?? （3）mpeg-4 systems elementary stream management.htm
字號:
<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<!-- saved from url=(0101)http://3c.nii.org.tw/3c/silicon/embedded/MPEG/MPEG-4%20Systems%20Elementary%20Stream%20Management.htm -->
<!-- saved from url=(0083)http://leonardo.telecomitalialab.com/icjfiles/mpeg-4_si/4-ESM_paper/4-ESM_paper.htm --><HTML><HEAD><TITLE>MPEG-4 Systems: Elementary Stream Management</TITLE>
<META http-equiv=Content-Type content="text/html; charset=windows-1252">
<META content="MSHTML 6.00.2800.1106" name=GENERATOR></HEAD>
<BODY><FONT size=5><B>
<P align=center>MPEG-4 Systems: Elementary Stream Management</P></B></FONT><FONT
size=2>
<P align=center>C. Herpel<SUP>a</SUP>,* A.
Eleftheriadis<SUP>b</SUP></P></FONT><SUP></SUP>
<P align=center><FONT size=1><SUP>a</SUP></FONT><FONT size=1><I>THOMSON
multimedia, Karl-Wiechert-Allee 74, 30625 Hannover, Germany</I></FONT></P><FONT
size=1>
<P align=center><SUP>b</SUP><I>Dept. of Electrical Engineering, Columbia
University, 500 West 120th Street, MC 4712, New York, NY 10027, USA</I></P><I>
<P align=center>*Corresponding author. Email:
herpelc@thmulti.com</P></I></FONT><FONT size=2><B>
<P align=justify> </P>
<P>Abstract </P>
<P align=justify>We describe the Elementary Stream Management (ESM) facilities
provided by MPEG-4 Systems. Within the extensive set of tools defined by MPEG-4,
the ESM tools play a critical role in joining several building blocks together.
ESM provides a dual to the scene description language (BIFS) in that it links
the streaming resources of a presentation to the scene. We also describe the
synchronization functionality as well as the system decoder model that defines
the timing behavior and buffer resource management of MPEG-4 receivers. The
paper concludes with considerations on data packaging in underlying delivery
layer protocols and a description of the MPEG-4 content access procedure.</P>
<P align=justify>Keywords: MPEG-4; Object descriptor; Synchronization;
Multiplex; Content access; Content delivery</P></B>
<OL><B>
<LI><A name=_Toc442152517>Introduction</A> </B>
<P align=justify>MPEG-4 is the first standard that views multimedia content as
a set of audio-visual objects that are presented, manipulated and transported
individually. This is achieved by a set of tools defined in several parts of
the standard. First, the media compression schemes defined in the Visual [2]
and Audio [3] parts of the MPEG-4 specification are object oriented so that
they can represent elemental audio-visual entities (e.g., arbitrarily shaped
visual objects). The overall architecture of MPEG-4 that provides the means
for the combined use of these elements is defined in MPEG-4 Systems [1]. The
<I>scene description </I>language, called Binary Format for Scenes (BIFS),
discussed in a companion paper in this Special Issue [19], expresses how
individual audio-visual objects are to be composed together for presentation
on the user抯 screen and speakers. In addition, a <I>stream description </I>of
the corresponding elementary streams that convey this data is needed. The term
<I>elementary stream management </I>is used to refer to the entire set of
functionalities needed to describe, express relations between, and effect
synchronization among such data streams. In addition to describing these
tools, we also provide some considerations about data packaging and
multiplexing on the underlying delivery infrastructure.</P>
<P align=justify>The organization of this paper is as follows. First, some
terms that are used throughout the MPEG-4 specification as well as in this
paper are briefly introduced and discussed. Then, we describe the object
description framework that allows the identification and characterization of a
related set of elementary streams. It serves as a guide through the
potentially large set of streams that may belong to a sophisticated MPEG-4
presentation. This guide must be first consulted before any of the content
conveyed within the streams can be presented. Next, the issue of
synchronization is discussed. Since many streams are quasi-concurrently
generated and consumed, time stamping is employed here, similar to MPEG-2, to
align them on the temporal axis. In this context, the system decoder model is
introduced; it describes in an abstract way how the notion of time is defined
in MPEG-4 and how some of the buffer memory in the system is modeled and
managed. Finally, some ongoing activities will be reported concerning delivery
of MPEG-4 data. The standard does not specify its own transport layer, but is
intended to be used with different delivery systems (including digital TV
broadcasting, the Internet, and stored files). Delivery functionality in
general is described in depth in a companion paper in this Special Issue [10].
Equipped with these building blocks, we finally present a walkthrough of the
process of accessing MPEG-4 content consisting of many different elementary
streams, possibly even originating from different locations.</P>
<LI><B><A name=_Toc442152518>Fundamental concepts</A></B>
<P align=justify>Audio or visual entities that participate as individual
elements in a scene are termed <I>audio-visual objects</I>. Such objects can
be either natural or synthetic. Time-variant data for each natural object is
conveyed separately, in different "channels". Synthetic objects may be
generated with the graphics and synthesized sound operations provided by BIFS.
BIFS is actually more than just a scene description language, in that it
integrates both natural and synthetic objects in the same composition space.
Some objects may therefore be fully described within the scene description
stream itself. Objects may also be animated using the BIFS-Anim tool. In that
case, the corresponding coded information will as well be conveyed in its own
channel. As a result, the term audio-visual object ?? 快捷鍵說明
復制代碼
Ctrl + C
搜索代碼
Ctrl + F
全屏模式
F11
切換主題
Ctrl + Shift + D
顯示快捷鍵
?
增大字號
Ctrl + =
減小字號
Ctrl + -