Author

Daniel Howard

Abstract

As computer use and capabilities have grown, people have become more interested in being able to create and access varies types of multimedia content. The MPEG video compression technique provides a method for compressing video content down to a size that computers and networks can handle. To properly make use of this algorithm it is necessary to understand the trade-offs that exist when choosing among the various options of the MPEG algorithm. Background information on the MPEG-1, MPEG-2 and MPEG-4 algorithms is presented. This thesis then provides an understanding of the trade-offs of applying different compression and decompression options of the MPEG-1 algorithm on various types of video streams. This allows recommendations on which options should be used for specific categories of video sequences to be made. The performance of an existing implementation of the MPEG compression and decompression algorithm is analysed to determine these resulting trade-offs . Various types of video sequences are used to observe the results of changing the various parameters of the algorithm. Some of the parameters that are investigated include the percentage of the I (only spatially compressed), P (forward predicted), and B (bi-directionally predicted) frames in the compressed stream and the individual quantization of each of these frames. The results from each of the video sequences when these parameters are modified and analysed with respect to overall CR (compression ratio), play rate, average compression ratio of the I, P, and B frames separately, file percentages of the I, P, and B frame separately, and image quality. Image quality is measured subjectively using results obtained by polling a group of individuals who have observed the various video sequences. The main variables that are dependent on each other are: play rate, image quality, and compression rate. This resulting trade off analyses leads to statement on which types of parameter settings should be used in each of the various types of video sequences. In order to complete this thesis first a working understanding of the MPEG algorithm was obtained. The various video sequences used were collected. The test video streams derived from these base cases were then created and analysed. As part of this analysis phase, a group of individuals viewed and rated these video streams with respect to an original base case. A systematic approach for reporting the effect of changing the MPEG parameters on image quality, play rate, and compression ratio was determined. These results are then presented along with suggestions on when to use the various parameter options. Areas for further research are then discussed.

Library of Congress Subject Headings

Video compression--Standards--Evaluation.; Image compression--Standards--Evaluation

Publication Date

6-1-1997

Document Type

Thesis

Department, Program, or Center

Computer Engineering (KGCOE)

Advisor

Czernikowski, Roy

Advisor/Committee Member

Zeck, Norman

Advisor/Committee Member

Chang, Tony

Comments

Note: imported from RIT’s Digital Media Library running on DSpace to RIT Scholar Works. Physical copy available through RIT's The Wallace Library at: QA76.575 .H693 1997

Campus

RIT – Main Campus

Share

COinS