Performance comparison of digital equalizers for magneto-optical storage channels

This paper compares performances of several linear and nonlinear equalization/detection methods for magnetooptical storage channels. The simulation data sets with linear and nonlinear distortion are generated based on the measured characteristics of a real magneto-optical disk, in which the data is (1,7) run-length limited coded and recorded in NRZI format. For linear signals, FDTS/DP of depth 3 has the lowest bit error rate. For nonlinear signals, the nonlinear detection methods for fixed delay tree search with decision (FDTS/DF) and multilevel decision feedback equaliser (MDFE) using look-up table achieve a significant improvement in performance over the linear methods. For PRML (partial response maximum likelihood) with PR II, however, the minimum run length constraint is the most dominant factor for the performance for both linear and nonlinear signals