This thesis studies lossless audio compression algorithms. The lossless audio compression enables the digital audio data encoding without any loss in quality due to a perfect reconstruction of the original signal.
In the domain of lossless compression, research works contain two broad development sections, signal predictive modeling techniques and coding algorithms. The signal predictive modeling techniques are concerned with the understanding of the source signal and utilizing the modeling method to decorrelate a signal. The coding algorithms focus on the processing tasks of efficiently representing a single symbol as a code, usually in binary, given a set of estimated symbol probabilities.
In this thesis, all two sections are investigated in depth and handled from the lossless viewpoint. There are four design principles proposed. The principles aim to improve the lossless audio compression efficiency. Base on these principles, a new lossless audio compression algorithm has been developed to achieve the goal. The algorithm consists of different processing modules including framing, decorrelation, smoothing, prediction, coding, and daptation for optimal compression. Finally, the simulation
results have shown the high compressing performance of the proposed lossless audio compression algorithm.
