| Speech Enhancement | |
| Constant Directivity Beamforming | p. 3 |
| Introduction | p. 3 |
| Problem Formulation | p. 6 |
| Theoretical Solution | p. 7 |
| Continuous sensor | p. 7 |
| Beam-shaping function | p. 8 |
| Practical Implementation | p. 9 |
| Dimension-reducing parameterization | p. 9 |
| Reference beam-shaping filter | p. 11 |
| Sensor placement | p. 12 |
| Summary of implementation | p. 12 |
| Examples | p. 13 |
| Conclusions | p. 16 |
| References | p. 16 |
| Superdirective Microphone Arrays | p. 19 |
| Introduction | p. 19 |
| Evaluation of Beamformers | p. 20 |
| Array-Gain | p. 21 |
| Beampattern | p. 22 |
| Directivity | p. 23 |
| Front-to-Back Ratio | p. 24 |
| White Noise Gain | p. 24 |
| Design of Superdirective Beamformers | p. 24 |
| Delay-and-Sum Beamformer | p. 26 |
| Design for spherical isotropic noise | p. 26 |
| Design for Cylindrical Isotropic Noise | p. 30 |
| Design for an Optimal Front-to-Back Ratio | p. 30 |
| Design for Measured Noise Fields | p. 32 |
| Extensions and Details | p. 33 |
| Alternative Form | p. 33 |
| Comparison with Gradient Microphones | p. 35 |
| Conclusion | p. 36 |
| References | p. 37 |
| Post-Filtering Techniques | p. 39 |
| Introduction | p. 39 |
| Multi-channel Wiener Filtering in Subbands | p. 41 |
| Derivation of the Optimum Solution | p. 41 |
| Factorization of the Wiener Solution | p. 42 |
| Interpretation | p. 45 |
| Algorithms for Post-Filter Estimation | p. 46 |
| Analysis of Post-Filter Algorithms | p. 47 |
| Properties of Post-Filter Algorithms | p. 49 |
| A New Post-Filter Algorithm | p. 50 |
| Performance Evaluation | p. 51 |
| Simulation System | p. 52 |
| Objective Measures | p. 52 |
| Simulation Results | p. 54 |
| Conclusion | p. 57 |
| Spatial Coherence Functions for Differential Microphones in Isotropic Noise Fields | p. 61 |
| Introduction | p. 61 |
| Adaptive Noise Cancellation | p. 61 |
| Spherically Isotropic Coherence | p. 65 |
| Cylindrically Isotropic Fields | p. 73 |
| Conclusions | p. 77 |
| References | p. 84 |
| Robust Adaptive Beamforming | p. 87 |
| Introduction | p. 87 |
| Adaptive Beamformers | p. 88 |
| Robustness Problem in the GJBF | p. 90 |
| Robust Adaptive Microphone Arrays - Solutions to Steering-Vector Errors | p. 92 |
| LAF-LAF Structure | p. 92 |
| CCAF-LAF Structure | p. 94 |
| CCAF-NCAF Structure | p. 95 |
| CCAF-NCAF Structure with an AMC | p. 97 |
| Software Evaluation of a Robust Adaptive Microphone Array | p. 99 |
| Simulated Anechoic Environment | p. 99 |
| Reverberant Environment | p. 101 |
| Hardware Evaluation of a Robust Adaptive Microphone Array | p. 104 |
| Implementation | p. 104 |
| Evaluation in a Real Environment | p. 104 |
| Conclusion | p. 106 |
| References | p. 106 |
| GSVD-Based Optimal Filtering for Multi-Microphone Speech Enhancement | p. 1ll |
| Introduction | p. 111 |
| GSVD-Based Optimal Filtering Technique | p. 113 |
| Optimal Filter Theory | p. 114 |
| General Class of Estimators | p. 116 |
| Symmetry Properties for Time-Series Filtering | p. 117 |
| Performance of GSVD-Based Optimal Filtering | p. 118 |
| Simulation Environment | p. 118 |
| Spatial Directivity Pattern | p. 119 |
| Noise Reduction Performance | p. 121 |
| Robustness Issues | p. 121 |
| Complexity Reduction | p. 122 |
| Linear Algebra Techniques for Computing GSVD | p. 122 |
| Recursive and Approximate GSVD-Updating Algorithms | p. 123 |
| Downsampling Techniques | p. 125 |
| Simulations | p. 125 |
| Computational Complexity | p. 126 |
| Combination with ANC Postprocessing Stage | p. 127 |
| Creation of Speech and Noise References | p. 127 |
| Noise Reduction Performance of ANC Postprocessing Stage | p. 128 |
| Comparison with Standard Beamforming Techniques | p. 129 |
| Conclusion | p. 129 |
| References | p. 130 |
| Explicit Speech Modeling for Microphone Array Speech Acquisition | p. 133 |
| Introduction | p. 133 |
| Model-Based Strategies | p. 136 |
| Example 1: A Frequency-Domain Model-Based Algorithm | p. 137 |
| Example 2: A Time-Domain Model-Based Algorithm | p. 140 |
| Conclusion | p. 148 |
| References | p. 151 |
| Source Localization | |
| Robust Localization in Reverberant Rooms | p. 157 |
| Introduction | p. 157 |
| Source Localization Strategies | p. 158 |
| Steered-Beamformer-Based Locators | p. 159 |
| High-Resolution Spectral-Estimation-Based Locators | p. 160 |
| TDOA-Based Locators | p. 161 |
| A Robust Localization Algorithm | p. 164 |
| The Impulse Response Model | p. 164 |
| The GCC and PHAT Weighting Function | p. 166 |
| ML TDOA-Based Source Localization | p. 167 |
| SRP-Based Source Localization | p. 169 |
| The SRP-PHAT Algorithm | p. 170 |
| Experimental Comparison | p. 172 |
| References | p. 178 |
| Multi-Source Localization Strategies | p. 181 |
| Introduction | p. 181 |
| Background | p. 184 |
| Array Signal Model | p. 184 |
| Incoherent Approach | p. 185 |
| Coherent Signal Subspace Method (CSSM) | p. 185 |
| Wideband Weighted Subspace Fitting (WB-WSF) | p. 186 |
| The Issue of Coherent Multipath in Array Processing | p. 187 |
| Implementation Issues | p. 188 |
| Linear Prediction-ROOT-MUSIC TDOA Estimation | p. 189 |
| Signal Pre-Whitening | p. 189 |
| An Approximate Model for Multiple Sources in Reverberant Environments | p. 191 |
| Robust TDOA Estimation via ROOT-MUSIC | p. 192 |
| Estimation of the Number of Relevant Reflections | p. 194 |
| Source Clustering | p. 195 |
| Experimental Results | p. 196 |
| References | p. 198 |
| Joint Audio-Video Signal Processing for Object Localiza-tion and Tracking | p. 203 |
| Introduction | p. 203 |
| Recursive State Estimation | p. 205 |
| Linear Kalman Filter | p. 206 |
| Extended Kalman Filter due to a Measurement Nonlinearity | p. 210 |
| Decentralized Kalman Filter | p. 212 |
| Implementation | p. 218 |
| System description | p. 218 |
| Results | p. 219 |
| Discussion and Conclusions | p. 221 |
| References | p. 222 |
| Applications | |
| Microphone-Array Hearing Aids | p. 229 |
| Introduction | p. 229 |
| Implications for Design and Evaluation | p. 230 |
| Assumptions Regarding Sound Sources | p. 230 |
| Implementation Issues | p. 231 |
| Assessing Performance | p. 232 |
| Hearing Aids with Directional Microphones | p. 233 |
| Fixed-Beamforming Hearing Aids | p. 234 |
| Adaptive-Beamforming Hearing Aids | p. 235 |
| Generalized Sidelobe Canceler with Modifications | p. 236 |
| Scaled Projection Algorithm | p. 242 |
| Direction of Arrival Estimation | p. 243 |
| Other Adaptive Approaches and Devices | p. 243 |
| Physiologically-Motivated Algorithms | p. 244 |
| Beamformers with Binaural Outputs | p. 245 |
| Discussion | p. 246 |
| References | p. 249 |
| Small Microphone Arrays with Postfilters for Noise and Acoustic Echo Reduction Rainer Martin | p. 255 |
| Introduction | p. 255 |
| Coherence of Speech and Noise | p. 257 |
| The Magnitude Squared Coherence | p. 257 |
| The Reverberation Distance | p. 258 |
| Coherence of Noise and Speech in Reverberant Enclosures | p. 259 |
| Analysis of the Wiener Filter with Symmetric Input Signals | p. 263 |
| No Near End Speech | p. 265 |
| High Signal to Noise Ratio | p. 265 |
| A Noise Reduction Application | p. 266 |
| An Implementation Based on the NLMS Algorithm | p. 266 |
| Processing in the 800 - 3600 Hz Band | p. 268 |
| Processing in the 240 - 800 Hz Band | p. 269 |
| Evaluation | p. 269 |
| Alternative Implementations of the Coherence Based Postfilter | p. 271 |
| Combined Noise and Acoustic Echo Reduction | p. 271 |
| Expermental Results | p. 274 |
| Conclusions | p. 275 |
| References | p. 276 |
| Acoustic Echo Cancellation for Beamforming Microphone Arrays | p. 281 |
| Introduction | p. 281 |
| Acoustic Echo Cancellation | p. 282 |
| Adaptation algorithms | p. 284 |
| AEC for multi-channel sound reproduction | p. 287 |
| AEC for multi-channel acquisition | p. 287 |
| Beamforming | p. 288 |
| General structure | p. 288 |
| Time-invariant beamforming | p. 290 |
| Time-varying beamforming | p. 291 |
| Computational complexity | p. 292 |
| Generic structures for combining AEC with beamforming | p. 292 |
| Motivation | p. 292 |
| Basic options | p. 293 |
| AEC first | p. 293 |
| Beamforming first | p. 296 |
| Integration of AEC into time-varying beamforming | p. 297 |
| Cascading time-invariant and time-varying beamforming | p. 297 |
| AEC with GSC-type beamforming structures | p. 301 |
| Combined AEC and beamforming for multi-channel recording and multi-channel reproduction | p. 302 |
| Conclusions | p. 303 |
| References | p. 303 |
| Optimal and Adaptive Microphone Arrays for Speech In-put in Automobiles | p. 307 |
| Introduction: Hands-Free Telephony in Cars | p. 307 |
| Optimum and Adaptive Beamforming | p. 309 |
| Common Signal Modeling | p. 309 |
| Constrained Minimum Variance Beamforming and the Gen-eralized Sidelobe Canceler | p. 310 |
| In Situ Calibrated Microphone Array (ICMA) | p. 312 |
| Time-Domain Minimum-Mean-Square-Error Solution | p. 313 |
| Frequency-Domain Minimum-Mean-Square-Error Solution | p. 314 |
| Optimal Near-Field Signal-to-Noise plus Interference Beam-former | p. 316 |
| Subband Implementation of the Microphone Array | p. 317 |
| Description of LS-Subband Beamforming | p. 318 |
| Multi-Resorution Time-Frequency Adaptive Beamforming | p. 319 |
| Memory Saving and Improvements | p. 319 |
| Evaluation and Examples | p. 320 |
| Car Environment | p. 320 |
| Microphone Configurations | p. 321 |
| Performance Measures | p. 321 |
| Spectral Performance Measures | p. 322 |
| Evaluation on car data | p. 323 |
| Evaluation Results | p. 323 |
| Summary and Conclusions | p. 324 |
| References | p. 326 |
| Speech Recognition with Microphone Arrays | p. 331 |
| Introduction | p. 331 |
| State of the Art | p. 332 |
| Automatic Speech Recognition | p. 332 |
| Robustness in ASR | p. 336 |
| Microphone Arrays and Related Processing for ASR | p. 337 |
| Distant-Talker Speech Recognition | p. 339 |
| A Microphone Array-Based ASR System | p. 342 |
| System Description | p. 342 |
| Speech Corpora and Task | p. 345 |
| Experiments and Results | p. 346 |
| Discussion and Future Trends | p. 348 |
| References | p. 349 |
| Blind Separation of Acoustic Signals | p. 355 |
| Introduction | p. 355 |
| The Cocktail Party Effect | p. 355 |
| Chapter Overview | p. 356 |
| Blind Signal Separation of Convolutive Mixtures | p. 357 |
| Problem Structure | p. 357 |
| Goal of Convolutive BSS | p. 359 |
| Relationship to Other Problems | p. 360 |
| Criteria for Blind Signal Separation | p. 362 |
| Table of Contents provided by Publisher. All Rights Reserved. |