Abstract
It is an issue that the iterative decoding system using the soft-output information based on a posteriori probability(APP) requires the huge memory. In this paper, the amount of required memory and the bit error rate performance are evaluated for two kinds of systems in a perpendicular magnetic recording channel model. A low density parity check(LDPC) code obtained by interleaving the overall code words and a structured random interleaved array(RIA)-LDPC code are combined with either soft output Viterbi algorithm(SOVA) decoder or APP decoder. The results show that the system adopting the RIA-LDPC code and the SOVA decoder reduces the memory by approximately 97% compared to the combined system of another LDPC code and the APP decoder, while the SNR degrades about 0.4dB.
