Multiplier-and memory-based digital FIR filters for nyquist channels: An assessment of quantization noise

The performance capability of digitally implemented FIR receiver filters designed to complement transmitter spectral shaping and thereby form Nyquist channels for bandlimited voiceband telephony is explored in this work. Assuming the use of receiver filters that match an approximately half-Nyquist transmitter transfer function, a computer simulation assesses the quantization noise contributed by multiplier- and memory-based digital FIR filters. These filter architectures, realized with just several taps and conventional digital hardware, are shown capable of affording minimal peak digital eye closure, with quantization noise of less than 10 percent for quaternary signaling.