Voice and Audio Compression for Wireless Communications

Lajos Hanzo has been with the Department of Electronics and Computer Science, University of Southampton, UK, since 1986 where he holds the Chair in Telecommunications. He received his degree in electronics in 1976 and his doctorate in 1983. During his career in telecommunications he has held various research and academic posts in Hungary, Germany and the UK. Lajos is a Fellow of the IEEE and the IEE as well as an IEEE Distinguished Lecturer. He has co-authored 11 books on mobile radio communications, published over 500 research papers and been awarded a number of distinctions. Clare Somerville received M.Eng and Ph.D degrees in Electronic Engineering in 1995 and 1999, respectively, from the University of Southampton. From 1995 to 1998 she performed research into low bit rate speech coders for wireless communications. Following a spell of research with the Global Wireless Systems Research department, Bell Laboratories, Swindon, U.K, she moved to PicoChip in Bath, UK. Her current research involves real-time services over wireless networks with a packet air interface.Jason Woodard received an MA in Physics from the University of Oxford in 1991, then an MSc in Electronics and a PhD in Speech Coding from the University of Southampton in 1992 and 1995 respectively. From 1995 to 1998 he worked as a Research Fellow at the University of Southampton, researching error correction coding, especially turbo codes. In 1998 he joined the Wireless Technology Practice at PA Consulting Group in Cambridge. At the start of 1999 he was transferred from PA to UbiNetics Ltd, where he is responsible for the development and implementation of bit rate processing algorithms, including turbo codes, for 3rd Generation mobile communications products.

• About the Authors xxiOther Wiley and IEEE Press Books on Related Topics xxiiiPreface and Motivation xxvAcknowledgements xxxvI Speech Signals and Waveform Coding 11 Speech Signals and an Introduction to Speech Coding 31.1 Motivation of Speech Compression 31.2 Basic Characterisation of Speech Signals 41.3 Classification of Speech Codecs 81.4 Waveform Coding 111.5 Chapter Summary 282 Predictive Coding 292.1 Forward-Predictive Coding 292.2 DPCMCodec Schematic 302.3 Predictor Design 312.4 Adaptive One-word-memory Quantisation 392.5 DPCM Performance 402.6 Backward-adaptive Prediction 422.7 The 32 kbps G.721 ADPCM Codec 472.8 Subjective and Objective Speech Quality 532.9 Variable-rateG.726andEmbeddedG.727ADPC542.10 Rate-distortion in Predictive Coding 622.11 Chapter Summary 67II Analysis-by-Synthesis Coding 693 Analysis-by-Synthesis Principles 713.1 Motivation 713.2 Analysis-by-Synthesis Codec Structure 723.3 The Short-term Synthesis Filter 733.4 Long-term Prediction 763.5 Excitation Models 853.6 Adaptive Short-term and Long-term Post-Filtering 883.7 Lattice-based Linear Prediction 903.8 Chapter Summary 974 Speech Spectral Quantisation 994.1 Log-area Ratios 994.2 Line Spectral Frequencies 1034.3 Vector Quantisation of Spectral Parameters 1154.4 Spectral Quantisers for Wideband Speech Coding 1234.5 Chapter Summary 1385 Regular Pulse Excited Coding 1395.1 Theoretical Background 1395.2 The 13 kbps RPE-LTP GSM Speech Encoder 1465.3 The 13 kbps RPE-LTP GSM Speech Decoder 1515.4 Bit-sensitivity of the 13 kbps GSM RPE-LTP Codec 1535.5 Application Example: A Tool-box Based Speech Transceiver 1545.6 Chapter Summary 1576 Forward-Adaptive Code Excited Linear Prediction 1596.1 Background 1596.2 The Original CELP Approach 1606.3 Fixed Codebook Search 1636.4 CELP Excitation Models 1656.5 Optimisation of  the CELP Codec Parameters 1746.6 The Error Sensitivity of CELP Codecs 1926.7 Application Example: A Dual-mode 3.1 kBd Speech Transceiver 2046.8 Multi-slot PRMA Transceiver 2186.9 Chapter Summary 2237 Standard Speech Codecs 2257.1 Background 2257.2 The US DoD FS-1016 4.8 kbps CELP Codec 2257.3 The 7.95 kbps Pan-American Speech Codec – Known as IS-54 DAMPS Codec 2317.4 The 6.7 kbps Japanese Digital Cellular System’s Speech Codec 2357.5 The Qualcomm Variable Rate CELPCodec 2377.6 Japanese Half-rate Speech Codec 2457.7 The Half-rate GSM Speech Codec 2537.8 The 8 kbps G.729 Codec 2577.9 The Reduced Complexity G.729 Annex A Codec 2787.10 The 12.2 kbps Enhanced Full-rate GSM Speech Codec 2827.11 The Enhanced Full-rate 7.4 kbps IS-136 Speech Codec 2877.12 The ITU G.723.1 Dual-rate Codec 2927.13 Advanced Multirate JD-CDMA Transceiver 3027.14 Chapter Summary 3278 Backward-adaptive Code Excited Linear Prediction 3318.1 Introduction 3318.2 Motivation and Background 3318.3 Backward-adaptiveG728CodecSchematic 3348.4 Backward-adaptiveG728CodingAlgorithm 3368.5 Reduced-rate G728-like Codec: Variable-length Excitation Vector 3518.6 The Effects of Long-term Prediction 3548.7 Closed-loop Codebook Training 3598.8 Reduced-rate G728-like Codec: Constant-length Excitation Vector 3648.9 Programmable-rate 8–4 kbps Low-delay CELP Codecs 3658.10 Backward-adaptive Error Sensitivity Issues 3818.11 A Low-delay Multimode Speech Transceiver 3888.12 Chapter Summary 392III Wideband Speech, MPEG-4 Audio and Their Transmission 3939 Wideband Speech Coding 3959.1 Sub-band-ADPCM Wideband Coding at 64 kbps 3959.2 Wideband Transform-coding at 32 kbps 4139.3 Sub-band-split Wideband CELPCodecs 4169.4 Fullband Wideband A CELPCoding 4209.5 A Turbo-coded Burst-by-burst Adaptive Wideband Speech Transceiver 4259.6 Turbo-detected Unequal Error Protection Irregular Convolutional Coded AMR-WB Transceivers 4429.7 The AMR-WB+AudioCodec 4549.8 Chapter Summary 46610 MPEG-4 Audio Compression and Transmission 46910.1 OverviewofMPEG-4Audio 46910.2 General Audio Coding 47110.3 Speech Coding in MPEG-4 Audio 49510.4 MPEG-4CodecPerformance 50310.5 MPEG-4 Space–time Block Coded OFDM Audio Transceiver 50510.6 Turbo-detected Space–time Trellis Coded MPEG-4 Audio Transceivers 51610.7 Turbo-detected Space–time Trellis Coded MPEG-4 Versus AMR-WB Speech Transceivers 52510.8 Chapter Summary 534IV Very Low-rate Coding and Transmission 53711 Overview of Low-rate Speech Coding 53911.1 Low-bitrate Speech Coding 53911.2 Linear Predictive Coding Model 55311.3 Speech Quality Measurements 55711.4 Speech Database 56011.5 Chapter Summary 56312 Linear Predictive Vocoder 56512.1 Overview of a Linear Predictive Vocoder 56512.2 Line Spectrum Frequencies Quantisation 56612.3 Pitch Detection 57112.4 Unvoiced Frames 58312.5 Voiced Frames 58412.6 Adaptive Postfilter 58512.7 Pulse Dispersion Filter 58812.8 Results for Linear Predictive Vocoder 59212.9 Chapter Summary 59713 Wavelets and Pitch Detection 59913.1 Conceptual Introduction to Wavelets 59913.2 Introduction to Wavelet Mathematics 60213.3 Preprocessing the Wavelet Transform Signal 60713.4 Voiced–unvoiced Decision 61013.5 Wavelet-based Pitch Detector 61213.6 Chapter Summary 61914 Zinc Function Excitation 62114.1 Introduction 62114.2 Overview of Prototype Waveform Interpolation Zinc Function Excitation 62214.3 Zinc Function Modelling 62714.4 Pitch Detection 63114.5 Voiced Speech 63514.6 Excitation Interpolation Between Prototype Segments 63914.7 Unvoiced Speech 64514.8 Adaptive Postfilter 64514.9 Results for Single Zinc Function Excitation 64614.10 Error Sensitivity of the 1.9 kbps PWI-ZFE Coder 64914.11 Multiple Zinc Function Excitation 65414.12 A Sixth-rate, 3.8 kbps GSM-like Speech Transceiver 66114.13 Chapter Summary 66515 Mixed-multiband Excitation 66715.1 Introduction 66715.2 Overview of Mixed-multiband Excitation 66815.3 Finite Impulse Response Filter 67115.4 Mixed-multiband Excitation Encoder 67315.5 Mixed-multiband Excitation Decoder 67615.6 Performance of the Mixed-multiband Excitation Coder 68015.7 A Higher Rate 3.85 kbps Mixed-multiband Excitation Scheme 68615.8 A 2.35 kbps Joint-detection-based CDMA Speech Transceiver 69115.9 Chapter Summary 69916 Sinusoidal Transform Coding Below 4 kbps 70116.1 Introduction 70116.2 Sinusoidal Analysis of Speech Signals 70216.3 Sinusoidal Synthesis of Speech Signals 70416.4 Low-bitrate Sinusoidal Coders 70516.5 Incorporating Prototype Waveform Interpolation 70916.6 Encoding the Sinusoidal Frequency Component 71016.7 Determining the Excitation Components 71216.8 Quantising the Excitation Parameters 72016.9 Sinusoidal Transform Decoder 72816.10 Speech Coder Performance 73016.11 Chapter Summary 73617 Conclusions on Low-rate Coding 73717.1 Summary 73717.2 Listening Tests 73817.3 Summary of Very-low-rate Coding 73917.4 Further Research 74118 Comparison of Speech Codecs and Transceivers 74318.1 Background to Speech Quality Evaluation 74318.2 Objective Speech Quality Measures 74418.3 Subjective Measures 75218.4 Comparison of Subjective and Objective Measures 75318.5 Subjective Speech Quality of Various Codecs 75518.6 Error Sensitivity Comparison of Various Codecs 75718.7 Objective Speech Performance of Various Transceivers 75718.8 Chapter Summary 76419 The Voice over Internet Protocol 76519.1 Introduction 76519.2 Session Initiation Protocol 76619.3 H.323Standards 77419.4 Real-time Transport Protocol 77819.5 Conclusion 781A Constructing the Quadratic Spline Wavelets 783B Zinc Function Excitation 787C Probability Density Function for Amplitudes 793Bibliography 797Index 825Author Index 834

"I recommend this book as a useful reference on voice and audio compression for wireless communication." (Computing Reviews, June 10, 2008)