On display were several new piezoelectric Micro-Electro-Mechanical Systems (MEMS) speakers, each of which is essentially a semiconductor — a chip-based driver 1/25th of an inch in height and small enough in width and depth to fit on the head of a pencil’s erasure. Eschewing the magnet, voice coil, and other components of the dynamic driver that has been a staple in audio for the last century, the MEMS speaker produces sound by applying voltage to a flexible silicon membrane, thus causing the membrane to vibrate.
While the purpose of the gathering was to explain and demonstrate this new speaker technology, xMEMS Labs also announced that three of its solid-state sound solutions are now available for integration into next-generation true wireless (TWS) earphones, in-ear-monitors (IEMs), digital hearing aids, and emerging personal audio categories such as smart glasses and sleep buds.
The company describes the new models as follows:
• “Cowell, now starting mass production shipments, is the world’s smallest solid-state micro speaker, delivering superior high-frequency response and clarity, plus a wider sound stage, for TWS earphones and hearing aids.
• “Montara Plus, the world’s highest output MEMS microspeaker, is capable of sound-pressure levels up to 120 dB @ 200 Hz, making it an ideal full-bandwidth transducer for audiophile-level, high-resolution in-ear monitors, creating a path to smaller, lighter, simpler IEM designs without the phase alignment and design complexities of multi-driver IEM implementations.
• “Skyline DynamicVent, the world’s first piezoMEMS, DSP-controllable vent blends the benefits of open- and closed-fit earbuds and hearing aids. Skyline ushers in a new era of active ambient control (AAC), offering enhanced adaptive ANC and transparency mode implementations with reduced occlusion effects, eliminating the uncomfortable amplification of one’s own voice or footsteps when their ears are covered.”
MEMS speaker technology is set to revolutionize the art of sound reproduction, according to xMEMS Labs, which claims it is perfectly suited to produce high-res audio in addition to offering other benefits, including excellent sound quality and the ability to be mass produced more reliably than traditional speaker designs. And since MEM speakers do not move the amount of air typically required to produce very low-frequency sounds, they can be paired with a dynamically driven driver during the production process to achieve full-range frequency coverage.
It’s important to note that MEMS speakers, which can be driven by the tiny battery in a true wireless IEM, can only be powered by amplifiers containing MEMS technology. A MEMS-compatible amp must be able to deliver up to 30 volts peak-to-peak so the speaker can play loudly, while applying a DC bias to the thin-film piezo diaphragm. So far, the British company Singularity Audio is offering ONI, the first in-ear monitors to feature xMEMS technology, and iFi audio, also based in England, has announced plans to ship the xMEMS-compliant Diablo X portable DAC/headphone in 2023. [Editor’s note: See Tech Watch: Silicon-Based xMEMS Microspeakers for a deeper dive on xMEMS technology.]
I had a chance to listen to a set of MEMS-based in-ear monitors. Did the sound quality surpass the best and most expensive IEMs that use dynamically driven drivers? No, but what I heard was good with excellent promise. Bass lacked extreme low-end extension but was free of bloat and overhang. There was a generous amount of detail and high frequencies lacked any signs of a fatiguing sharpness, though some listeners may prefer a more harmonically dense sound. And while soundstaging was acceptable, I have heard IEMs that deliver a more open presentation. I am intrigued to see where xMEMS technology — which is already quite advanced — goes.
Following the demo, I sat down for a chat with Mike Householder, vice president of marketing for xMEMS Labs.
Howard Kneller: Why don’t we start with a little background on xMEMS Labs and the new speaker technology you’re announcing today.
Mike Householder: Sure. Thanks, Howard. Yeah, so we’re a startup semiconductor company in Silicon Valley. And what we’ve created is the world’s first all-silicon microspeaker. I can show you the raw materials: This is a production wafer [holds it up] for one of our one of our microspeakers. This wafer has thousands of micro speakers on it, and it’s manufactured using an all solid-state semiconductor process. We replace coil and magnet actuation [found in conventional dynamic drivers] with a piezo MEMS actuator and the paper or plastic speaker diaphragm with silicone. So this is a fully functional speaker at the wafer level. We dice it, package it just like any other chip in in your phone, and it produces next level audio.
Kneller: Just so people understand, when you say piezo, if you pass an AC current through it, it’s going to vibrate just like any other driver would. Correct?
Householder: Correct. Yeah, so the piezo material replaces your coil and magnet — it is the drive layer and it’s voltage driven, instead of being capacitive like a traditional dynamic speaker. So this makes it extremely fast in terms of impulse response and transient response. You apply a voltage and there’s an instantaneous response. It’s super-fast.
Kneller: From my understanding, there’s a wide variety of benefits in using silicon chips to create sound. Maybe you can explain a little bit about that.
Householder: Sure. You know, in general in consumer electronics there’s just a natural gravitational pull toward solid-state components. Your hard drive has gone to solid state, your microphone has gone from mechanical to MEMS. And the speaker is just the next thing. So there are both manufacturing benefits to the OEM (original equipment manufacturer) as well as enhancements in audio quality.
Let’s start with audio first. Being a voltage-driven capacitor device, the impulse response is instantaneous so you’re getting a very fast driver with excellent separation and excellent detail. What you’re also getting is basically a flat, near-zero phase response or phase shift. There is no artificial coloring of the sound in the critical audio regions — you’re getting pulse true reproduction through the solid-state driver.
In addition, the part-to-part consistency in a semiconductor process is next-level. Your SPL or loudness is consistent part-to-part and your phase consistency is about plus or minus 1 degree, part-to-part. So as we move from stereo to spatial audio, that phase consistency is going to pay dividends.
On the manufacturing front, solid-state devices are just inherently more reliable, more scalable, more consistent, and repeatable. So if you’re a manufacturer trying to match left/right speakers to go into earbuds, these are perfectly matched out of our factory. So we’re reducing testing. MEMS drivers are also next level in terms of reliability. We’ve removed the magnet from the speaker so there is reduced weight and less electromagnetic interference. Quality, reliability, repeatability, less weight — it’s all next level.
Kneller: And these are actually automated processes that are used to manufacture these drivers, correct?
Householder: Correct. For a century now, speakers have been manufactured and assembled in a factory, wrapping coil around a magnet being attached to a frame being attached to a membrane. Some of it’s been automated, some of it is still manual labor. Moving to solid-state semiconductor, the process is all automated. You have a fully automated fab — we partner with TSMC to manufacture our devices — and then it’s all semiconductor, automated packaging, and automated testing. So the manpower required to manufacture these speakers is significantly reduced.