Joint Source-Channel Coding of Discrete-Time Signals with Continuous Amplitudes (Communications and Signal Processing)

✍ Scribed by Norbert Goertz

• Publisher: Imperial College Press
• Year: 2007
• Tongue: English
• Leaves: 207
• Category: Library

✦ Synopsis

This book provides the first comprehensive and easy-to-read discussion of joint source-channel encoding and decoding for source signals with continuous amplitudes. It is a state-of-the-art presentation of this exciting, thriving field of research, making pioneering contributions to the new concept of source-adaptive modulation. The book starts with the basic theory and the motivation for a joint realization of source and channel coding. Specialized chapters deal with practically relevant scenarios such as iterative source-channel decoding and its optimization for a given encoder, and also improved encoder designs by channel-adaptive quantization or source-adaptive modulation.Although Information Theory is not the main topic of the book - in fact, the concept of joint source-channel coding is contradictory to the classical system design motivated by a questionable practical interpretation of the separation theorem - this theory still provides the ultimate performance limits for any practical system, whether it uses joint source-channel coding or not. Therefore, the theoretical limits are presented in a self-contained appendix, which is a useful reference also for those not directly interested in the main topic of this book.

✦ Table of Contents

Contents......Page 10
Preface......Page 8
1. Introduction......Page 14
2.1 System Model......Page 18
2.1.1 Channel......Page 19
2.2 System Distortion......Page 20
2.4 Optimal Encoder......Page 22
2.5.2.1 System Example......Page 24
2.5.2.2 Comparison with Results from Information Theory......Page 26
2.5.3 Channels with Binary Input: Channel-Optimized Vector Quantization......Page 27
2.5.3.1 Binary Symmetric Channel (BSC)......Page 31
2.5.3.2 Channels with Binary Input but with Real Output......Page 32
2.5.3.3 Optimization of the Encoder Mapping: Codebook Training......Page 33
2.5.3.4 Some Simulation Results for COVQ......Page 35
2.6.1 Systems for Multimedia Transmission......Page 37
2.6.2 Separation of Source and Channel Coding......Page 38
2.6.2.1 Channel Codes have a Non-Zero Residual Bit Error Rate......Page 39
2.6.2.2 Source Encoder Output Bits Contain Redundancies......Page 41
2.6.2.3 It Matters, which Bit is in Error......Page 42
2.6.3.1 Unequal Error Protection and Error Concealment......Page 43
2.6.3.2 Source-Controlled Channel Decoding......Page 44
2.6.3.3 Estimation-Based Source Decoding......Page 45
2.6.3.4 Iterative Source-Channel Decoding......Page 46
2.6.4 Approaches to Joint Source-Channel Encoding......Page 47
3.1 Introduction and System Model......Page 50
3.2 Near Optimum Joint Source-Channel Decoding......Page 53
3.2.1 Specialization and Generalization......Page 58
3.3.1 Principle and Derivation......Page 59
3.3.2 Efficient Implementation of ISCD by L-values......Page 62
3.3.3 Simulation Results for ISCD......Page 64
3.4.1 Basic Considerations......Page 68
3.4.2 Optimization by Binary Switching......Page 71
3.4.3 Simulation Results with Optimized Bit Mappings......Page 74
3.5 Conclusions......Page 78
4.1 Introduction......Page 80
4.2 Memory and Complexity Issues for Vector Quantization (VQ) and Channel-Optimized VQ......Page 81
4.3.1 Basic Principle......Page 85
4.3.2 Optimization of CASVQ......Page 87
4.3.3 Complexity and Memory Requirements of CASVQ for Transmission over Time-Varying Channels......Page 89
4.4 Simulation Results......Page 91
4.5 Conclusions......Page 95
5.1 Introduction......Page 98
5.2 System Model......Page 100
5.3 Optimal Decoder for a Given Index Assignment......Page 101
5.4 Quality Criterion for the Index Assignments......Page 103
5.5.2 Index Optimization by the Binary Switching Algorithm for a System with a Single Description......Page 105
5.5.3 Binary Switching for Multiple Descriptions......Page 106
5.6 Simulation Results......Page 107
5.7 Conclusions......Page 110
6.1 Introduction......Page 112
6.2 Conventional System Model......Page 113
6.2.1 Conventional Hard-Decision Receiver......Page 115
6.2.2 Conventional Soft-Decision Receiver......Page 116
6.3 Principle of Source-Adaptive Modulation (SAM)......Page 117
6.4.1 Derivation of the Optimal Solution......Page 119
6.4.2 Test-Point Method......Page 126
6.4.3 Analytical Approximation......Page 127
6.4.4 Simulation Results......Page 129
6.5.1 Discussion of Potential Signal-Point Locations......Page 135
6.5.2 Simulation Results for SAM with QAM .......Page 139
6.6 Conclusions......Page 141
7.1 Introduction......Page 142
7.2 System Model......Page 143
7.3 Principle of Source-Adaptive Power Allocation......Page 144
7.5 Simulation Results......Page 148
7.6 Conclusions......Page 150
8. Concluding Remarks......Page 152
A.1 Preliminary Remarks......Page 154
A.2.1.1 Memoryless Gaussian Source......Page 155
A.2.1.2 Memoryless Non-Gaussian Sources......Page 156
A.2.2.1 SNR-Values for Optimal Scalar Quantization......Page 158
A.2.2.2 Comparison of the DRF with Optimal Scalar and Vector Quantization......Page 160
A.2.3 Gaussian Sources with Memory......Page 161
A.2.3.1 Simplification for High Rate......Page 162
A.2.3.2 Simplification for High Rate and a Linearly Filtered Gaus- sian Source......Page 163
A.2.3.3 Quality of the High-Rate Approximation......Page 165
A.3.1 Binary Symmetric Channel (BSC)......Page 166
A.3.3 Additive White Gaussian Noise (AWGN) Channel......Page 168
A.3.3.1 Discrete-Time AWGN Channel......Page 169
A.3.3.2 Continuous-Time Band-Limited AWGN Channel......Page 170
A.3.3.3 Discrete-Time Two-Dimensional AWGN Channel......Page 174
A.3.3.4 Equivalent Power-Normalized Channel Model for Discrete Input Alphabets......Page 175
A.3.3.6 Discrete-Time AWGN Channel with Discrete Input......Page 178
A.3.3.7 Discrete Time Two-Dimensional AWGN Channel with Discrete Inputs......Page 179
A.3.3.8 Some Capacity-Curves for M-PSK......Page 180
A.4 Performance Limits for the Transmission of Sources with Continuous-Values over Noisy Channels......Page 181
A.4.1 Gaussian Source/AWGN Channel and BSC......Page 182
A.4.2 Correlated Gaussian Source/AWGN Channel with Binary Input......Page 183
A.4.3 Gaussian Source/Two-Dimensional AWGN Channel with PSK Input......Page 185
Appendix B Optimal Decoder for a Given Encoder......Page 188
Appendix C Symbol Error Probabilities for M-PSK......Page 190
Appendix D Derivative of the Expected Distortion for SAM......Page 192
List of Symbols......Page 194
List of Acronyms......Page 198
Bibliography......Page 200
Index......Page 206