identical, except for a single wall the devices being tested are mounted on, “because most TVs and sound bars are put against or near a wall”.
A microphone on a moving arm takes measurements of the same sound output at various angles, and devices being tested are rotated so the resultant measurements form a spiral shape around the device and no two measurements are taken from the exact same spot.
“There’s almost aways acoustic interference between the woofer and the tweeter. Measurements underestimate that problem,” Devantier explains. “We found you need at least 250 measurements to get that. But we rounded up to 360 because it was a nice round number,” he adds with a smile. In reality, he points out, there are 361 points because the start and end point are the same.
“The notion that ‘you can hear things you can’t measure’ is one proliferated by people who can’t measure their products properly,” Devantier says. “We can measure everything we can hear.”
Why all the testing? Because listener preferences tend to match the test results. In addition to the somewhat clinical testing, Devantier and his team also conduct blind listening tests with a mixture of “trained listeners” and amateurs. For these, the devices being tested are put on a rotating platform behind a screen so that each tested device is “in the same place, at the same volume and unseen”.
Where amateur listeners are used, a lot of them are put through the listen-and-rank procedure. Because, as Devantier puts it, there “needs to be enough data for statistical significance”. But the team is also aware that how a product sounds is only part of what will make or break it in the market. “For us to make a successful product it has to sound good, it has to look like it sounds good, and it has to be easy to use.”
So, when does Samsung hope to climb to the top of the audio equipment charts? “If the Harman deal goes through, on sheer numbers Samsung could be the top audio brand by the end of 2018,” Devantier says with confidence. Unit sales are the metric Samsung uses to measure its position and its success, so the Harman deal would give it a massive boost.
But is the target of dominance in such a hyper-competitive sector something Devantier considers realistic? “When we started out it felt like we were five guys in a dingy being told to redirect a cruise liner. But now we’re 20 plus in a tugboat,” he says with a grin.