“…where the parameter measures the density of the recording. It was shown in [1] that for x = (1; 1; 01; 01) and 1:5 2:75, a target response and a density range of practical interest, the minimum effective distance difference pattern is contained in the set D = f000(+0) k 00; 000 + (0+) k 000g where ( 1 ) k denotes k 1 repetitions of ( 1 ), and we use a shorthand notation to represent the ternary difference patterns, i.e., +00 is used to denote +1, 0, 01. The bit-error probability is well approximated over this range of densities by taking k = 1, i.e., D = f000 + 000; 000 + 0 + 000g:…”