| Introduction | p. 1 |
| Acoustic MIMO Signal Processing | p. 1 |
| Organization of the Book | p. 4 |
| Theory | |
| Acoustic MIMO Systems | p. 9 |
| Signal Models | p. 9 |
| SISO Model | p. 9 |
| SIMO Model | p. 11 |
| MISO Model | p. 12 |
| MIMO Model | p. 12 |
| Characteristics of Acoustic Channels | p. 13 |
| Linearity and Shift-Invariance | p. 14 |
| FIR Representation | p. 14 |
| Time-Varying Channel Impulse Responses | p. 14 |
| Frequency Selectivity | p. 15 |
| Reverberation Time | p. 15 |
| Channel Invertibility and Minimum-Phase Filter | p. 16 |
| Multichannel Diversity and the Common-Zero Problem | p. 18 |
| Sparse Impulse Response | p. 19 |
| Measurement and Simulation of MIMO Acoustic Systems | p. 21 |
| Direct Measurement of Acoustic Impulse Responses | p. 22 |
| Image Model for Acoustic Impulse Response Simulation | p. 24 |
| Summary | p. 29 |
| Wiener Filter and Basic Adaptive Algorithms | p. 31 |
| Introduction | p. 31 |
| Wiener Filter | p. 32 |
| Impulse Response Tail Effect | p. 34 |
| Condition Number | p. 35 |
| Decomposition of the Correlation Matrix | p. 36 |
| Condition Number with the Frobenius Norm | p. 37 |
| Fast Computation of the Condition Number | p. 39 |
| Basic Adaptive Algorithms | p. 41 |
| Deterministic Algorithm | p. 41 |
| Stochastic Algorithm | p. 44 |
| Sign Algorithms | p. 46 |
| MIMO Wiener Filter | p. 48 |
| Numerical Examples | p. 53 |
| Summary | p. 56 |
| Sparse Adaptive Filters | p. 59 |
| Introduction | p. 59 |
| Notation and Definitions | p. 60 |
| The NLMS, PNLMS, and IPNLMS Algorithms | p. 61 |
| Universal Criterion | p. 64 |
| Linear Update | p. 65 |
| Non-Linear Update | p. 67 |
| Exponentiated Gradient Algorithms | p. 68 |
| The EG Algorithm for Positive Weights | p. 68 |
| The EG[plusminus] Algorithm for Positive and Negative Weights | p. 69 |
| The Exponentiated RLS (ERLS) Algorithm | p. 71 |
| The Lambert W Function Based Gradient Algorithm | p. 72 |
| Some Important Links Among Algorithms | p. 74 |
| Link Between NLMS and EG[plusminus] Algorithms | p. 74 |
| Link Between IPNLMS and EG[plusminus] Algorithms | p. 75 |
| Link Between LWG and EG[plusminus] Algorithms | p. 77 |
| Numerical Examples | p. 78 |
| Summary | p. 83 |
| Frequency-Domain Adaptive Filters | p. 85 |
| Introduction | p. 85 |
| Derivation of SISO FD Adaptive Algorithms | p. 86 |
| Criterion | p. 86 |
| Normal Equations | p. 89 |
| Adaptive Algorithms | p. 91 |
| Convergence Analysis | p. 93 |
| Approximation and Special Cases | p. 96 |
| Approximation | p. 96 |
| Special Cases | p. 98 |
| FD Affine Projection Algorithm | p. 99 |
| Generalization to the MISO System Case | p. 101 |
| Numerical Examples | p. 104 |
| Summary | p. 106 |
| Blind Identification of Acoustic MIMO Systems | p. 109 |
| Introduction | p. 109 |
| Blind SIMO Identification | p. 111 |
| Identifiability and Principle | p. 111 |
| Constrained Time-Domain Multichannel LMS and Newton Algorithms | p. 113 |
| Unconstrained Multichannel LMS Algorithm with Optimal Step-Size Control | p. 120 |
| Frequency-Domain Unnormalized and Normalized Multichannel LMS Algorithms | p. 122 |
| Adaptive Multichannel Exponentiated Gradient Algorithm | p. 135 |
| Numerical Examples | p. 141 |
| Blind MIMO Identification | p. 147 |
| Problem Formulation and Background Review | p. 148 |
| Memoryless MIMO System with White Inputs | p. 151 |
| Memoryless MIMO System with Colored Inputs | p. 152 |
| Convolutive MIMO Systems with White Inputs | p. 154 |
| Gonvolutive MIMO Systems with Colored Inputs | p. 156 |
| Frequency-Domain Blind Identification of Convolutive MIMO Systems and Permutation Inconsistency | p. 157 |
| Convolutive MIMO Systems with White but Quasistationary Inputs | p. 158 |
| Summary | p. 160 |
| Appendix. Blind SIMO Identification: A Derivation Directly from the Covariance Matrices of the System Outputs | p. 161 |
| Separation and Suppression of Co-Channel and Temporal Interference | p. 169 |
| Introduction | p. 169 |
| Separating Co-Ghannel and Temporal Interference | p. 170 |
| Example: Conversion of a 2 x 3 MIMO System to Two SIMO Systems | p. 170 |
| Generalization to M x N MIMO Systems with M [Greater than] 2 and M [Less than] N | p. 174 |
| Suppressing Temporal Interference | p. 177 |
| Direct Inverse (Zero-Forcing) Equalizer | p. 178 |
| MMSE Equalizer | p. 179 |
| MINT Equalizers | p. 179 |
| Summary | p. 182 |
| Applications | |
| Acoustic Echo Cancellation and Audio Bridging | p. 185 |
| Introduction | p. 185 |
| Network Echo Problem | p. 186 |
| Single-Channel Acoustic Echo Cancellation | p. 188 |
| Multichannel Acoustic Echo Cancellation | p. 190 |
| Multi versus Mono | p. 190 |
| Multichannel Identification and the Nonuniqueness Problem | p. 192 |
| Impulse Response Tail Effect | p. 194 |
| Some Different Solutions for Decorrelation | p. 195 |
| Hybrid Mono/Stereo Acoustic Echo Canceler | p. 199 |
| Double-Talk Detection | p. 200 |
| Basics | p. 200 |
| Double-Talk Detection Algorithms | p. 202 |
| Performance Evaluation of DTDs | p. 206 |
| Audio Bridging | p. 206 |
| Principle | p. 206 |
| Interchannel Differences for Synthesizing Stereo Sound | p. 209 |
| Choice of Interchannel Differences for Stereo AEC | p. 211 |
| Summary | p. 212 |
| Time Delay Estimation and Acoustic Source Localization | p. 215 |
| Time Delay Estimation | p. 215 |
| Cross-Correlation Method | p. 217 |
| Magnitude-Difference Method | p. 219 |
| Maximum Likelihood Method | p. 220 |
| Generalized Cross-Correlation Method | p. 223 |
| Adaptive Eigenvalue Decomposition Algorithm | p. 226 |
| Multichannel Cross-Correlation Algorithm | p. 227 |
| Forward Spatial Linear Prediction | p. 228 |
| Backward Spatial Linear Prediction | p. 230 |
| Spatial Linear Interpolation | p. 231 |
| Time Delay Estimation Using Spatial Linear Prediction | p. 232 |
| Spatial Correlation Matrix and Its Properties | p. 233 |
| Multichannel Cross-Correlation Coefficient | p. 235 |
| Time Delay Estimation Using MCCC | p. 235 |
| Adaptive Multichannel Time Delay Estimation | p. 236 |
| Acoustic Source Localization | p. 238 |
| Measurement Model and Cramer-Rao Lower Bound | p. 239 |
| Algorithm Overview | p. 242 |
| Maximum Likelihood Estimator | p. 243 |
| Least-Squares Estimators | p. 244 |
| Least-Squares Error Criteria | p. 245 |
| Spherical Intersection (SX) Estimator | p. 247 |
| Spherical Interpolation (SI) Estimator | p. 247 |
| Linear-Correction Least-Squares Estimator | p. 248 |
| Example System Implementation | p. 254 |
| Summary | p. 259 |
| Speech Enhancement and Noise Reduction | p. 261 |
| Introduction | p. 261 |
| Noise-Reduction and Speech-Distortion Measures | p. 263 |
| Noise-Reduction Factor and Noise-Reduction Gain Function | p. 204 |
| Speech-Distortion Index and Attenuation Frequency Distortion | p. 265 |
| Signal-to-Noise Ratio | p. 265 |
| Log-Spectral Distance | p. 266 |
| Itakura Distance | p. 266 |
| Itakura-Saito Distance | p. 268 |
| Mean Opinion Score | p. 269 |
| Single-Channel Noise-Reduction Algorithms: a Brief Overview | p. 269 |
| Time-Domain Wiener Filter | p. 270 |
| Estimation of the Clean Speech Samples | p. 270 |
| Estimation of the Noise Samples | p. 273 |
| Noise Reduction versus Speech Distortion | p. 274 |
| A Priori SNR versus a Posteriori SNR | p. 277 |
| Bounds for Noise Reduction and Speech Distortion | p. 281 |
| Particular Case: White Gaussian Noise | p. 282 |
| A Suboptimal Filter | p. 283 |
| Frequency-Domain Wiener Filter | p. 287 |
| Estimation of the Clean Speech Spectrum | p. 287 |
| A Priori SNR versus a Posteriori SNR | p. 290 |
| Noise Reduction Through Spectral Magnitude Restoration | p. 292 |
| Spectral Subtraction | p. 293 |
| Estimation of the Spectral Magnitude of the Clean Speech | p. 293 |
| Estimation of the Noise Spectrum | p. 295 |
| Relationship Between Spectral Subtraction and Wiener Filtering | p. 296 |
| Estimation of the Wiener Gain Filter | p. 299 |
| Simulations | p. 300 |
| Adaptive Noise Cancellation | p. 302 |
| Estimation of the Clean Speech | p. 302 |
| Ideal Noise Cancellation Performance | p. 304 |
| Signal Cancellation Problem | p. 305 |
| Simulations | p. 307 |
| Noise Reduction with a Microphone Array | p. 309 |
| Delay-and-Sum Algorithm | p. 310 |
| Linearly Constrained Algorithms | p. 312 |
| Summary | p. 317 |
| Source Separation and Speech Dereverberation | p. 319 |
| Cocktail Party Effect | p. 319 |
| Source Separation | p. 323 |
| Microphone Array Beamforming | p. 323 |
| Independent Component Analysis and Blind Source Separation | p. 331 |
| A Synergistic Solution to Source Separation and Speech Dereverberation | p. 341 |
| Summary | p. 350 |
| References | p. 353 |
| Index | p. 373 |
| Table of Contents provided by Ingram. All Rights Reserved. |
