Speech Enhancement with Minimum Gating

A speech enhancement system enhances transitions between speech and non-speech segments. The system includes a background noise estimator that approximates the magnitude of a background noise of an input signal that includes a speech and a non-speech segment. A slave processor is programmed to perform the specialized task of modifying a spectral tilt of the input signal to match a plurality of expected spectral shapes selected by a Codec.