Digital MEMS microphone for future standalone voice command
There is high demand for voice command functions in IoT and portable device sectors thanks to speech recognition engines and powerful voice assistants like Siri, ”OK Google,” and Amazon Echo. Current speech recognition systems are typically executing all of the time as they consume quite a lot of power listening and recognizing speech.
Future voice command function will be expected to be operated standalone and will be turn on when it is activated by voice. Low power digital MEMS microphone technology will be suitable for future standalone voice trigger solutions – it will perform very well, extremely low power, and they can be added to an existing design relatively easily.
Algorithms such as noise cancellation and beam forming, which analyse signals from multiple microphones, need to rely on close matching of the sensitivity of individual microphones in the array, ideally to within ±1dB. Although screening, or binning, is one potential solution, microphone designers are looking for ASICs to provide adjustable gain that enables tuning out of process-related variations in MEMS fabrication.
The LC706200 product family provides high performance solution. It has several other features that ensure enhanced linear performance over a wide operating range, including low input-referred noise of -106 dBFS with the advantage of an 8 kHz low-pass filter for peaking compensation, and a low noise internal bias and regulator circuit that leverages ON Semiconductor’s giga-ohm resistor process. The device also has a high Power-Supply Rejection Ratio (PSRR) that prevents unwanted noise entering the signal chain, and power management including sleep mode and a low-power mode that remains responsive to voice commands.
Changes in the way people use computers and smart devices are driving demand for reliable and high-performing MEMS microphones. Digital ASICs now in the marketplace maximise freedom for microphone developers to deliver best-in-class products to satisfy these demands.
About the author:
Masahito Kanaya is Product Marketer at ON Semiconductor – www.onsemi.com