A Sound Activity Detector Embedded Low-Power MEMS Microphone Readout Interface for Speech Recognition

This paper presents a sound activity detector embedded low-power MEMS microphone readout interface for speech recognition. The proposed readout interface exploits the sound activity detector to automatically switch to active/standby mode depending on whether a sound activity is present or not. Since voice recognition applications are mostly in standby mode, standby power consumption is of greater importance than active power consumption. Our readout interface consumes only 14 μA in standby mode, which can significantly extend battery usage time. Also, a fast wake-up feature is provided. In active mode, the readout interface converts the microphone input signal to a high-resolution digital signal. The proposed circuit is fabricated in a 0.18 μm CMOS process. The measurement is performed using a differential piezo MEMS transducer. It achieves A-weighted signal-to-noise ratio of 62.6 dBA and a dynamic range of 104.5 dB.

Paper

Full text

PDF

A Sound Activity Detector Embedded Low-Power MEMS Microphone Readout Interface for Speech Recognition

Semantic Scholar · Engineering · 2019

Abstract

This paper presents a sound activity detector embedded low-power MEMS microphone readout interface for speech recognition. The proposed readout interface exploits the sound activity detector to automatically switch to active/standby mode depending on whether a sound activity is present or not. Since voice recognition applications are mostly in standby mode, standby power consumption is of greater importance than active power consumption. Our readout interface consumes only 14 μA in standby mode, which can significantly extend battery usage time. Also, a fast wake-up feature is provided. In active mode, the readout interface converts the microphone input signal to a high-resolution digital signal. The proposed circuit is fabricated in a 0.18 μm CMOS process. The measurement is performed using a differential piezo MEMS transducer. It achieves A-weighted signal-to-noise ratio of 62.6 dBA and a dynamic range of 104.5 dB.

Similar papers

© 2026 NYSGPT2525 LLC