Source modeling techniques for quality enhancement in statistical parametric speech synthesis. (c2019)

• Record Type: Book
• Further Information: Note: K. Sreenivasa Rao, N. P. Narendra.
• Other Names: Rao, K. Sreenivasa (Krothapalli Sreenivasa); Narendra, N. P
• Contents: Intro; Preface; Contents; Acronyms; 1 Introduction; 1.1 Introduction; 1.2 Speech Synthesis Methods; 1.2.1 Formant Synthesis; 1.2.2 Articulatory Synthesis; 1.2.3 Concatenative Synthesis; 1.2.4 Statistical Parametric Speech Synthesis; 1.2.5 Hybrid Synthesis Methods; 1.3 Objectives and Scope of the Work; 1.4 Contributions of the Book; 1.4.1 Robust Voicing Detection and F0 Estimation Method; 1.4.2 Parametric Approach of Modeling the Excitation Signal; 1.4.3 Hybrid Approach of Modeling the Excitation Signal; 1.4.4 Generation of Creaky Voice; 1.5 Organization of the Book; References 2 Background and Literature Review2.1 HMM-Based Speech Synthesis; 2.1.1 Hidden Markov Model; 2.1.2 System Overview; 2.1.3 Duration Modeling; 2.1.4 Decision Tree-Based Context Clustering; 2.1.5 Synthesis; 2.2 Voicing Detection and F0 Estimation: A Review; 2.3 Source Modeling Approaches: A Review; 2.4 Generation of Creaky Voice: A Review; 2.5 Summary; References; 3 Robust Voicing Detection and F0 Estimation Method; 3.1 F0 Modeling and Generation in HTS; 3.2 Proposed Method for Voicing Detection and F0 Estimation 3.2.1 Zero-Frequency Filtering Method for Detecting the Instants of Significant Excitation3.2.2 Voicing Detection; 3.2.3 Influence of Window Size on the Strength of Excitation; 3.2.4 F0 Estimation; 3.2.5 Performance Evaluation; 3.3 Implementation of the Proposed Voicing Detection and F0 Extraction in HTS Framework; 3.4 Evaluation; 3.4.1 Evaluation of Voicing Detection; 3.4.2 Subjective Evaluation; 3.5 Summary; References; 4 Parametric Approach of Modeling the Source Signal; 4.1 Parametric Source Modeling Method Based on Principal Component Analysis 4.1.1 Generation of Pitch-Synchronous Residual Frames4.1.2 Parameterization of Residual Frame Using PCA; 4.1.3 Speech Synthesis Using the Proposed PCA-Based Parametric Source Model; 4.1.4 Evaluation; 4.2 Parametric Source Modeling Method Based on the Deterministic and Noise Components of Residual Frames; 4.2.1 Analysis of Characteristics of Residual Frames; 4.2.2 Overview of Proposed Parametric Source Model; 4.2.3 Parameterization of Deterministic Component; 4.2.4 Parameterization of Noise Component 4.2.5 Speech Synthesis Using the Proposed Deterministic and Noise Component-Based Parametric Source Model4.2.6 Evaluation; 4.3 Summary; References; 5 Hybrid Approach of Modeling the Source Signal; 5.1 Optimal Residual Frame-Based Hybrid Source Modeling Method; 5.1.1 Computation of Optimal Residual Frame for a Phone; 5.1.2 Clustering the Optimal Residual Frames; 5.1.3 Speech Synthesis Using the Proposed Optimal Residual Frame-Based Hybrid Source Model; 5.1.4 Evaluation; 5.2 Time-Domain Deterministic Plus Noise Model-Based Hybrid Source Model
• Publisher Details: Cham : Springer
• Publication Date: 2019
• Copyright Date: 2019
• Extent: 1 online resource (145 p.)
• Subjects: 006.54; Speech synthesis; Electronic books
• Languages: English
• ISBNs: 9783030027599; 3030027597
• Related ISBNs: 9783030027582
• Notes: Note: Includes bibliographical references and index.; Note: Description based on online resource; title from digital title page (viewed on January 30, 2019).
• Access Rights: Legal Deposit; Only available on premises controlled by the deposit library and to one user at any one time; The Legal Deposit Libraries (Non-Print Works) Regulations (UK).
• Access Usage: Restricted: Printing from this resource is governed by The Legal Deposit Libraries (Non-Print Works) Regulations (UK) and UK copyright law currently in force.