Minimum Analog-Digital Quantization Resolution Requirements for Digital Communications Systems

This paper considers the effects of front-end quantization on the performance of digital communications systems. Instead of modeling the quantizer as a source of additive noise, we use a novel geometric approach to understand the interactions between the quantizer and received constellation points. We first find the minimum number of bins a uniform quantizer must possess in order to recover all symbols from a fading channel without error, then show that although all symbols passed through the minimum precision quantizer are recoverable at high SNR, they may not be equally reliable in realistic SNR regimes. We conclude by showing that, for the case of a uniform quantizer, the QAM modulation scheme is optimal in a symbol-reliability sense.