Speech and Audio Processing in Adverse Environments

Front Cover
Eberhard Hänsler, Gerhard Schmidt
Springer Science & Business Media, Jul 22, 2008 - Technology & Engineering - 736 pages
0 Reviews
Users of signal processing systems are never satis?ed with the system they currently use. They are constantly asking for higher quality, faster perf- mance, more comfort and lower prices. Researchers and developers should be appreciative for this attitude. It justi?es their constant e?ort for improved systems. Better knowledge about biological and physical interrelations c- ing along with more powerful technologies are their engines on the endless road to perfect systems. This book is an impressive image of this process. After “Acoustic Echo 1 and Noise Control” published in 2004 many new results lead to “Topics in 2 Acoustic Echo and Noise Control” edited in 2006 . Today – in 2008 – even morenew?ndingsandsystemscouldbecollectedinthisbook.Comparingthe contributions in both edited volumes progress in knowledge and technology becomesclearlyvisible:Blindmethodsandmultiinputsystemsreplace“h- ble” low complexity systems. The functionality of new systems is less and less limited by the processing power available under economic constraints. The editors have to thank all the authors for their contributions. They cooperated readily in our e?ort to unify the layout of the chapters, the ter- nology, and the symbols used. It was a pleasure to work with all of them. Furthermore, it is the editors concern to thank Christoph Baumann and the Springer Publishing Company for the encouragement and help in publi- ing this book.
 

What people are saying - Write a review

We haven't found any reviews in the usual places.

Contents

1034 Talking and Listening Tests
347
1035 Listeningonly Tests LOT and Third Party Listening Tests
348
1036 Experts Tests for Assessing Real Life Situations
349
104 Test Environment
350
1041 The Acoustical Environment
351
1043 Positioning of the HandsFree Terminal
352
1045 Influence of the Transmission System
354
1051 Speech and Perceptual Speech Quality Measures
356

225 Low Delay FilterBanks
26
23 The FilterBank Equalizer
29
232 Prototype Filter Design
31
233 Relation between GDFT and GDCT
33
234 Realization for Different Filter Structures
35
235 Polyphase Network Implementation
37
236 The NonUniform FilterBank Equalizer
41
237 Comparison between FBE and AS FB
43
24 Further Measures for Signal Delay Reduction
44
241 Concept
45
243 Approximation by an AutoRegressive Filter
46
244 Algorithmic Complexity
47
245 Warped Filter Approximation
48
25 Application to Noise Reduction
49
252 Instrumental Quality Measures
50
253 Simulation Results for the Uniform FilterBanks
51
254 Simulation Results for the Warped FilterBanks
53
26 Conclusions
55
References
56
A PreFilter for HandsFree Car Phone Noise Reduction Suppression of Harmonic Engine Noise Components
62
32 Analysis of the Different Car Noise Components
64
321 Wind Noise
65
323 Engine Noise
66
33 Engine Noise Removal Based on Notch Filters
68
34 Compensation of Engine Harmonics with Adaptive Filters
73
341 StepSize Control
75
342 Calculating the Optimal StepSize
78
343 Results of the Compensation Approach
80
35 Evaluation and Comparison of the Results Obtained by the Notch Filter and the Compensation Approach
84
36 Conclusions and Summary
85
362 Summary
86
References
87
ModelBased Speech Enhancement
89
42 Conventional Speech Enhancement Schemes
91
43 Speech Enhancement Schemes Based on Nonlinearities
93
44 Speech Enhancement Schemes Based on Speech Reconstruction
97
441 Feature Extraction and Control
99
442 Reconstruction of Speech Signals
110
45 Combining the Reconstructed and the Noise Suppressed Signal
124
451 Adding the Fully Reconstructed Signal
125
452 Adding only the Voiced Part of the Reconstructed Signal
129
46 Summary and Outlook
133
Bandwidth Extension of Telephony Speech
135
52 Organization of the Chapter
137
53 Basics
138
531 Human Speech Generation
139
532 SourceFilter Model
141
533 Parametric Representations of the Spectral Envelope
143
534 Distance Measures
147
54 NonModelBased Algorithms for Bandwidth Extension
149
542 Spectral Shifting
151
543 Application of NonLinear Characteristics
153
551 Generation of the Excitation Signal
155
552 Vocal Tract Transfer Function Estimation
159
56 Evaluation of Bandwidth Extension Algorithms
176
561 Objective Distance Measures
177
562 Subjective Measures
180
57 Conclusions
181
References
182
Dereverberation and Residual Echo Suppression in Noisy Environments
185
61 Introduction
186
62 Problem Formulation
188
63 OMLSA Estimator for Multiple Interferences
191
632 A priori SIR Estimator
193
64 Dereverberation of Noisy Speech Signals
195
642 Problem Formulation
197
643 Statistical Reverberation Model
199
644 Late Reverberant Spectral Variance Estimator
200
645 Summary and Discussion
203
651 Problem Formulation
204
652 Late Residual Echo Spectral Variance Estimator
206
653 Parameter Estimation
208
654 Summary
210
67 Experimental Results
212
671 Experimental Setup
214
673 Suppression of Residual Echo
216
674 Joint Suppression of Reverberation Residual Echo and Noise
221
68 Summary and Outlook
223
References
224
Low Distortion Noise Cancellers Revival of a Classical Technique
228
72 Distortions in Widrows Adaptive Noise Canceller
230
722 Distortion by Crosstalk
232
73 Paired Filter PF Structure
233
732 Evaluations
235
74 Crosstalk Resistant ANC and CrossCoupled Structure
239
741 Crosstalk Resistant ANC
240
742 CrossCoupled Structure
241
75 CrossCoupled Paired Filter CCPF Structure
242
752 Evaluations
245
76 Generalized CrossCoupled Paired Filter GCCPF Structure
247
761 Algorithm
250
762 Evaluation by Recorded Signals
251
77 Demonstration in a Personal Robot
261
References
263
Echo Cancellation
265
Nonlinear Echo Cancellation Based on Spectral Shaping
267
82 FrequencyDomain Model of Highly Nonlinear Residual Echo
268
821 Spectral Correlation Between Residual Echo and Echo Replica
269
822 Model of Residual Echo Based on Spectral Correlation
273
83 Echo Canceller Based on the New Residual Echo Model
274
832 Estimation of NearEnd Speech
275
833 Spectral Gain Control
276
84 Evaluations
277
842 Subjective Evaluation
279
85 DSP Implementation and RealTime Evaluation
280
References
281
Signal and System Quality Evaluation
284
TelephoneSpeech Quality
287
92 SpeechSignal Quality
289
922 SpeechSound Quality
290
93 SpeechQuality Assessment
292
933 Instrumental Quality Assessment
293
95 Auditory TotalQuality Assessment
294
952 Listening Tests
296
954 AbsoluteCategory Rating ACR LOTs
297
96 Auditory QualityAttribute Analysis
298
963 Search for Suitable Attributes
302
964 IntegralQuality Estimation from Attributes
305
97 Instrumental TotalQuality Measurement
306
973 Psychoacoustically Motivated Measures
312
98 Instrumental AttributeBased Quality Measurements
320
982 Loudness
322
983 Sharpness
323
985 DirectnessFrequency Content DFC
324
986 Continuity
326
987 Noisiness
329
988 Combined Direct and AttributeBased Total Quality Determination
331
References
332
Evaluation of Handsfree Terminals
338
102 Quality Assessment of Handsfree Terminals
340
103 Subjective Methods for Determining the Communicational Quality
342
1031 General Setup and Opinion Scales Used for Subjective Performance Evaluation
343
1032 Conversation Tests
345
1033 Double Talk Tests
346
1053 Background Noise
360
1054 Applications
363
106 Result Representation
365
1061 Interpretation of HFT Quality Pies
366
1062 Examples
368
1072 Intelligibility Outside Vehicles
372
References
375
MultiChannel Processing
378
CorrelationBased TDOAEstimation for Multiple Sources in Reverberant Environments
379
112 Analysis of TDOA Ambiguities
383
1122 Multipath Ambiguity
384
1124 Ambiguity due to Periodic Signals
386
113 Estimation of Direct Path TDOAs
390
1132 Raster Matching
392
114 Consistent TDOA Graphs
397
1142 Strategies of Consistency Check
398
1143 Properties of TDOA Graphs
399
1144 Efficient Synthesis Algorithm
402
1145 Initialization and Termination
404
1146 Estimating the Number of Active Sources
405
115 Experimental Results
406
1152 TDOA Estimation of a Single Signal Block
408
1153 Source Position Estimation
412
116 Summary
414
References
415
Microphone Calibration for MultiChannel Signal Processing
417
122 Beamforming with Ideal Microphones
418
1222 Evaluation of Beamformers
421
1223 Statistically Optimum Beamformers
424
123 Microphone Mismatch and its Effect on Beamforming
427
1231 Model for NonIdeal Microphone Characteristics
428
1232 Effect of Microphone Mismatch on Fixed Beamformers
429
1233 Effect of Microphone Mismatch on Adaptive Beamformers
430
1234 Comparison of Fixed and Adaptive Beamformers
432
1242 Analysis of Fixed Beamformers
440
1243 Analysis of Adaptive Beamformers
444
1244 Comparison of Fixed and Adaptive Beamformers
448
125 SelfCalibration Techniques
449
1251 Basic Unit
451
1252 Configurations for Array Processing
452
1253 Recursive Configuration
455
1254 Adaptation Control
457
1255 Experimental Results
458
126 Summary
459
12A Experimental Determination of the Directivity Index
460
12A2 Definition of the Coordinate System
462
12A3 Determination of the Directivity using the Normalized Cross Power Spectral Densities of Microphone Signals
464
References
465
Convolutive Blind Source Separation for Noisy Mixtures
469
132 Blind Source Separation for Acoustic Mixtures Based on the TRINICON Framework
473
1322 Optimization Criterion and Coefficient Update
474
1323 Approximations Leading to Special Cases
478
1324 Estimation of the Correlation Matrices and an Efficient Normalization Strategy
484
1325 On Broadband and Narrowband BSS Algorithms in the DFT Domain
485
1326 Experimental Results for Reverberant Environments
488
133 Extensions for Blind Source Separation in Noisy Environments
490
1331 Model for Background Noise in Realistic Environments
491
1332 PreProcessing for NoiseRobust Adaptation
493
1333 PostProcessing for Suppression of Residual Crosstalk and Background Noise
498
134 Conclusions
518
References
519
Binaural Speech Segregation
525
142 TF Masks for CASA
528
143 Anechoic Binaural Segregation
529
144 Reverberant Binaural Segregation
533
145 Evaluation
536
146 Concluding Remarks
543
147 Acknowledgments
546
SpatioTemporal Adaptive Inverse Filtering in the Wave Domain
550
152 Problem Description
553
1523 Multichannel Active Listening Room Compensation
556
1524 Multichannel Active Noise Control
559
1525 Unified Representation of SpatioTemporal Adaptive Filtering Problems
561
1526 FrequencyDomain Notation
563
1531 Classification of Algorithms
564
1533 Adaptive Solution
565
1534 Problems of the Adaptive Solution
567
154 Eigenspace Adaptive Filtering
568
1542 Eigenspace Adaptive Filtering
569
1543 Problems
570
1551 Concept
571
1552 The Circular Harmonics Decomposition
573
1553 The Circular Harmonics Expansion Using Boundary Measurements
574
156 Application of WDAF to Adaptive Inverse Filtering Problems
576
1561 Application of WDAF to Active Listening Room Compensation
577
1562 Application of WDAF to Active Noise Control
578
References
580
Selected Applications
584
Virtual Hearing
585
161 Previous Work
588
162 VirtualHearing
592
163 Room Acoustic Model
593
164 HRTF Simulation
597
165 Neural Model
600
166 The Software and Interface
605
167 Software Testing
609
168 Future Work and Conclusions
610
References
613
Dynamic Sound Control Algorithms in Automobiles
615
1711 Introduction of Dynamic Volume Control Systems
616
1712 Introduction of Dynamic Equalization Control Systems
618
172 Previous Systems Description and Analysis
619
1722 Microphone Based Dynamic Volume Control Sound Systems
621
1723 NonAcoustic Sensor Based Sound Systems
643
173 SpectrumBased Dynamic Equalization Control
645
1731 Frequency Domain Adaptive Filter
646
1732 Generalized Multidelay Adaptive Filter
649
1733 StepSize Control
652
1734 MultiChannel Systems
653
1735 Estimating the Power Spectral Density of the Background Noise
654
1736 Psychoacoustic Basics
658
1737 The Psychoacoustic Masking Model According to Johnston
661
174 Conclusion and Outlook
670
175 Acknowledgement
673
References
674
Towards Robust DistantTalking Automatic Speech Recognition in Reverberant Environments
679
182 The DistantTalking ASR Scenario
680
183 How to Deal with Reverberation in ASR Systems?
683
184 Effect of Reverberation in the Feature Domain
691
185 Signal Dereverberation and Beamforming
695
186 Robust Features
699
187 Model Training and Adaptation
700
188 Reverberation Modeling for Speech Recognition
702
1881 Feature Production Model
703
1882 Reverberation Model
704
1883 Training of the Reverberation Model
705
1884 Decoding
708
1885 Inner Optimization
713
1886 Solution of the Inner Optimization Problem in the Melspec Domain for Single Gaussian Densities
714
1887 Simulations
717
189 Summary and Conclusions
722
References
723
Index
729
Copyright

Other editions - View all

Common terms and phrases

About the author (2008)

EBERHARD HANSLER, Dr.-Ing., is Professor of Electrical Engineering at the Darmstadt University of Technology, Darmstadt, Germany.

GERHARD SCHMIDT, Dr.-Ing., is a Research Engineer at Temic Speech Dialog Systems in Ulm, Germany.

Bibliographic information