RLL (d,k) coding using structured vector quantization

Run-length-limited (RLL) codes, also known as (d,k) RLL codes, are typically used in high-density digital magnetic recording systems to increase the recording density and control self clocking. We design encoders that map unconstrained information binary sequences into (d, k) RLL constrained sequences. A direct approach to the design of encoders and decoders for our RLL codes can be realized using some ideas from the fixed-rate structured vector quantizer. We show that this approach is well suited to shift, insertion and deletion error correction. For the multistage approach, we prove that our code can be encoded and decoded using p simple parallel codes realized using the same idea. We show also that we can incorporate easily multilevel coding with multistage decoding in the multistage RLL encoder/decoder. We show also that, for both direct and multistage approaches, the optimum thresholds of our RLL code are modified when error-correcting codes are inserted