344 Pages
English

System and cross-layer design for mobile video transmission [Elektronische Ressource] / Thomas Stockhammer

Gain access to the library to view online
Learn more

Subjects

Informations

Published by
Published 01 January 2008
Reads 9
Language English
Document size 9 MB

¨ ¨TECHNISCHE UNIVERSITAT MUNCHEN
Lehrstuhl fu¨r Nachrichtentechnik
System and Cross–Layer Design
for
Mobile Video Transmission
Thomas Stockhammer
Vollsta¨ndiger Abdruck der von der Fakulta¨t fu¨r Elektrotechnik und Informationstechnik der Tech-
nischen Universita¨t Mu¨nchen zur Erlangung des akademischen Grades eines
Doktor–Ingenieurs
genehmigten Dissertation.
Vorsitzender: Univ.–Prof. Dr.–Ing. habil. Gerhard Rigoll
Pru¨fer der Dissertation:
1. Univ.–Prof. Dr.–Ing., Dr.–Ing. E.h. Joachim Hagenauer, i.R.
2. Univ.–Prof. Dr.–Ing. Eckehard Steinbach
Die Dissertation wurde am 19.06.2008 bei der Technischen Universita¨t Mu¨nchen
eingereicht und durch die Fakulta¨t fu¨r Elektrotechnik und Informationstechnik
am 24.09.2008 angenommen.Preface
I have made this letter longer than usual, only
because I have not had the time to make it shorter.
Blaise Pascal, (1623-1662), Lettres provinciales.
This thesis has been written during my time as a research assistant at the Institute for Commu-
nications Engineering (Lehrstuhl fu¨r Nachrichtentechnik, LNT). Many people at the Institute and
elsewhere have contributed to this work. Firstly, I would like to thank my supervisor Professor
Dr.-Ing. Dr.-Ing. E.h. Joachim Hagenauer for providing me with the opportunity to work at his
institute, for all his support of my thesis and for the patience to await its completion. I am also
very grateful to Professor Dr.-Ing. Eckehard Steinbach for acting as a co-supervisor and Professor
Dr.-Ing. habil. Gerhard Rigoll for heading the commission. At the same time, many thanks also to
Professor Antonio Ortega from the University of Southern California (USC) for accepting a place
on the commission - despite logistical reasons that made it impossible in the end and to Professor
Ken Zeger from the University of California, San Diego (UCSD) for making possible my visit to
the UCSD during spring 2000. Also many thanks to all colleagues and partners for exciting col-
laborations during my standardization work for the ITU-T, MPEG, IETF, 3GPP, DVB, and JVT.
In particular, I am grateful to Professor Dr. Thomas Wiegand and the other video experts from the
Fraunhofer HHI in Berlin for introducing me to the secrets and mysteries of video coding.
The LNT has always been a place with an enjoyable atmosphere and I was surrounded by ex-
traordinary people. The collaboration with many diploma and master students made the time at the
Institute a memorable experience, and many of the colleagues and students have become friends. I
would like to mention in this respect Professor Dr. Gu¨nter So¨der, Dr. Alexander Seeger, Dr. Rainer
Bauer, Dr. Christian Weiß, Dr. Volker Franz, Thomas Effern, Daniel Pfeifer and Houda Kamoun.
Today, I am delighted to collaborate with some of my former colleagues and partners from the
LNT, in particular Professor Hagenauer, Dr. Ingo Viering, Gu¨nther Liebl and Christian Buchner.
I am very grateful to all my colleagues at Nomor Research for their courage to found and join a
company that is heavily based on research work started at the LNT. Also, my thanks to Dr. Michael
Luby and all other colleagues at Digital Fountain as well as all my friends at home and all over the
world. Thank you all for your support and for encouraging me to complete this thesis.
While generating this work, I became friends with Dr. Michael Mecking and Dr. Hrvoje Jenkacˇ.
They truly proved their friendship by not leaving anything undone to incite me on the final stages
of this work and proofread faster than I could type. Micki, H, thank you so much. Finally, this
thesis would not have been possible without the continuous support of my parents and family, I
am deeply grateful for providing me with all the opportunities. I apologize to Lisa-Maria and
Maximilian for spending so much time on this work and not with them. Last, but not least, I
deeply thank the rock of our family, my wife Christina, for her understanding and love during the
generation of this thesis.
Bergen, November 2008 THOMAS STOCKHAMMERContents v
Contents
1 Introduction 1
2 Technical Preliminaries: Networked Video 7
2.1 System Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.1.1 Video Transmission System . . . . . . . . . . . . . . . . . . . . . . . . . 7
2.1.2 Video Quality Measures . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
2.1.3 Information Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.1.4 Timing, Delays, and Buffering in Video Transmission . . . . . . . . . . . 14
2.1.5 Video Applications and Services: Requirements and Objectives . . . . . . 17
2.2 Transmission Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.2.1 Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.2.2 Internet Principles and Characteristics . . . . . . . . . . . . . . . . . . . . 21
2.2.3 Multimedia Transmission over IP . . . . . . . . . . . . . . . . . . . . . . 22
2.3 Multimedia Transmission in Packet-Radio Networks . . . . . . . . . . . . . . . . 25
2.3.1 Background and History . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.3.2 Integration of Multimedia Services . . . . . . . . . . . . . . . . . . . . . 28
2.4 System Abstractions and Channel Models . . . . . . . . . . . . . . . . . . . . . . 30
2.4.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
2.4.2 The Mobile Radio Channel . . . . . . . . . . . . . . . . . . . . . . . . . . 32
2.4.3 Signaling and Discrete–Time Models . . . . . . . . . . . . . . . . . . . . 35
2.4.4 Link-Layer Channel Models . . . . . . . . . . . . . . . . . . . . . . . . . 41
2.4.5 Application Layer Channel Characteristics and Models . . . . . . . . . . . 45
2.5 Motivation for the Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
3 Source and Video Coding Toolbox 53
3.1 Source Coding Preliminaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
3.1.1 Theoretical Background . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
3.1.2 Rate and Quality Adaptive Source Coders . . . . . . . . . . . . . . . . . . 56
3.1.3 Scalable and Progressive Source Coding . . . . . . . . . . . . . . . . . . . 58
3.1.4 Example: Still Image Progressive Source Coding . . . . . . . . . . . . . . 61
3.2 Practical Issues in Video Source Coding . . . . . . . . . . . . . . . . . . . . . . . 62
3.2.1 Video Source Coding Basics . . . . . . . . . . . . . . . . . . . . . . . . . 62
3.2.2 Video Coding Standards . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
3.2.3 H.264/AVC Compression Tools . . . . . . . . . . . . . . . . . . . . . . . 68
3.2.4 H.264/AVC Functionalities . . . . . . . . . . . . . . . . . . . . . . . . . . 70
3.2.5 Operational Encoder and Rate Control . . . . . . . . . . . . . . . . . . . . 72
3.3 H.264/AVC in Network Environment . . . . . . . . . . . . . . . . . . . . . . . . . 78vi Contents
3.3.1 Network Abstraction Layer . . . . . . . . . . . . . . . . . . . . . . . . . . 78
3.3.2 H.264/AVC VCL in Error-Prone Environments . . . . . . . . . . . . . . . 80
3.3.3 H.264/AVC Error-Resilience Features . . . . . . . . . . . . . . . . . . . . 81
3.3.4 Stream Switching with H.264/AVC . . . . . . . . . . . . . . . . . . . . . 84
3.4 Scalable Video Coding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
3.4.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
3.4.2 Progressive Texture Video Coding . . . . . . . . . . . . . . . . . . . . . . 92
3.4.3 Rate Control and Reference Frame Layer Selection . . . . . . . . . . . . . 95
3.4.4 Group-of-Picture Interleaving . . . . . . . . . . . . . . . . . . . . . . . . 97
3.5 Performance of Video Coders . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
3.5.1 Variable Bit-rate Coding . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
3.5.2 Low-Delay Constant Bit-rate Coding . . . . . . . . . . . . . . . . . . . . 99
3.5.3 Unknown Transmission Bit-rate . . . . . . . . . . . . . . . . . . . . . . . 99
3.5.4 GOP Transmission Rate . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
3.6 Formalization of Pre-Encoded Data . . . . . . . . . . . . . . . . . . . . . . . . . 101
3.6.1 Abstraction of Progressively Coded Sources . . . . . . . . . . . . . . . . . 101
3.6.2 General Framework for Transmitting Pre-Encoded Data . . . . . . . . . . 103
3.6.3 Dependencies and Versions . . . . . . . . . . . . . . . . . . . . . . . . . . 103
3.6.4 Summary: Abstract Representation of Pre-encoded Video Data . . . . . . . 107
3.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
4 Reliable Transmission Toolbox 109
4.1 Error Protection Preliminaries . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
4.1.1 Theoretical Background . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
4.1.2 Rate-Compatible Punctured Channel Coding . . . . . . . . . . . . . . . . 113
4.1.3 Unequal Error Protection . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
4.1.4 Channel State Information and Power Control . . . . . . . . . . . . . . . . 117
4.1.5 Automatic Repeat Request Schemes . . . . . . . . . . . . . . . . . . . . . 118
4.2 Channel Coding and Decoding Methods . . . . . . . . . . . . . . . . . . . . . . . 122
4.2.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
4.2.2 Convolutional Codes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
4.2.3 Sequential Decoding of Convolutional Codes . . . . . . . . . . . . . . . . 130
4.3 Convolutional Codes in Systems with Fading . . . . . . . . . . . . . . . . . . . . 139
4.3.1 System Design and Radio Access for Block Fading Channels . . . . . . . . 139
4.3.2 Outage Probabilities for FEC . . . . . . . . . . . . . . . . . . . . . . . . . 140
4.4 Regressive UEP and the Far End Error Decoder . . . . . . . . . . . . . . . . . . . 142
4.4.1 System Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
4.4.2 Optimum Trellis-Based Solution . . . . . . . . . . . . . . . . . . . . . . . 145
4.4.3 Complexity-Reduced Computation of Reliability . . . . . . . . . . . . . . 147
4.4.4 Performance Comparison for Progressively Coded Sources . . . . . . . . . 153
4.5 Forward Error Correction for Packet-Lossy Channels . . . . . . . . . . . . . . . . 156
4.5.1 Introduction and Framework . . . . . . . . . . . . . . . . . . . . . . . . . 156
4.5.2 Code Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
4.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161Contents vii
5 Progressive Source Coding and Unequal Error Protection 163
5.1 System Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
5.2 Information-Theoretic Considerations . . . . . . . . . . . . . . . . . . . . . . . . 164
5.2.1 Problem Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 164
5.2.2 Assumptions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
5.2.3 Distortion for Fixed Channel Coding Rates . . . . . . . . . . . . . . . . . 167
5.2.4 Optimized Channel Coding Rates . . . . . . . . . . . . . . . . . . . . . . 168
5.2.5 Equal Transmission Block Size . . . . . . . . . . . . . . . . . . . . . . . 170
5.2.6 Equal Information Part Length . . . . . . . . . . . . . . . . . . . . . . . . 171
5.2.7 One Practical Example . . . . . . . . . . . . . . . . . . . . . . . . . . . . 172
5.2.8 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
5.3 Unequal Erasure Protection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175
5.3.1 Framework and Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . 175
5.3.2 Optimized Rate Allocation . . . . . . . . . . . . . . . . . . . . . . . . . . 177
5.4 Progressive Image Transmission over Fading Channels . . . . . . . . . . . . . . . 178
5.4.1 System Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 178
5.4.2 Optimized Rate Allocation . . . . . . . . . . . . . . . . . . . . . . . . . . 179
5.4.3 Selected Experimental Results . . . . . . . . . . . . . . . . . . . . . . . . 181
5.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
6 Video Transmission in Packet-Lossy Environment 187
6.1 H.264/AVC-Compliant End-to-End Systems . . . . . . . . . . . . . . . . . . . . . 187
6.1.1 Formalization of H.264/AVC Packetized Video . . . . . . . . . . . . . . . 187
6.1.2 Error Concealment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
6.1.3 Performance of Basic Error Resilience Tools . . . . . . . . . . . . . . . . 189
6.1.4 Operational Encoder Control in Error–Prone Environment . . . . . . . . . 192
6.1.5 Interactive Error Control . . . . . . . . . . . . . . . . . . . . . . . . . . . 200
6.1.6 Performance Results for Test Conditions . . . . . . . . . . . . . . . . . . 205
6.2 End-to-End Forward Error Correction for Low-Delay Applications . . . . . . . . . 207
6.3 FEC for Moderate-Delay Applications . . . . . . . . . . . . . . . . . . . . . . . . 210
6.3.1 Forward Error Correction with H.264/AVC . . . . . . . . . . . . . . . . . 210
6.3.2 PTVC with Unequal Erasure Protection . . . . . . . . . . . . . . . . . . . 212
6.4 Summary: Major Observations for Video Transmission over Packet Loss Channels 215
7 Video Services in UMTS-like Environments 219
7.1 System Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219
7.2 Experimental Results for Conversational Video . . . . . . . . . . . . . . . . . . . 221
7.3 Experimental Results for Moderate-Delay Applications . . . . . . . . . . . . . . . 225
7.4 System Design Guidelines . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 226
7.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 227
8 Mobile Conversational Video 229
8.1 Common Channel and Service Environment . . . . . . . . . . . . . . . . . . . . . 229
8.2 H.264/AVC-based Source-Channel Coding System . . . . . . . . . . . . . . . . . 232
8.2.1 System Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 232
8.2.2 Constant Code Rate FEC and H.264 Single Layer System . . . . . . . . . 235
8.2.3 Source-Adaptive Code Rate Selection . . . . . . . . . . . . . . . . . . . . 239
8.3 PTVC-based Source-Channel Coding System . . . . . . . . . . . . . . . . . . . . 244viii Contents
8.3.1 System Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244
8.3.2 Forward Error Correction . . . . . . . . . . . . . . . . . . . . . . . . . . 246
8.3.3 Non-Persistent ARQ Schemes . . . . . . . . . . . . . . . . . . . . . . . . 249
8.3.4 Channel-State Information at Transmitter . . . . . . . . . . . . . . . . . . 254
8.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259
9 Video Streaming over Variable Bit-rate Mobile Channels 263
9.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263
9.2 System Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265
9.3 A Framework for the Description of Mobile Links and Source Representation . . . 266
9.3.1 Media Encoding and Abstraction . . . . . . . . . . . . . . . . . . . . . . . 266
9.3.2 Abstract Channel Representation . . . . . . . . . . . . . . . . . . . . . . . 267
9.3.3 Simplified Description for EGPRS-like channels . . . . . . . . . . . . . . 267
9.4 Optimized Packet Scheduling and Bitstream Switching . . . . . . . . . . . . . . . 269
9.4.1 Transmitter Assumptions . . . . . . . . . . . . . . . . . . . . . . . . . . . 269
9.4.2 Periodic Update of Side Information at the Transmitter . . . . . . . . . . . 271
9.4.3 The Scheduling Process . . . . . . . . . . . . . . . . . . . . . . . . . . . 272
9.4.4 Implementation Aspects and Complexity Reduction . . . . . . . . . . . . 273
9.5 Experimental Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
9.5.1 Simulation Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
9.5.2 Local Decisions - Temporal Scalability . . . . . . . . . . . . . . . . . . . 279
9.5.3 Influence of Scheduler Options . . . . . . . . . . . . . . . . . . . . . . . . 279
9.5.4 System Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279
9.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 282
10 Summary and Outlook 285
A Outline of Proofs for Importance and Quality Definitions 295
A.1 Received Quality as Sum of Importances . . . . . . . . . . . . . . . . . . . . . . . 295
A.2 Expected Quality as Weighted Sum of Importance . . . . . . . . . . . . . . . . . . 297
B Acronyms 299
Bibliography 307
Supervised Diploma and Master Theses 331Kurzfassung
Diese Arbeit untersucht Systementwurfs- und -optimierungsverfahren fu¨r die Verwendung von
Videoapplikationen im Mobilfunk. Besonderer Wert wird dabei auf das Zusammenspiel und
¨die Optimierung von Funktionen auf verschiedenen Schichten des Ubertragungsmodells unter
Beru¨cksichtung der Bedingungen unterschiedlicher Applikationen gelegt. Basierend auf ver-
¨schiedenen Ubertragungsmodellen, die an aktuelle und zuku¨nftige Mobilfunksysteme angelehnt
¨sind, werden ada¨quate Ubertragungs- und Videocodierverfahren vorgestellt. Speziell wird der
mitentwickelte H.264/AVC Videocodierstandard sowie eine proprieta¨r entwickelte, skalierbare Er-
¨weiterung auf die Verwendbarkeit in Mobilfunk- und Internet-Ubertragungsumgebungen unter-
sucht. Hierzu werden neuartige Fehlerschutzverfahren entwickelt, die sich speziell auf die im
Mobilfunk zu erwartenden zeit- und ortsvariablen Empfangsbedingungen anpassen ko¨nnen und
ein Zusammenspiel mit den vorgeschlagenen Videocodier-Verfahren erlauben. Die Auswahl von
Optionen innerhalb der Videocodier- und Fehlerschutzverfahren sowie deren Zusammenspiel wird
durch qualita¨tsoptimierende Auswahlverfahren unterstu¨tzt. Die Arbeit wird begleitet von infor-
mationstheoretischen Betrachtungen, praktisch relevanten Systementwu¨rfen sowie ausfu¨hrlichen
Simulationsergebnissen, die die Vorzu¨ge der vorgeschlagenen Methoden im Vergleich zu ex-
istierenden Systemen belegen und quantifizieren.
Abstract
This work investigates system designs and optimizations for the application of video services in
mobile communication systems. Special focus is put on the cooperation and optimization of func-
tions in different layers of a transmission system taking into account the service requirements of
different video applications. Based on channel models that are derived from state-of-the-art and
emerging mobile communication systems, we introduce suitable transmission and video coding
methods. In particular, the co-developed H.264/AVC video coding standard as well as a pro-
prietary scalable extension are intensively analyzed for their applicability in mobile and Inter-
net communication environments. Innovative error protection tools are designed which enable
adaptation to the varying reception conditions in mobile communication environments and enable
the cross-layer optimization with the proposed video coding tools. The selection of video cod-
ing and error protection options as well as their cooperation is supported by the development of
quality-optimizing selection and rate allocation schemes. The findings in this work are verified by
information-theoretic justifications, practically relevant system designs, as well as extensive sim-
ulation results. The benefits of the proposed methods with respect to existing systems are shown
and the realizable gains are quantified.