On source decoding based on finite-bandwidth soft information

Designing a communication system using joint source-channel coding in general makes it possible to achieve a better performance than when the source and channel codes are designed separately, especially under strict delay-constraints. The majority of work done in joint source-channel coding uses a discrete channel model, corresponding to an analog channel in conjunction with a hard decision modulation scheme. The performance of such a system can however be improved by using soft decision modulation. The main cost is a higher decoding complexity. An alternative is to quantize the soft information and store the pre-calculated soft decision values in a lookup table. In this paper we propose new methods for quantizing soft channel information, to be used in conjunction with soft-decision source decoding. We achieve a performance close to that of a system using unquantized soft information