Method and apparatus for improving effective signal to noise ratios in

Abstract

Improved signal to noise ratio to help speech comprehension in a noisy environment is accomplished by selectively downwardly expanding a speech and noise composite signal when the speech signal is absent, thereby lowering signal components which represent noise. An expansion control signal is extracted from the composite input signal. Operation is based on the assumption that when noise alone is present, the input signal amplitude is less than some reference level and that when speech and noise are present together, the input signal amplitude is greater than the reference level. The response rates of gain changes are quite rapid, and do not introduce distortion or other audibly noticeable artifacts of the processing. The amount of downward expansion of the noise alone is small compared to noise gates to further reduce processing artifacts. The methods of realization include use of, in combination and alone, analog compressors and expanders, analog expanders in combination with voltage clamps and/or automatic level control circuits, two-quadrant multipliers in conjunction with digital control, entirely digital means for obtaining the requisite sensing and gain control, and expandor designs that are analogs of conventional filter designs where the notion of amplitude replaces frequency. Automatic noise suppression may be employed to pre-process the input signal, thereby rendering the control circuit self adjusting for better performance over a wide range of background sound levels. Various microphone, preferably providing directional characteristics, may be used to reduce noise levels in the received input signal.