Nonlinear Analyses and Algorithms for Speech Processing: International Conference on Non-Linear Speech Processing, NOLISP 2005, Barcelona, Spain,

Front matter......Page 1
Rationale for a Speech Processing COST Action......Page 13
Management Committee Meetings......Page 16
Short Term Scientific Missions......Page 17
Collaboration with Other COST Actions......Page 18
Collaboration Between Different Countries......Page 19
Acknowledgement......Page 20
References......Page 21
Introduction......Page 22
Formulation of the Fuzzy Interference Canceller......Page 24
Expert Knowledge in the Fuzzy Interference Canceller......Page 25
Simulations......Page 26
Conclusions......Page 27
Fuzzy-Inference-Based Robust Beamforming......Page 28
Problem Statement......Page 29
Fuzzy Inference Based Beamformer......Page 30
Parameter Design......Page 32
Simulations......Page 33
Conclusions......Page 34
Fuzzy Unsupervised Classifiers......Page 36
Separation of Seismic Signals......Page 39
Conclusions......Page 43
Fuzzy Logic at the Protocol Level: Horizontal Hand-Off......Page 44
References......Page 46
Introduction......Page 49
Classical Filtering Approaches......Page 50
Definitions and Basic Properties......Page 53
Early Examples of Connected Operators......Page 54
Anti-extensive Reconstruction and Connected Operators......Page 55
Self-dual Reconstruction and Levelings......Page 59
Tree Representations and Connected Operators......Page 61
Example of Connected Operators Based on Tree Representations......Page 71
Pruning Strategies Involving Global Optimization Under Constraint......Page 73
Conclusions......Page 75
Introduction......Page 78
ALISP N-Gram System......Page 79
Experimental Setup......Page 80
Experimental Results......Page 81
Conclusions......Page 82
Introduction......Page 84
Speaker Identification Baseline......Page 86
MLP Design and Training......Page 87
Test Protocol......Page 88
Test Results......Page 89
Discussion......Page 90
References......Page 91
Introduction......Page 93
Theoretical Approach......Page 95
Threshold Estimation Based on Weighting Scores......Page 97
The BioTech Database......Page 98
Results......Page 99
Conclusions......Page 102
References......Page 103
Introduction......Page 104
Speech Processing and Phoneme Feature Generation......Page 106
Support Vector Machine......Page 107
Experimental Methodology and Results......Page 108
Conclusion......Page 110
Introduction......Page 112
Bandwidth Extended Database......Page 113
Watermarked Database......Page 114
Algorithm Evaluation......Page 116
Speaker Verification......Page 117
References......Page 119
Introduction......Page 120
Definitions......Page 121
Preliminary Results......Page 122
Acknowledgements......Page 126
References......Page 127
Methodology......Page 128
First Analysis: F0 Results......Page 129
First Analysis: Intensity Results......Page 132
Second Analysis: F0 Results......Page 133
Second Analysis: Intensity Results......Page 134
Conclusion......Page 135
References......Page 136
Introduction......Page 137
Extraction of the Fundamental Drive......Page 140
Entrainment of the Primary Response......Page 143
Long Range Correlation in a Vowel – Nasal Diphone......Page 146
Discussion and Conclusion......Page 148
References......Page 149
Introduction......Page 151
Closed Phase Glottal Inverse Filtering......Page 152
CPIF with a Second Channel......Page 153
Model-Based Approaches......Page 154
Adaptive Inverse Filtering Approaches......Page 155
Discussion......Page 157
Introduction......Page 162
Method......Page 164
Human Voice Signals......Page 166
Results......Page 167
Discussion......Page 169
Conclusion......Page 171
References......Page 172
Introduction......Page 173
Generalized Cepstral Analysis......Page 174
Deconvolution in the Pseudo Cepstral Domain......Page 178
Spectral Domain......Page 179
Time Domain......Page 180
Concentration Measure for Male and Female Voices......Page 181
References......Page 184
Introduction......Page 186
Model Assumptions......Page 187
Bispectrum Estimators......Page 188
Detection Tests for Voice Activity......Page 189
Noise Reduction Block......Page 190
Experimental Framework......Page 192
Conclusions......Page 195
Introduction......Page 198
The MOCHA Database......Page 199
$\epsilon$-Support Vector Regression......Page 200
Data Processing......Page 201
Training and Results......Page 202
Conclusion......Page 204
Introduction......Page 208
Models......Page 209
Lattice Filter Implementation......Page 210
Multi-step Linear Predictive Analysis as a Paradigm for the Analysis of Vocal Dysperiodicities......Page 211
Generalized Variogram......Page 212
Methods......Page 213
Noise Marker......Page 214
Sustained Vowels and Running Speech......Page 215
References......Page 217
Introduction......Page 218
Correlation Dimension......Page 219
New Voice Disorder Parameterisation......Page 220
Evaluation of the Parameterization......Page 221
Parameterization......Page 225
Classifier......Page 226
Results......Page 228
References......Page 229
Introduction......Page 231
Database......Page 232
Parameterization......Page 233
Pattern Classification: The SVM Detector......Page 236
Evaluation Procedure......Page 237
Results......Page 239
Conclusions......Page 240
References......Page 241
Introduction......Page 243
Model I: Fourier Series Representation......Page 244
Model II: Distortion Function Representation......Page 245
Phonatory Excitation......Page 247
Vocal Jitter and Microtremor......Page 249
Random Cycle Lengths......Page 250
Results......Page 251
Conclusion......Page 252
References......Page 253
Introduction......Page 254
Estimating Cord Dynamics......Page 256
Estimation of the Body Biomechanical Parameters......Page 259
Results for Synthetic Voice......Page 260
Results from Natural Voice......Page 263
Conclusions......Page 266
References......Page 267
Introduction......Page 269
SVM Formulation......Page 270
Feature Extraction and Dimensional Normalization......Page 272
Non-uniform Distribution of Analysis Instants......Page 273
Baseline System and Database......Page 274
Selecting Parameters for the SVM-Based Recognizer......Page 275
Conclusions and Further Work......Page 277
Introduction......Page 279
Data......Page 280
Boundaries and Latency......Page 281
Boundaries and Entropy......Page 282
Entropy......Page 283
Boundaries and Latency......Page 285
Conclusions......Page 287
Introduction......Page 289
Third-Order Moment Feature Computation......Page 290
Experiments......Page 291
Conclusion......Page 293
References......Page 294
Introduction......Page 296
Non-linear Predictive Sub-band Feature Extractor......Page 297
Description......Page 298
Experimental Conditions......Page 299
Results......Page 300
Conclusion......Page 301
Introduction......Page 303
Quantile Based Noise Estimation and Spectral Subtraction......Page 304
Adaptive QBNE......Page 306
Speech Band Emphasizing Filter Bank......Page 308
Experimental Results......Page 310
Conclusion......Page 313
Introduction......Page 315
Detection of Anchor Points......Page 317
Speech Data and Representation......Page 319
System for Detection of VOPs in Continuous Speech Utterances......Page 321
Classification System for Recognition of Multilingual CV Units......Page 323
Spotting CV Units in Continuous Speech......Page 326
Summary and Conclusions......Page 328
Introduction......Page 330
MMSBA Schemes Employing WF......Page 331
Diverse SBP Options......Page 332
Wiener Filtering (WF)......Page 333
Recursive Magnitude Squared Coherence (MSC) Metric for Selecting SBP......Page 334
Simulation Results......Page 335
Conclusions......Page 336
References......Page 338
Introduction......Page 340
Double-Gamma Modeling of Speech and Noise Spectra......Page 341
Actual Adaptation of the Modeled Distribution Parameters......Page 342
MMSE Estimation......Page 343
Model 2: Gamma Speech and Gaussian Noise......Page 344
Cumulative Distribution Function Equalization......Page 345
Model 3: Gamma Speech and Gamma Noise......Page 346
Experiment......Page 347
Conclusion......Page 348
Nonlinear Prediction......Page 350
Prediction Based on Vectors......Page 351
Synthesis Structure of Volterra Systems......Page 352
Synthesis of Stationary Sounds......Page 354
Synthesis of Words......Page 356
Conclusions......Page 358
References......Page 359
Introduction......Page 360
Deriving the Continuous Model......Page 363
Deriving the Discrete Model......Page 364
Quasi-linear Prediction for Parametric Identification......Page 367
Discussion and Conclusions......Page 368
Introduction......Page 369
Source Separation Review......Page 370
Model and Assumptions......Page 371
Summary of the Deconvolution Algorithm......Page 372
Whitening of the Signal......Page 373
$Des-Gaussianity$ of the Signal......Page 374
Experimental Results......Page 375
Summary......Page 378
References......Page 379
Introduction......Page 380
Motivation and Aim of the Method......Page 382
Important Theorems Derived from the DTWT Algorithm......Page 383
The Algorithm of Segmented Wavelet Transform......Page 386
Corollaries and Limitations of the SegWT Algorithm......Page 388
Conclusion......Page 389
Back matter......Page 391