New types of adaptive quantizers

Adaptive quantizers that are able to adjust each of the quantizer decision levels are presented. Two hardware schemes are proposed. One uses a high resolution analog-to-digital converter (A/D) and the other uses a high resolution digital-to-analog converter (D/A). In both schemes, by adapting the content of a look-up table, the quantizer's decision level can be adapted to the input signal. The adaptation process can be on data blocks, using the histogram of the previous data block for the calculation of the optimum decision level setting for the upcoming data block. An alternative is the new sample-to-sample decision level adaptation algorithm presented in this paper. Finally, we demonstrate the performance of the proposed schemes resulting from simulation and actual hardware implementation. The schemes are basically simple to implement and can lead to significant bit rate reduction and signal compression. Such a system can be used as a simple stand-alone data compression scheme or for preprocessing in a more complicated compression scheme.