Abstract

Digital audio broadcasting (DAB) systems for the AM band are being developed to provide higher-quality radio broadcasts, broader coverage, and data services. Transmission is typically done by means of multistreaming with QAM/OFDM in the high bits/sec/Hz regime in which Reed-Solomon codes, either alone or in concatenation with inner trellis coded modulation (TCM), are natural choices for error control. For the AM DAB application, Reed-Solomon codes with suboptimal decoders of the bounded-distance type offer error correction important for bringing down the error probability, but also offers error detection important for generating block detected-error flags for use in the error concealment, or error mitigation, algorithm in the audio decoder. For multistream systems, the block detected-error flags can also be used to select the component streams retained by the audio decoder. In this paper, we evaluate the performance of Reed-Solomon codes in terms of both detected (flag) and undetected error rates for DAB applications. We provide simulation results for a variety of decoding algorithms, including hard-decision decoding (HDD), successive-erasure decoding (SED), and fixed-erasure decoding (FED), and discuss the channel environments in which each of these algorithms may be appropriate. Among other results, for example, we address the issue of matching the blocklength of the Reed-Solomon code to the audio decoder with its variable frame structure. From our simulation results for HDD, we conclude that the increased error correction capability of longer Reed-Solomon codes more than compensates for the mismatch in terms of error mitigation in the audio decoder. We also observe that SED provides only small improvements for uniform interference channels, while FED offers considerable improvements for some partial-band interference channels. We describe an appealing two-mode adaptive configuration using HDD and FED for mitigating partial-band interference caused by some second-adjacent AM broadcasts.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call