Apple, Amazon.com, Google, and Microsoft are among the companies trying to get you talking—to your phone, to your TV remote, to the funny-looking speaker on your desk. Amazon’s Alexa can order a cookbook and Apple’s Siri can set an oven timer for the cake, while Google’s Home silences the smoke alarm and Microsoft’s Cortana texts party guests to bring a dessert, all via voice commands. It’s impressive right up until the virtual assistants start responding with a familiar chorus along the lines of: “I’m sorry, I didn’t get that.”
These kinds of features test the limits of the microphones they require. The mics in most consumer technology haven’t kept pace with the advances in, say, cameras. They still aren’t great at focusing on faraway voices or filtering out background noise, and they often require too much power to be listening at all times. So the race into voice control by device makers is putting fresh pressure on the handful of obscure companies leading the $1 billion global market for microelectromechanical systems (MEMS) microphones. The message: We need better hardware, software, or both.
“No doubt, there is an arms race,” says Peter Cooney, an analyst at SAR Insight & Consulting. The big tech companies are thinking a lot more about mics than they have for the past few years. Since the 2012 launch of the iPhone 5, “microphone performance has not really improved that much,” says Marwan Boustany, an analyst with research firm IHS Markit.
Apple and its rivals have challenging, albeit straightforward, demands. They want a higher signal-to-noise ratio, meaning the mic can isolate voices more clearly and from farther away, and a higher acoustic overload point, the threshold at which the mic can no longer distinguish signal from noise. And the chips have to improve in both areas without getting bigger, becoming less reliable, or using more power than before.
Those factors are becoming more important as device makers add more mics. There’s one in the first iPhone, three in 2014’s iPhone 6, and four in last year’s 6S. Motorola’s Droid Turbo smartphone has five mics, and Amazon’s smart speaker, Echo, has seven. The extra mics boost clarity for voice controls or recording when some are muffled, overwhelmed, or pointed the wrong way. The trade-off: More mics cost more money and power and, in some cases, can add their own noise, making for diminishing returns. For now, Samsung’s Galaxy phones are sticking with two.
Market leader Knowles, which shipped about 1.4 billion MEMS mics last year, has turned to software. The company is building audio-processing algorithms into the mic chips themselves, which can recognize when to activate a device’s other audio processors. Greg Doll, Knowles’s vice president of product management for mobile consumer electronics, says the company hopes that will speed voice recognition and reduce power consumption.
Other companies, like upstart Vesper, are experimenting with entirely new designs. A conventional microphone condenses sound waves into electrical signals based on the movement of a metal plate in relation to a second, static plate—which tends to collect dust and moisture over time, reducing its sensitivity. Vesper’s flexible piezoelectric technology generates its own voltage and eliminates the need for the static plate, improving signal and power use, says Chief Executive Officer Matt Crowley.
Vesper says U.S. consumers won’t be using that design until at least mid-2017; it’s in talks with manufacturers, but declined to name them. For now, the smart money’s on tech leaders to keep crowding their devices with more mics, and hoping the virtual assistants can make sense of them. “The next major step,” says IHS’s Boustany, “is not until a new technology comes along.”
The bottom line: The $1 billion digital microphone industry is racing to accommodate the growing demands on its technology.