Quantifying information rate increase by constrained codes for recording channels with media noise

The achievable information rates for magnetic recording channels with media noise, when constrained codes are employed, are computed. It is shown that higher information rates can be achieved by using specific constrained codes compared to the unconstrained system with independent identically distributed (i.i.d.) equiprobable binary inputs. The results are compared with the optimised information rates of such channels over general Markov inputs.