Original TM-5 Table of Contents
1 INTRODUCTION 6
2 GENERAL CODEC OUTLINE 7
2.1 Arithmetic Precision 7
2.2 Coder block diagram 8
2.3 Profiles 10
2.3.1 MAIN profile 11
2.3.2 Hierarchical profile 11
2.3.3 Professional profile 12
3 SOURCE FORMATS 13
3.1 Input Formats 13
3.2 Definition of fields and frames 14
3.3 Conversion of CCIR 601 to the Input formats 15
3.3.1 Conversion of CCIR 601 to the 4:2:0 format 15
3.3.2 Conversion of CCIR 601 to SIF 16
3.3.3. Conversion of CCIR 601 to SIF Odd and SIF Even 16
3.3.4 Conversion of CCIR 601 to HHR 17
3.3.5 Conversion of CCIR 601 to SIF Interlaced (SIF-I) 18
3.4 Conversion of the Input Formats to CCIR 601 18
3.4.1 Conversion of the 4:2:0 Format to CCIR 601 18
3.4.2 Conversion of SIF to CCIR 601 19
3.4.3 Conversion of SIF Odd and SIF Even to CCIR 601 20
3.4.4 Conversion of HHR to CCIR 601 20
3.4.5 Conversion of SIF interlaced to CCIR 601 21
3.5 Down conversion from interlaced to interlaced 21
3.6 Upconversion from interlaced to interlaced for display purposes 22
4 LAYERED STRUCTURE OF VIDEO DATA 24
4.1 Sequence 24
4.2 Group of pictures 24
4.3 Picture 24
4.4 Macro block Slice 25
4.5 Macroblock 25
4.6 Block 26
4.6.2 Block in harmonised scalable solution, for low level decoders 26
5 MOTION ESTIMATION AND COMPENSATION 27
5.1 Motion Vector Estimation 27
5.1.1 Full Search 27
5.1.2 Half pel search 28
5.1.3 Motion estimation for Special prediction mode 28
5.2 Motion Compensation 29
5.2.1 Frame Motion Compensation 30
5.2.2 Field Motion Compensation 30
5.3 Special prediction mode 31
5.3.1. Overview of Special Prediction mode 31
5.3.2. Specification of Dual-prime vectors 32
5.3.3. Temporal Scaling of the Field Motion Vector 32
5.3.4. Prediction of Chrominance Blocks 33
6 MODES AND MODE SELECTION 34
6.1 Picture types 34
6.2 Macroblock type decision 34
6.2.1 Modification of Decision for Field-based Prediction 34
7 TRANSFORMATION AND QUANTIZATION 36
7.1 Quantization of Intra Macroblocks 36
7.1.1 DC Coefficients 36
7.1.2 AC Coefficients 36
7.2 Quantization Non Intra Macroblocks 37
7.3 Dequantization 37
8 CODING 38
8.1 Macroblock Addressing 38
8.2 Macroblock Type 39
8.2.1 Compatible Prediction 39
8.3 Motion Vectors 39
8.3.1 Prediction of Motion Vectors for special prediction modes 41
8.4 Coded Block Pattern 42
8.5 Intra picture Coefficient Coding 42
8.5.1 DC Prediction 42
8.5.2 AC Coefficients 43
8.6 Non-Intrapicture Coefficient Coding 43
8.6.1 Intra blocks 43
8.6.2 Non intra blocks 43
8.7 Coding of Transform Coefficients 44
9 VIDEO MULTIPLEX CODER 45
9.1 Sequence Frequency Extension 45
9.2 Sequence spatial extension 47
9.3 Picture frequency extension 48
9.4 Picture spatial extension 49
9.5 Scaled Slice Layer 51
9.6 Slave Macroblock Layer 52
9.7 Scaled Macroblock Layer 53
9.8 Spatial Macroblock layer 55
9.8 CBP 57
9.8 Scaled Block Layer 58
9.9 Slave Block Layer 59
9.10 SNR Block layer 60
10 RATE CONTROL AND QUANTIZATION CONTROL 61
Step 1 - Bit Allocation 61
Complexity estimation 61
Picture Target Setting 61
Step 2 - Rate Control 63
Step 3 - Adaptive Quantization 64
Known Limitations 65
APPENDIX A: DISCRETE COSINE TRANSFORM (DCT) 66
APPENDIX B: VARIABLE LENGTH CODE TABLES 67
Introduction 67
B.1 Macroblock Addressing 67
B.2 Macroblock Type and Compatible Macroblock Type 67
B.3 Macroblock Pattern 69
B.4 Motion Vectors 70
B.5 DCT Coefficients 70
APPENDIX C : VIDEO BUFFER VERIFIER 71
APPENDIX D: FREQUENCY DOMAIN SCALABILITY EXTENSION 72
D.1 INTRODUCTION 72
D.2 LAYERED STRUCTURE OF VIDEO DATA AND MULTIPLEXING OF FREQUENCY SCALES 75
D.3 MOTION ESTIMATION AND COMPENSATION 76
D.4 MODES AND MODE SELECTION 77
D.4.1 No Scalable Side Information 77
D.4.2 Scalable Side Information 77
D.4.3 Motion Refinement 78
D.5 INTER-SCALE DCT COEFFICIENT PREDICTION 78
D.6 TRANSFORMATION AND QUANTIZATION 78
D.6.1 Transformation 78
D.6.2 Upward Coefficient Prediction and Quantization 78
D.6.3 BANDWIDTH CONTROL OF RESOLUTION LAYERS 79
D.7 DCT COEFFICIENT CODING 79
D.8 VIDEO MULTIPLEX CODER 79
D.9. RATE CONT ROL AND QUANTIZATION CONTROL 79
D.10 Drift Correction 80
D.11 Interlaced quality on lower resolution layers 81
APPENDIX F: CELL LOSS EXPERIMENTS 82
F.1 Cell loss 82
F.1.1 Bitstream specification 82
F.1.2 Calculation of cell loss probabilities 82
F.1.3 Calculation of mean cell loss rate 83
F.1.4 Calculation of mean burst of consecutive cells lost 83
F.1.5 Calculation of cell loss probabilities 84
F.1.6 Simulation of cell loss 84
F.1.7 Random number generation 85
F.2 Parameters 85
F.3 Concealment techniques 86
F.4 AC-Leaky Prediction 86
F.4.1 Description of AC Leak 86
F.4.2 Core Experiments 87
F.4.3 Syntax 88
F.5 Data Partitioning versus 1-layer cell loss experiment 88
APPENDIX G: COMPATIBILITY AND SPATIAL SCALABILITY 90
G.1 Spatial/Temporal Prediction 90
G.1.1 Introduction 90
G.1.2 Detail 91
G.2 Upsampling for Prediction 92
G.2.1 Introduction 92
G.2.2 Interlace to Interlace Up-sampling 92
G.2.3 Progressive to Interlace Up-sampling 93
G.3 Coding of spatial scalable/compatible macroblocks 93
G.3.2 Summary of previous scalability results 94
G.3.2 Remaining spatial scalability experiments. 94
G.4 SNR Scalability 95
G.4.1 Experiment on SNR-scalability: 97
G.7 Chrominance scalability 98
APPENDIX H: LOW DELAY CODING 101
H.1 Simulation guidelines for low delay coding. 101
H.1.1 Coding structure. 101
H.1.2 Handling of scene change to maintain low delay. 101
H.1.3 How to handle the first picture using forced intra slice 106
H.1.4 Definition of intra slice/column coding. 106
H.1.5 Rate control. 107
H.1.6 Influence of leaky prediction on low delay coding. 108
APPENDIX I : FREQUENCY DOMAIN SCALABILITY CORE EXPERIMENTS 109
I.4: Scalable Side Information 109
I.4.1. Application 109
I.4.2. Experiment details 109
I.4.3. Syntax extensions 109
I.4.4. Coding 109
I.12 Single loop decoder structure 109
I.14 Comparison of different encoder filters for inter-layer prediction. 111
APPENDIX J: HARMONISED HYBRID SCALABILITY 113
J.1 Introduction 113
J.2 Examples of the proposals currently on the table 113
J.3 The concise generic decoder solution 114
J.4 Decoder conclusion 116
J.5 The encoder 117
J.6 Encoder conclusion 119
J.8 Upconversion prediction filter 120
J.9 Harmonisation of syntax 120
APPENDIX K: FAST FORWARD AND FAST REVERSE MODES 121
K.1 Concept of FF/FR mode in DSMs 121
K.2 Intra_slice approach 121
K.3 Requirements for syntax extension 121
K.4 Data Partitioning approach 122
K.4.1 FF/FR Operation by use of data partition 122
K.4.2 VCR FF/FR Bitstream Delivery 122
K.4.3 Syntax for FF/FR by Data Partition 122
APPENDIX L: DATA PARTITIONING 125
L.1. Syntax for Data Partitioning 125
L.1.1 Sequence Layer Syntax 125
L.2 Description 125
L.3 Rate control 127
L.4 Experiments 127
APPENDIX Q: QUANTISATION 128
Q 5. EXPERIMENT OF NON-8 x 8 DCT (Revised at Sydney) 128
Back to Contents