In this letter, we propose a simple but effective technique that improves statistical model-based voice activity detection (VAD) by both reducing computational complexity and increasing detection accuracy. The improvements are made by applying Taylor series approximations to the exponential and logarithmic functions in the VAD algorithm based on an in-depth analysis of the algorithm. Experiments performed on a smartphone as well as on a desktop computer with various background noises confirm the effectiveness of the proposed technique.

A CMOS current-mode nth-switchable-root circuit composed of a compact logarithm circuit, a divide-by-n circuit, and a compact exponential circuit is proposed. The n can be selected from 5 values by three switches. Simulation results indicate that the compact nth-switchable-root circuit has a wide input-current range for relative errors less than 3%, low power dissipations below 630 µW, and high bandwidth over 330 MHz.

A two-quadrant CMOS current divider using a two-variable second-order Taylor series approximation is proposed. The approximation divider is realized with a compact circuit. The simulation results indicate that the compact divider has with sufficient accuracy, small distortion, and high bandwidth for only 1.8 V supply voltage.

A CMOS current-mode companding divider is presented. Currents of both dividend and divisor are compressed into log-domain. Then the logarithm current of divisor is subtracted from the logarithm current of dividend. After expanding the subtraction result, the division function could be achieved. The simulation results indicate that the proposed divider has with good performance at only 1.8 V supply voltage.