Audio Signal Processing and Coding

This book splits the difference between a purely academic and a practical approach. It does talk a great deal about the history of various audio representations, and it has some derivations, but it also has some practical numerical examples inserted into the narrative that help explain some of the audio concepts. The exercises at the end of each chapter are on the practical side, stressing numerical problems and MATLAB computer exercises over pure derivations. The book spends the first six chapters going over the basics that you need to know to understand or implement audio coding schemes. Chapter two reviews basic signal processing concepts associated with audio coding. Chapter 3 provides introductory material to waveform quantization and entropy coding schemes. Some of the key topics covered in that chapter include scalar quantization, uniform and nonuniform quantization, pulse code modulation, differential PCM, adaptive DPCM, vector quantization, bit allocation schemes, and entropy coding techniques such as Huffman, Rice, and arithmetic methods. Chapter 4 provides information on linear prediction and its application in narrow and wideband coding. In chapter 5, where psychoacoustic principles are described, Johnston's notion of perceptual entropy is presented as a measure of the fundamental limit of transparent compression for audio. Chapter 6, on filter bank design issues and algorithms, places particular emphasis on the modified discrete cosine transform which is widely used in several perceptual audio coding algorithms. The chapter also addresses pre-echo artifacts and control strategies. Chapters 7,8, and 9 review established and emerging techniques for transparent coding of FM and CD-quality audio signals, including several algorithms that have become international standards. Transform coding methodologies are described in chapter 7, subband algorithms are discussed in chapter 8, and sinusoidal algorithms are presented in chapter 9. Chapter 10 discusses the standardization activities in audio coding. It describes coding standards and products such as the ISO/IEC MPEG family. Details on popular standards, such as the MP3 and MPEG-4 AAC algorithms, are provided. Chapter 11 focuses on lossless audio coding and digital audio watermarking techniques. In particular, the SHORTEN, the DVD algorithm, the MUSICompress, the AudioPaK, and other such coding schemes are described in detail. Chapter 12 provides information on subjective quality measures for perceptual codecs. The five-point absolute and differential subjective quality scales are addressed. A set of subjective benchmarks is provided for the various standards in both stereophonicand multichannel modes so that algorithms can be more easily compared. If you've never been exposed to the subjects discussed in chapters one through six, you'll find this book rough going, since there are entire books written on the subjects that this book is covering in just one chapter each. However, I think it is a good review and a good way for those that are accustomed to looking at these problems from a purely mathematical perspective to see them from the viewpoint of audio processing and coding and to see problems solved using MATLAB. Chapters 7 through 9 are very good at presenting the various algorithms and illustrating them, but the quality seems to drop off as far as details go in the final three chapters starting with the sections on the MPEG standards in chapter 10. This book is good for background and reference, but don't expect to be able to decode or encode anything based solely on what's presented here. AUDIO SIGNAL PROCESSING AND CODING......Page 4 CONTENTS......Page 10 PREFACE......Page 18 1.1 Historical Perspective......Page 24 1.2 A General Perceptual Audio Coding Architecture......Page 27 1.3 Audio Coder Attributes......Page 28 1.3.3 Complexity......Page 29 1.4 Types of Audio Coders – An Overview......Page 30 1.5 Organization of the Book......Page 31 1.6 Notational Conventions......Page 32 Computer Exercises......Page 34 2.2 Spectra of Analog Signals......Page 36 2.3 Review of Convolution and Filtering......Page 39 2.4 Uniform Sampling......Page 40 2.5.1 Transforms for Discrete-Time Signals......Page 43 2.5.2 The Discrete and the Fast Fourier Transform......Page 45 2.5.4 The Short-Time Fourier Transform......Page 46 2.6 Difference Equations and Digital Filters......Page 48 2.7 The Transfer and the Frequency Response Functions......Page 50 2.7.1 Poles, Zeros, and Frequency Response......Page 52 2.7.2 Examples of Digital Filters for Audio Applications......Page 53 2.8.1 Down-sampling by an Integer......Page 56 2.8.2 Up-sampling by an Integer......Page 58 2.8.4 Quadrature Mirror Filter Banks......Page 59 2.9 Discrete-Time Random Signals......Page 62 2.9.1 Random Signals Processed by LTI Digital Filters......Page 65 2.10 Summary......Page 67 Problems......Page 68 Computer Exercises......Page 70 3.1 Introduction......Page 74 3.1.1 The Quantization–Bit Allocation–Entropy Coding Module......Page 75 3.2 Density Functions and Quantization......Page 76 3.3.1 Uniform Quantization......Page 77 3.3.2 Nonuniform Quantization......Page 80 3.3.3 Differential PCM......Page 82 3.4 Vector Quantization......Page 85 3.4.1 Structured VQ......Page 87 3.4.2 Split-VQ......Page 90 3.4.3 Conjugate-Structure VQ......Page 92 3.5 Bit-Allocation Algorithms......Page 93 3.6 Entropy Coding......Page 97 3.6.1 Huffman Coding......Page 100 3.6.2 Rice Coding......Page 104 3.6.3 Golomb Coding......Page 105 3.6.4 Arithmetic Coding......Page 106 Problems......Page 108 Computer Exercises......Page 109 4.1 Introduction......Page 114 4.2 LP-Based Source-System Modeling for Speech......Page 115 4.3 Short-Term Linear Prediction......Page 117 4.3.1 Long-Term Prediction......Page 118 4.4 Open-Loop Analysis-Synthesis Linear Prediction......Page 119 4.5 Analysis-by-Synthesis Linear Prediction......Page 120 4.5.1 Code-Excited Linear Prediction Algorithms......Page 123 4.6.1 Wideband Speech Coding......Page 125 4.6.2 Wideband Audio Coding......Page 127 4.7 Summary......Page 129 Problems......Page 130 Computer Exercises......Page 131 5.1 Introduction......Page 136 5.2 Absolute Threshold of Hearing......Page 137 5.3 Critical Bands......Page 138 5.4 Simultaneous Masking, Masking Asymmetry, and the Spread of Masking......Page 143 5.4.1 Noise-Masking-Tone......Page 146 5.4.4 Asymmetry of Masking......Page 147 5.4.5 The Spread of Masking......Page 148 5.5 Nonsimultaneous Masking......Page 150 5.6 Perceptual Entropy......Page 151 5.7 Example Codec Perceptual Model: ISO/IEC 11172-3 (MPEG - 1) Psychoacoustic Model 1......Page 153 5.7.2 Step 2: Identification of Tonal and Noise Maskers......Page 154 5.7.3 Step 3: Decimation and Reorganization of Maskers......Page 158 5.7.4 Step 4: Calculation of Individual Masking Thresholds......Page 159 5.8 Perceptual Bit Allocation......Page 161 Problems......Page 163 Computer Exercises......Page 164 6.1 Introduction......Page 168 6.2 Analysis-Synthesis Framework for M-band Filter Banks......Page 169 6.3 Filter Banks for Audio Coding: Design Considerations......Page 171 6.3.2 The Role of Frequency Resolution in Perceptual Bit Allocation......Page 172 6.3.3 The Role of Time Resolution in Perceptual Bit Allocation......Page 173 6.4 Quadrature Mirror and Conjugate Quadrature Filters......Page 178 6.5 Tree-Structured QMF and CQF M-band Banks......Page 179 6.6 Cosine Modulated “Pseudo QMF” M-band Banks......Page 183 6.7 Cosine Modulated Perfect Reconstruction (PR) M-band Banks and the Modified Discrete Cosine Transform (MDCT)......Page 186 6.7.2 MDCT Window Design......Page 188 6.7.3 Example MDCT Windows (Prototype FIR Filters)......Page 190 6.8 Discrete Fourier and Discrete Cosine Transform......Page 201 6.9 Pre-echo Distortion......Page 203 6.10.2 Window Switching......Page 205 6.10.3 Hybrid, Switched Filter Banks......Page 207 6.10.5 Temporal Noise Shaping......Page 208 6.11 Summary......Page 209 Problems......Page 211 Computer Exercises......Page 214 7.1 Introduction......Page 218 7.2 Optimum Coding in the Frequency Domain......Page 219 7.3 Perceptual Transform Coder......Page 220 7.3.1 PXFM......Page 221 7.3.2 SEPXFM......Page 222 7.4 Brandenburg-Johnston Hybrid Coder......Page 223 7.5.2 CNET MDCT Coder 1......Page 224 7.5.3 CNET MDCT Coder 2......Page 225 7.6 Adaptive Spectral Entropy Coding......Page 226 7.7 Differential Perceptual Audio Coder......Page 227 7.8 DFT Noise Substitution......Page 228 7.9 DCT with Vector Quantization......Page 229 7.10 MDCT with Vector Quantization......Page 230 Problems......Page 231 Computer Exercises......Page 233 8.1 Introduction......Page 234 8.1.1 Subband Algorithms......Page 235 8.2 DWT and Discrete Wavelet Packet Transform (DWPT)......Page 237 8.3.1 DWPT Coder with Globally Adapted Daubechies Analysis Wavelet......Page 241 8.3.2 Scalable DWPT Coder with Adaptive Tree Structure......Page 243 8.3.4 DWPT Coder with Adaptive Tree Structure and Locally Adapted Analysis Wavelet......Page 246 8.3.5 DWPT Coder with Perceptually Optimized Synthesis Wavelets......Page 247 8.4.1 Switched Nonuniform Filter Bank Cascade......Page 249 8.5 Hybrid WP and Adapted WP/Sinusoidal Algorithms......Page 250 8.5.1 Hybrid Sinusoidal/Classical DWPT Coder......Page 251 8.5.2 Hybrid Sinusoidal/M-band DWPT Coder......Page 252 8.5.3 Hybrid Sinusoidal/DWPT Coder with WP Tree Structure Adaptation (ARCO)......Page 253 8.6 Subband Coding with Hybrid Filter Bank/CELP Algorithms......Page 256 8.6.1 Hybrid Subband/CELP Algorithm for Low-Delay Applications......Page 257 8.6.2 Hybrid Subband/CELP Algorithm for Low-Complexity Applications......Page 258 Problems......Page 260 Computer Exercise......Page 263 9.1 Introduction......Page 264 9.2.1 Sinusoidal Analysis and Parameter Tracking......Page 265 9.2.2 Sinusoidal Synthesis and Parameter Interpolation......Page 268 9.3 Analysis/Synthesis Audio Codec (ASAC)......Page 270 9.3.3 ASAC Bit Allocation, Quantization, Encoding, and Scalability......Page 271 9.4 Harmonic and Individual Lines Plus Noise Coder (HILN)......Page 272 9.4.1 HILN Sinusoidal Analysis-by-Synthesis......Page 273 9.5 FM Synthesis......Page 274 9.5.2 Perceptual Audio Coding Using an FM Synthesis Model......Page 275 9.6 The Sines + Transients + Noise (STN) Model......Page 277 9.7 Hybrid Sinusoidal Coders......Page 278 9.7.1 Hybrid Sinusoidal-MDCT Algorithm......Page 279 9.7.2 Hybrid Sinusoidal-Vocoder Algorithm......Page 280 Problems......Page 281 Computer Exercises......Page 282 10.1 Introduction......Page 286 10.2.1 MIDI Synthesizer......Page 287 10.2.3 MIDI Applications......Page 289 10.3.1 The Evolution of Surround Sound......Page 290 10.3.3 The ITU-R BS.775 5.1-Channel Configuration......Page 291 10.4 MPEG Audio Standards......Page 293 10.4.1 MPEG-1 Audio (ISO/IEC 11172-3)......Page 298 10.4.2 MPEG-2 BC/LSF (ISO/IEC-13818-3)......Page 302 10.4.3 MPEG-2 NBC/AAC (ISO/IEC-13818-7)......Page 306 10.4.4 MPEG-4 Audio (ISO/IEC 14496-3)......Page 312 10.4.5 MPEG-7 Audio (ISO/IEC 15938-4)......Page 332 10.4.6 MPEG-21 Framework (ISO/IEC-21000)......Page 340 10.5 Adaptive Transform Acoustic Coding (ATRAC)......Page 342 10.6.1 Perceptual Audio Coder (PAC)......Page 344 10.6.3 Multichannel PAC (MPAC)......Page 346 10.7.1 Dolby AC-2, AC-2A......Page 348 10.7.2 Dolby AC-3/Dolby Digital/Dolby SR · D......Page 350 10.8 Audio Processing Technology APT-x100......Page 358 10.9.1 Framing and Subband Analysis......Page 361 10.9.3 ADPCM – Differential Subband Coding......Page 362 10.9.4 Bit Allocation, Quantization, and Multiplexing......Page 364 Computer Exercise......Page 365 11.1 Introduction......Page 366 11.2 Lossless Audio Coding (L(2)AC)......Page 367 11.2.1 L(2)AC Principles......Page 368 11.2.2 L(2)AC Algorithms......Page 369 11.3 DVD-Audio......Page 379 11.4 Super-Audio CD (SACD)......Page 381 11.4.2 Sigma-Delta Modulators (SDM)......Page 385 11.4.3 Direct Stream Digital (DSD) Encoding......Page 387 11.5 Digital Audio Watermarking......Page 391 11.5.1 Background......Page 393 11.5.2 A Generic Architecture for DAW......Page 397 11.5.3 DAW Schemes – Attributes......Page 400 11.6 Summary of Commercial Applications......Page 401 Computer Exercise......Page 405 12.1 Introduction......Page 406 12.2 Subjective Quality Measures......Page 407 12.3 Confounding Factors in Subjective Evaluations......Page 409 12.4 Subjective Evaluations of Two-Channel Standardized Codecs......Page 410 12.5 Subjective Evaluations of 5.1-Channel Standardized Codecs......Page 411 12.6 Subjective Evaluations Using Perceptual Measurement Systems......Page 412 12.6.2 NSE Perceptual Measurement Schemes......Page 413 12.7 Algorithms for Perceptual Measurement......Page 414 12.7.1 Example: Perceptual Audio Quality Measure (PAQM)......Page 415 12.7.2 Example: Noise-to-Mask Ratio (NMR)......Page 419 12.7.3 Example: Objective Audio Signal Evaluation (OASE)......Page 422 12.8 ITU-R BS.1387 and ITU-T P.861: Standards for Perceptual Quality Measurement......Page 424 12.9 Research Directions for Perceptual Codec Quality Measures......Page 425 REFERENCES......Page 428 INDEX......Page 482 Master algorithms and standards for transparent coding of high-fidelity audio Here is an in-depth treatment of algorithms and standards for transparent coding of high-fidelity audio. Readers learn how algorithms for high-quality digital audio deliver transparent signal reproduction with a minimum number of bits. The unique features of the book include detailed coverage of topics such as filter banks, transform coding, sinusoidal analysis, linear prediction, hybrid algorithms, perceptual evaluation methods, scalable algorithms, Internet applications, MP3 and MP4 stereo systems, and current international and commercial audio standards. Following a general introduction, the authors present fundamental signal processing concepts relevant to audio coding and then introduce waveform and entropy quantization schemes. Next are thorough treatments of the following topics: . Linear prediction, ADPCM, and CELP algorithms for narrowband and wideband coding. Cellular telephony vocoders versus CD-quality audio coders. Perceptual modules including the ISO/IEC 11172-3 (MPEG-1) psychoacoustic analysis model. Detailed descriptions of the MPEG-1 Layer III (MP3) and the AAC (MP4) algorithms. Descriptions of the algorithms behind successful products such as the Apple iPod. Filter bank design and algorithms and the Modified Discrete Cosine Transform (MDCT). Established and emerging standards for transparent coding of CD-quality stereo audio signals. Standardization activities in high-fidelity audio coding, including DVD-Audio, Super Audio CD (SACD), Dolby AC3, Digital, Digital Theater Systems (DTS), and Sony SDDS surround sound. Lossless audio coding and digital audio watermarking techniques, including SHORTEN, DVD-algorithm, MUSICompress, AudioPaK, C-LPAC, LTAC, and IntMDCT lossless coding schemes. Surround sound compression algorithms for cinema and super audio CD applications. Digital audio watermarking, content protection, and copyright issues. Complexity, delay, error control, and subjective quality in perceptual audio coding Computer exercises and MATLAB(R) hands-on projects complement the algorithm theory and reinforce concept learning. A comprehensive bibliography with more than 600 references to additional sources of information to explore individual topics in greater depth. This textbook includes all the right elements and topics for a senior and/or graduate-level course in speech/audio processing and multimedia. Moreover, it is highly recommended for practitioners, scientists, and audio engineers who want to master coding algorithms for high-fidelity audio An in-depth treatment of algorithms and standards for perceptual coding of high-fidelity audio, this self-contained reference surveys and addresses all aspects of the field. Coverage includes signal processing and perceptual (psychoacoustic) fundamentals, details on relevant research and signal models, details on standardization and applications, and details on performance measures and perceptual measurement systems. It includes a comprehensive bibliography with over 600 references, computer exercises, and MATLAB-based projects for use in EE multimedia, computer science, and DSP courses. An ftp site containing supplementary material such as wave files, MATLAB programs and workspaces for the students to solve some of the numerical problems and computer exercises in the book can be found at ftp://ftp.wiley.com/public/sci_tech_med/audio_signal Signal Processing Essentials -- Quantization And Entropy Coding -- Linear Prediction In Narrownband And Wideband Coding -- Psychoacoustic Principles -- Time-frequency Analysis: Filter Banks And Transforms -- Transform Coders -- Subband Coders -- Sinusoidal Coders -- Audio Coding Standards And Algorithms -- Lossless Audio Coding And Digital Watermarking -- Quality Measures For Perceptual Audio Coding. Andreas Spanias, Ted Painter, Venkatraman Atti. Includes Bibliographical References And Index.