Digital Audio: from Lossless to Transparent Coding

We have seen rapid progress in high-quality compression of wideband audio signals. Today’s coding algorithms can achieve substantially more compression than was thought possible only a few years ago. In the case of audio coding with its bandwidth of 20 kHz and more, the concept of perceptual coding has paved the way for significant bit rate reductions. However, multiple codings can reveal originally masked distortions. In addition, reproduction of critical music items shows that even the best systems can not be considered as truly transparent. Therefore lossless audio coding has become a topic of high interest both for professional and customer applications. This paper will explain approaches to lossless and lossy compressions, in the latter case with emphasis on existing MPEG standards which have found a wide range of communications-based and storage-based applications. It will be shown that the very recent MPEG2 Advanced Audio Coding (AAC) standard outperforms many other coding algorithms (including MPEG-1 coders). Finally, we will address the current MPEG-4 speech and audio coding standardization work which merges the whole range of audio from high fidelity audio coding and speech coding down to synthetic audio, synthetic speech and text-to-speech conversion.