The effects of changing the allocation of filtering on the performance of a modified duobinary partial response system are examined. Two classes of filtering as well as two types of decoders are considered. The precoded system is, in general, superior to the system with decision feedback decoding, except at low values of the signal-to-noise-ratio and large fractions of the filtering at the receiver.

Pseudoerror monitoring techniques, besides providing a precursive, measure of system degradation, can also be used in extrapolation schemes to estimate the actual bit error rate of a digital link. This study examines pseudoerror techniques as applied to a precoded, modified duobinary system. Two known methods of pseudoerror generation, namely, threshold modification and sampling offset, are compared with two new monitors using controlled amounts of amplitude and phase distortion. The threshold modification monitor is most robust to channel degradations, while amplitude or phase distortion monitors seem to be more accurate in estimating the actual error rate but less robust. The sampling offset monitor seems to be less robust than the other monitors due to its unavoidable sensitivity to degradations affecting the system phase. Methods of improving the estimation accuracy by the choice of degradation parameters are also examined.