Noise blanker using an adaptive all-pole predictor and method therefor

摘要

A noise blanker ( 40, 106 ) monitors and removes noise from a sampled signal by adaptive filtering ( 98, 150 ) the sampled signal to generate trained adaptive filter prediction coefficients. The sampled signal is provided as an output signal when the noise blanker is in a training mode. A noise monitor ( 34, 154 ) detects whether noise contained within the sampled signal exceeds a predetermined threshold and provides a control signal in response to the detecting. The noise blanker is placed in a prediction mode for a predetermined amount of time in response to asserting the control signal. A prediction output signal is generated using a plurality of prediction coefficients as an all-pole filter. The prediction output signal has minimal noise content.