The Analysis And Simulation On Lossless Audio Coding Of Mpeg-4 Standards

Lossless audio compression gradually becomes a focus of research on digital audio compression. A lot of research institutions and companies have proposed some lossless audio compression arithmetic. Because of its good performance in compression efficiency,complexity and flexibility, more attention are laid on MPEG-4 ALS(MPEG- 4 Standard for Lossless Audio Coding).After a study on the MPEG-4 ALS coder, this thesis proposes two improvements. They are very easy to implement. One is progressive-order prediction of the starting samples at the random access points,the other is the improvement of quantization of predictor coefficients. Then data and diagram thoroughly demonstrate the superiority of progressive-order prediction. Eventually, we compare some MPEG-4 ALS's capabilities, such as prediction order, prediction error's entropy and spectrum,minimum, maximum, average value and variance per frame, with that of AudioPak. Experimental studies indicate that MPEG-4 ALS coder has better compression performance.Based on the study of design of framing, predictor and entropy coding, we design, simulate and analyze lossless audio coder under MATLAB. Meanwhile, audio coder and decoder are implemented and a great deal of experimental data is obtained, which offers guarantee for analyzing the lossless coder and decoder's performance.