Contrasting performance of faster binary signaling with QAM

In this paper we determine the performance of Faster Binary Signaling (FBS), an alternative method to Quadrature Amplitude Modulation (QAM) for achieving a high bit rate over an ideal, bandlimited, noisy channel. With this method, signaling is faster than the Nyquist rate. Consequently, there are fewer points in the signal constellation, resulting in a greater separation of the points when the average transmitter power is the same as for QAM. Thus, at the expense of introducing Intersymbol Interference (ISI), there is an apparent improvement in noise immunity. The ISI can be mitigated with maximum likelihood sequence detection. We explore the advisability of trading freedom from ISI for added noise immunity for the extreme case where the system with faster signaling uses a four-point constellation. The question of the efficacy of FBS has been difficult to approach, but FBS has loomed as a possibly strong competitor among alternatives to QAM. We show here how to analyze FBS, and we give examples involving FBS operating at up to five times the QAM rate. In the examples, FBS is revealed to be, at best, of marginal value even if one allows for implementation capabilities far beyond those of forthcoming processors.