Quantization and quantization sensitivity of soft-output product codes for fast-speed applications

This paper investigates two quantization schemes on three soft-output message-passing decoding algorithms for 2-dimensional product codes. Quantization sensitivity on code rates and channel conditions is also investigated. The surprising yet encouraging result is that a simple (3,1) uniform quantization scheme on the min-sum algorithm results in the best overall quality in terms of space, performance and complexity.