The influence of quantization of optimal FTN signals with PAM on correlation properties and level of out-of-band emissions

Abstract

Usage of faster than Nyquist signaling (FTN) in practice is limited by the complexity of the formation and detection devices. Optimal FTN pulses may be formed via digital non-recursive filters. One of the important parameters that affect the performance of such devices is the number of quantization levels. In this paper, the influence of the number of quantization levels for optimal FTN signals with pulse-amplitude modulation on spectral characteristics, correlation properties and the probability of error detection is investigated. It is shown that at least 64 quantization levels should be chosen to provide almost potential bit error rate performance and additional spectral quantization noise about - 40 dB for volumes of channel alphabet upto 8. Decreasing the volume of channel alphabet allows reducing the requirements for the number of quantization levels.