MSDL-S Representation of the MPEG-2 Video IS

Yihan Fang and Alexandros Eleftheriadis
Department of Electrical Engineering, Columbia University

Contribution M1138, 35th MPEG Meeting, Tampere, Finland, July 1996

Introduction

This document is the complete MSDL-S representation of the bitstream syntax of the MPEG-2 Video International Standard [1]. The definition of MSDL-S used is as given in [2], with the addition of parameter types as detailed in [3]. For readability, we adopted the typographic convention of setting MSDL-S keywords in italics and bitstream elements in boldface.

The purpose of this document is to demonstrate, using a very familiar and non-trivial example, how MSDL-S can be used to express complex bitstream structures. Note that the Video_Sequence class has to be subclassed from the appropriate MSDL Architecture superclass. In order to obtain a complete MSDL-based specification of a functioning decoder, one only has to extend the class definitions to also include methods.

For simulation purposes, this MSDL-S file together with the associated methods (written in regular C++ or Java) can be translated to an equivalent C++ or Java one, in which bitstream parsing is handled within the object constructor(s). In addition to being important for testing and validating MSDL-S, this is also useful when writing decoder simulation software, as it automatically takes care of bitstream parsing. Essentially, the definition of the bitstream in MSDL-S can be directly used as part of a software implementation of its decoder. Such a translator is currently under implementation.

In an actual MPEG-4 environment, or its simulation, the bitstream specification is translated (by the same translator) to a binary format, and is downloaded to the decoder. Note that this downloading capability can be applied to just the syntax; in other words, the method declarations and definitions (tools) need not change. This allows flexibility in modifying the syntax even in Level 1 MPEG-4 systems.

PostScript (34 KB)
Word (31 KB)