Graduate Engineering Seminar
May 16th, 2002
ENGR 326, 4:10 PM

Basic Compression Techniques in MPEG-4

Gunnar Hovden
Ph.D. Candidate
Department of Computer Engineering
Santa Clara University

Abstract

Compression standards for audio and video are getting increasingly important as new products are developed. Products like DVD (Digital Versatile Disk), streaming video over Internet, HDTV (High Definition TV) and MOD (Movie On Demand) would not be possible without data compression.

The most videly used standards for video compression today is MPEG-1, MPEG-2 and MPEG-4, all defined by the Motion Picture Experts Group (MPEG). MPEG-1 and MPEG-2 are often used for video and audio (often referred to as MP3) on Internet. MPEG-2 is used for HDTV. Microsoft's new ASF (Advanced Streaming Format) is based on MPEG-4 (and myriads of other formats to be (somewhat) backward compatible).

Although MPEG-4 is the newest and by far the most advanced standard, it shares many basic compression techniques with its older cousins MPEG-1 and MPEG-2.

In this talk we will cover the most basic compression techniques and some more advanced techniques in MPEG-4, including DCT (Discrete Cosine Transform), quantization, run-length-code, motion compensation, arbitrary shape coding and sprites.