ADAPTATIVE ENTROPY CODINGFOR MPEG-4 AUDIO LOSSLESS CODING

• Main Authors: Jung-Hau Chen, 陳仲壕
• Other Authors: Ching-Kuen Lee
• Format: Others
• Language: en_US
• Online Access: http://ndltd.ncl.edu.tw/handle/3j675t

碩士 === 大同大學 === 通訊工程研究所 === 95 === There are two major techniques for audio compression; one technique is loss compression system, and another is lossless compression system. Recently, due to the improvement of storing space, networking speed, and people ask for better sound quality, the lossless compression system becomes more valuable. Although the compression ratio for the lossless compression system is not as good as the compression ratio for the loss compression system, the lossless compression system focuses more on presenting the original sound to the audience. This is also the major reason that the lossless compression system is the most popular topic now. The basic structure of ALS coding uses original signal as Linear Prediction, and it uses the relationship between signals and signals as Prediction Signal. Then, the structure uses the original signal and the difference of Prediction Signal as Prediction Error. The Prediction Error is the required input signal for Entropy Coding. In MPEG-4 standard, there are two kinds of Entropy Coding: Golomb-Rice Code and Arithmetic Code. This thesis is going to talk about the relationship between open design threshold and the two kinds of Entropy Coding in the MPEG-4 ALS standard. With proper adjusted threshold, Entropy Coding can function the most efficiently, and it can find out the balance between compression ratio and compression time. The inappropriate threshold will cause against the result. We simulate the results in Chapter 4, the entropy coding method that we proposed will have better compression ratio than inappropriate adjustment threshold.