A low-power, fixed-point, front-end feature extraction for a distributed speech recognition system | IEEE Conference Publication | IEEE Xplore