Linked e-resources
Details
Table of Contents
Chapter 1: Introduction; 1.1 Lower Limit of Bit-Rate; 1.1.1 Using Text Accompanying Speech; 1.2 Vocoder Framework; 1.3 Clustered Codebook to Continuous Codebook; 1.3.1 Clustered Segment Codebook; 1.3.2 Vector Quantization Performance Trends and Limits; 1.3.3 Random Segment Codebooks; 1.3.4 Vector to Segment Quantization Performance Retention; 1.3.5 A Converging Viewpoint; 1.3.5.1Reasoning I; Part (a); Part (b); 1.3.5.2Reasoning II; Part (a); Part (b); Part (c); 1.4 Speech-to-Speech Synthesis by Unit-Selection; 1.5 Alternate Perspectives for Ultra Low Bit-Rate Speech Coding
1.6 Applications of Ultra Low Bit-Rate Speech Coding1.7 Organization of the Book; Chapter 2: Ultra Low Bit-Rate Coders; 2.1 Vector and Matrix Quantization; 2.2 Segment Vocoders; 2.2.1 Automatic Segmentation; 2.2.1.1Spectral Transition Measure; 2.2.1.2Maximum-Likelihood Segmentation; 2.2.1.3ML Segmentation: Duration Constrained (ML(DC)); 2.2.1.4ML Segmentation: A Generalized Basis; 2.2.1.5 Syllable-like units and other segmentations; 2.2.1.6Temporal Decomposition; 2.2.2 Segment Quantization; 2.2.3 Joint Segmentation Quantization; 2.2.3.1Basic framework
2.2.3.2Shiraki and Honda Variable-Length Segment Quantization2.2.3.3 2-LevelDP Framework for Joint Segmentation and Quantization; 2.2.3.4One-Pass DP Algorithm; 2.2.3.5Phoneme Recognition and Phonetic Vocoders; 2.2.4 Segment Codebook; 2.2.4.1Template Segment Codebooks; 2.2.4.2HMM Segment Codebook; 2.2.5 Duration Modification; 2.2.6 Residual Parameterization and Quantization; 2.2.7 Synthesis; 2.3 R/D Optimal Linear Prediction; 2.3.1 Prandoni and Vetterli R/D Optimal Linear Prediction; 2.3.2 Variable-to-Variable Length Vector Quantization; 2.3.3 Multigrams Quantization
2.3.4 Distortion Constrained Segmentation2.4 HMM Based Recognition-Synthesis Paradigm; 2.4.1 HTS Based Framework; 2.4.2 Speaker Adaptive HMM Recognition-Synthesis; 2.4.3 Ergodic HMM Framework; 2.4.4 Ismail and Ponting HMM Based Vocoders; 2.4.5 Formant Trajectory Model Based Recognition-Synthesis; 2.5 ALISP Units and Refinements; 2.5.1 Basic ALISP Framework; 2.5.2 Re-segmented Long Synthesis Units; 2.5.3 Short Synthesis Units by Dynamic Unit Selection; 2.5.4 Pre-selection of Units; 2.5.5 Noise Robustness; 2.6 Speaker Adaptation in Phonetic Vocoders; 2.7 Unit-Selection Paradigms
2.8 Performance Measures for Segment QuantizationChapter 3: Unit Selection Framework; 3.1 Lee-Cox Single-Frame Unit Selection Quantization; 3.1.1 An Alternate `5 ms Segment ́Single-Frame Unit-Selection Algorithm; 3.2 Lee-Cox Segmental Unit Selection Quantization; 3.3 Run-Length Coding and Effective Bit-Rate; 3.4 Sub-optimality of Lee-Cox Segmental Unit-Selection Algorithm; Chapter 4: Unified and Optimal Unit-Selection Framework; 4.1 Unified Unit-Selection Framework; 4.1.1 Proposed One-Pass DP Algorithm; 4.1.1.1Comparison with Lee and Cox Single-Frame and Segmental Unit-Selection
1.6 Applications of Ultra Low Bit-Rate Speech Coding1.7 Organization of the Book; Chapter 2: Ultra Low Bit-Rate Coders; 2.1 Vector and Matrix Quantization; 2.2 Segment Vocoders; 2.2.1 Automatic Segmentation; 2.2.1.1Spectral Transition Measure; 2.2.1.2Maximum-Likelihood Segmentation; 2.2.1.3ML Segmentation: Duration Constrained (ML(DC)); 2.2.1.4ML Segmentation: A Generalized Basis; 2.2.1.5 Syllable-like units and other segmentations; 2.2.1.6Temporal Decomposition; 2.2.2 Segment Quantization; 2.2.3 Joint Segmentation Quantization; 2.2.3.1Basic framework
2.2.3.2Shiraki and Honda Variable-Length Segment Quantization2.2.3.3 2-LevelDP Framework for Joint Segmentation and Quantization; 2.2.3.4One-Pass DP Algorithm; 2.2.3.5Phoneme Recognition and Phonetic Vocoders; 2.2.4 Segment Codebook; 2.2.4.1Template Segment Codebooks; 2.2.4.2HMM Segment Codebook; 2.2.5 Duration Modification; 2.2.6 Residual Parameterization and Quantization; 2.2.7 Synthesis; 2.3 R/D Optimal Linear Prediction; 2.3.1 Prandoni and Vetterli R/D Optimal Linear Prediction; 2.3.2 Variable-to-Variable Length Vector Quantization; 2.3.3 Multigrams Quantization
2.3.4 Distortion Constrained Segmentation2.4 HMM Based Recognition-Synthesis Paradigm; 2.4.1 HTS Based Framework; 2.4.2 Speaker Adaptive HMM Recognition-Synthesis; 2.4.3 Ergodic HMM Framework; 2.4.4 Ismail and Ponting HMM Based Vocoders; 2.4.5 Formant Trajectory Model Based Recognition-Synthesis; 2.5 ALISP Units and Refinements; 2.5.1 Basic ALISP Framework; 2.5.2 Re-segmented Long Synthesis Units; 2.5.3 Short Synthesis Units by Dynamic Unit Selection; 2.5.4 Pre-selection of Units; 2.5.5 Noise Robustness; 2.6 Speaker Adaptation in Phonetic Vocoders; 2.7 Unit-Selection Paradigms
2.8 Performance Measures for Segment QuantizationChapter 3: Unit Selection Framework; 3.1 Lee-Cox Single-Frame Unit Selection Quantization; 3.1.1 An Alternate `5 ms Segment ́Single-Frame Unit-Selection Algorithm; 3.2 Lee-Cox Segmental Unit Selection Quantization; 3.3 Run-Length Coding and Effective Bit-Rate; 3.4 Sub-optimality of Lee-Cox Segmental Unit-Selection Algorithm; Chapter 4: Unified and Optimal Unit-Selection Framework; 4.1 Unified Unit-Selection Framework; 4.1.1 Proposed One-Pass DP Algorithm; 4.1.1.1Comparison with Lee and Cox Single-Frame and Segmental Unit-Selection