A new piezoelectric voice accelerometer claims to have enabled drastic improvements in voice quality for the earbud, AR glass and true wireless stereo (TWS) designs by adding wake-on-voice (WoV) and hardware voice activity detection (HVAD) features.
Vesper, a supplier of MEMS-based piezoelectric sensors, also claims that its second-generation voice accelerometer, VA1210, has significantly refined voice pickup compared to its first-generation analog sensor, VA1200. The VA1210 voice accelerometer offers higher sensitivity (+17 dB) and a higher signal-to-noise ratio (SNR) of +6 dB than the VA1200.
The VA1210 accelerometer offers robust immunity to ambient sounds and only picks up the user’s voice. Source: Vesper
With WoV capability, the piezoelectric sensor can continuously monitor the TWS wearer for voice activity via bone conduction while consuming a scant 10 µA of current. That allows the sensor to accurately detect voice activity and power on the system processor only when the voice is present. As a result, it enables system designers to dramatically cut average power consumption in devices like TWS.
In the VA1210 voice accelerometer, Vesper’s patented ZeroPower Sensing (ZPS) engine can be programmed to ideally suit the applications serving large area touch sensing, machine monitoring, leak detection, and many others. That makes this low-power accelerometer highly suitable for detecting vibration in Internet of Things (IoT), automotive, and industrial applications.
The VA1210 piezoelectric sensor—available in a tiny package of 2.90 × 2.76 × 0.9 mm—has a differential analog output that facilitates more robust hardware designs. Moreover, it’s reflow solder compatible with no sensitivity degradation.