Data Compression For Multimedia Computing

• Main Author: Mohd Tarmizi, Musa
• Format: Thesis
• Language: English; English
• Published: 1995
• Subjects: QA76 Computer software
• Online Access: http://etd.uum.edu.my/721/; https://ueaeprints.uea.ac.uk

This is a library based study on data compression for multimedia computing. Multimedia information needs a large storage capacity as they contain vast amount of data. This would mean multimedia information would be out of reach of most computer users as their PCs would not be able to store the enormous amount of data accumulated on such programs. However, it is not necessary to keep these data in its original form as there are techniques that could compressed multimedia data to a more manageable level. Therefore, the main objective of this study is to provide information on the availability of compression techniques that would enable PC users the opportunity to use such programs. The review of related literature reveals that there are two basic compression techniques available - lossless and lossy. Under the lossless technique, the Huffman Coding, Arithmetic Coding and Lempel-Ziv Welch Coding are discussed. On the other hand, the Predictive, Frequency Oriented and Importance Oriented techniques are discussed under the lossy technique. Besides these two main techniques, Hybrid techniques such as the JPEG, MPEG and Px64 are also discussed. In order to bind the discussion between compression and storage media, a description of popular storage media such as magnetic disk storage and optical disc storage are also included. Although the data are of secondary source, the writer uses a formula derived from Howard and Vitter (1992) to measure compression efficiency. Based on the data collection and analysis it is found that different types of data (text, audio, video etc.)should be compressed using different techniques in order to obtain the ideal compression ratio and quality. Although the writer believes that the secondary data obtained is sufficient to show the best compression techniques for the different types of multimedia data, he also believes that real experiment using real data, software application and hardware would give better and more precise results.