Performance-complexity tradeoffs of the MPEG-4 ALS lossless coding standard

MPEG-4 audio lossless coding (ALS) is a standard compression tool for audio signals based on time- domain prediction. In this paper, enhanced encoder algorithms and implementation examples of MPEG-4 ALS are described and performance-complexity tradeoffs are discussed. Compression performance can be enhanced at the cost of computational complexity, using integrated tools, such as hierarchical block switching, multi-channel coding, long-term prediction, and backward adaptive prediction. To reduce the computational complexity of the encoder, simplified algorithms have been developed for the multichannel prediction coding and the long-term prediction tools. As a result, encoder execution time has been reduced to 1/6 of that by the MPEG reference software. It is also found that ALS outperforms other compression tools.