ADPCM encoding and decoding systems.

摘要

An ADPCM encoding and decoding system consists of an encoder (50) having an adaptive predictor (54) and an adaptive quantizer (55), and a decoder (100) having an adaptive predictor (107) and an inverse adaptive quantizer (102), the adaptive predictors each comprising an adaptive-zero predictor (65,121) of an arbitrary order and an adaptive-pole predictor (71,127) of an arbitrary order, a fixed-pole predictor (73,129) of an arbitrary order, these elements being connected in series, and an offset predictor (68) for rejecting an offset component of the input signals. The adaptive-zero (65,121) and adaptive-pole (71,127) predictors generate estimates for voice signals with time-varying frequency components. The fixed-pole predictor (73,129) generates estimates for voiceband data signals the frequency components of which remain substantially invariant over time.