We derive an upper bound on the average MAP decoding error probability of random linear SW codes for arbitrary correlated stationary memoryless sources de ned on Galois elds. By using this tool, we analyze the performance of SW codes based on LDPC codes and random permutations, and show that under some conditions, all but diminishingly small proportion of LDPC encoders and permutations are good enough for the design of practical SW systems when the coding length is very large.