Speech Enhancement: Theory and Practice

Description

The first book to provide comprehensive and up-to-date coverage of all major speech enhancement algorithms proposed in the last two decades, Speech Enhancement: Theory and Practice is a valuable resource for experts and newcomers in the field. The book covers traditional speech enhancement algorithms, such as spectral subtraction and Wiener filtering algorithms as well as state-of-the-art algorithms including minimum mean-squared error algorithms that incorporate signal-presence uncertainty and subspace algorithms that incorporate psychoacoustic models. The coverage includes objective and subjective measures used to evaluate speech quality and intelligibility. Divided into three parts, the book presents the digital-signal processing and speech signal fundamentals needed to understand speech enhancement algorithms, the various classes of speech enhancement algorithms proposed over the last two decades, and the methods and measures used to evaluate the performance of speech enhancement algorithms. The text is supplemented with examples and figures designed to help readers understand the theory. MATLAB® implementations of all major speech enhancement algorithms and a speech database that can be used for evaluation of noise reduction algorithms are available for download on the book's description page at the CRC Press website. Providing clear and concise coverage of the subject, the author brings together a large body of knowledge about how human listeners compensate for acoustic noise when in noisy environments. This book is a valuable resource not only for engineers who want to implement the latest speech enhancement algorithms but also for speech practitioners who want to incorporate some of these algorithms into hearing aid applications for speech intelligibility and/or quality improvement. A download is available for those that purchase this book and can be obtained by contacting [email protected], providing proof of purchase. Content: IntroductionUnderstanding the Enemy: NoiseClasses of Speech Enhancement AlgorithmsBook OrganizationReferencesPart I FundamentalsDiscrete-Time Signal Processing and Short-Time Fourier AnalysisDiscrete-Time SignalsLinear Time-Invariant Discrete-Time Systemsz-TransformDiscrete-Time Fourier TransformShort-Time Fourier TransformSpectrographic Analysis of Speech SignalsSummaryReferencesSpeech Production and PerceptionSpeech SignalSpeech Production ProcessEngineering Model of Speech ProductionClasses of Speech SoundsAcoustic Cues in Speech PerceptionSummaryReferencesNoise Compensation by Human ListenersIntelligibility of Speech in Multiple-Talker ConditionsAcoustic Properties of Speech Contributing to RobustnessPerceptual Strategies for Listening in NoiseSummaryReferencesPart II AlgorithmsSpectral-Subtractive AlgorithmsBasic Principles of Spectral SubtractionGeometric View of Spectral SubtractionShortcomings of the Spectral Subtraction MethodSpectral Subtraction Using OversubtractionNonlinear Spectral SubtractionMultiband Spectral SubtractionMMSE Spectral Subtraction AlgorithmExtended Spectral SubtractionSpectral Subtraction Using Adaptive Gain AveragingSelective Spectral SubtractionSpectral Subtraction Based on Perceptual PropertiesPerformance of Spectral Subtraction AlgorithmsSummaryReferencesWiener FilteringIntroduction to Wiener Filter TheoryWiener Filters in the Time DomainWiener Filters in the Frequency DomainWiener Filters and Linear PredictionWiener Filters for Noise ReductionIterative Wiener FilteringImposing Constraints on Iterative Wiener FilteringConstrained Iterative Wiener FilteringConstrained Wiener FilteringEstimating the Wiener Gain FunctionIncorporating Psychoacoustic Constraints in Wiener FilteringCodebook-Driven Wiener FilteringAudible Noise Suppression AlgorithmSummaryReferencesStatistical-Model-Based MethodsMaximum-Likelihood EstimatorsBayesian EstimatorsMMSE EstimatorImprovements to the Decision-Directed ApproachImplementation and Evaluation of the MMSE EstimatorElimination of Musical NoiseLog-MMSE EstimatorMMSE Estimation of the pth-Power SpectrumMMSE Estimators Based on Non-Gaussian DistributionsMaximum A Posteriori (Map) EstimatorsGeneral Bayesian EstimatorsPerceptually Motivated Bayesian EstimatorsIncorporating Speech Absence Probability in Speech EnhancementMethods for Estimating the A Priori Probability of Speech AbsenceSummaryReferencesSubspace AlgorithmsIntroductionUsing SVD for Noise Reduction: TheorySVD-Based Algorithms: White NoiseSVD-Based Algorithms: Colored NoiseSVD-Based Methods: A Unified ViewEVD-Based Methods: White NoiseEVD-Based Methods: Colored NoiseEVD-Based Methods: A Unified ViewPerceptually Motivated Subspace AlgorithmsSubspace-Tracking AlgorithmsSummaryReferencesNoise-Estimation AlgorithmsVoice Activity Detection vs. Noise EstimationIntroduction to Noise-Estimation AlgorithmsMinimal-Tracking AlgorithmsTime-Recursive Averaging Algorithms for Noise EstimationHistogram-Based TechniquesOther Noise-Estimation AlgorithmsObjective Comparison of Noise-Estimation AlgorithmsSummaryReferencesPart III EvaluationEvaluating Performance of Speech Enhancement AlgorithmsQuality vs. IntelligibilityEvaluating Intelligibility of Processed SpeechEvaluating Quality of Processed SpeechEvaluating Reliability of Quality Judgments: Recommended PracticeSummaryReferencesObjective Quality and Intelligibility MeasuresObjective Quality MeasuresEvaluation of Objective Quality MeasuresQuality Measures: Summary of Findings and Future DirectionsSpeech Intelligibility MeasuresEvaluation of Intelligibility MeasuresIntelligibility Measures: Summary of Findings and Future DirectionsSummaryReferencesComparison of Speech Enhancement AlgorithmsNOIZEUS: A Noisy Speech Corpus for Quality Evaluation of Speech Enhancement AlgorithmsComparison of Enhancement Algorithms: Speech QualityComparison of Enhancement Algorithms: Speech IntelligibilitySummaryReferencesPart IV Future StepsAlgorithms That Can Improve Speech IntelligibilityReasons for the Absence of Intelligibility Improvement with Existing Noise-Reduction AlgorithmsAlgorithms Based on Channel Selection: A Different Paradigm for Noise ReductionChannel-Selection CriteriaIntelligibility Evaluation of Channel-Selection-Based Algorithms: Ideal ConditionsImplementation of Channel-Selection-Based Algorithms in Realistic ConditionsEvaluating Binary Mask Estimation AlgorithmsChannel Selection and Auditory Scene AnalysisSummaryReferencesAppendicesAppendix A: Special Functions and IntegralsAppendix B: Derivation of the MMSE EstimatorAppendix C: MATLAB (R) Code and Speech/Noise DatabasesIndex