Students can master the basic theory of digital speech signal processing and mathematical analysis and modeling method of sound object through this course. Also, they will understand the specific applications and cutting-edge technology in the field of speech signal processing. This course aims to improve the ability of using basic theory to solve practical problems. The knowledge includes digital representations of speech signals, overview of phonetics, speech production model, auditory systems and speech perception, short-time analysis of speech signal and key techniques of speech signal processing(linear predictive analysis, homomorphic analysis, vector quantization and Hidden Markov Model);specific application areas of digital speech signal processing include speech coding, speech recognition, speech synthesis and speech enhancement; applications of cutting-edge technologies such as deep learning in the field of speech signal processing.
Overview
Syllabus
- 1. Introduction
- 1.1 Outline
- 1.2 Application of digital speech signal processing
- 1.3 Development history of speech signal processing
- 1.4 Basic Knowledge of speech signal
- Test
- 2. Fundamentals of Speech processing
- 2.1 Speech processing techniques
- 2.2 Short-time analysis and windowing
- 2.3 Short-time parameters of speech signal
- 2.4 Pitch estimation
- Test
- 3. Speech production and perception
- 3.1 Speech chain
- 3.2 Speech production
- 3.3 Speech perception
- 3.4 Speech quality test
- Test
- 4. Linear predictive analysis（LPA）
- 4.1 Basic concepts of LP
- 4.2 Estimation of LP coefficients
- 4.3 LP spectrum and basic features
- 4.4 Alternative Representations of LP parameters
- Test
- 5. Homomorphic analysis（HA）
- 5.1 Homomorphic systems
- 5.2 Cepstrum and complex cepstrum
- 5.3 Homomorphic filtering
- 5.4 Homomorphic analysis of speech signal
- Test
- 6. Vector quantization（VQ）
- 6.1 Basic concept of VQ
- 6.2 VQ system design
- 6.3 Codebook design algorithms
- 6.4 Different types of VQ
- Test
- 7. Hidden Markov Models（HMM）
- 7.1 Prerequisite knowledge
- 7.2 Basic concept of HMM
- 7.3 Three basic problems of HMM
- 7.4 Insights of HMM
- Test
- 8. Speech Coding
- 8.1 Overview of speech coding
- 8.2 Speech coding strategies
- 8.3 Code Excited Linear Prediction (CELP)
- 8.4 Key technologies of speech coding
- Test
- 9. Speech Recognition
- 9.1 Overview of speech recognition
- 9.2 Basic speech recognition systems
- 9.3 Models in speech recognition
- 9.4 Challenges for speech recognition
- Test
- 10. Speech Enhancement
- 10.1 Overview of speech enhancement
- 10.2 Speech separation problem
- 10.3 Monaural speech denoising
- 10.4 Other speech enhancement techniques
- Test
- 11. Speech Synthesis
- 11.1 Overview of speech synthesis
- 11.2 Methods of speech synthesis(PartI)
- 11.3 Methods of speech synthesis(PartII)
- 11.4 Prosody and PSOLA
- Test
- 12. Latest development in speech signal processing
- 12.1 Background of Deep Learning
- 12.2 Outline of neural networks
- 12.3 Deep learning based speech recognition
- 12.4 Deep learning based speech separation
- Test
- Exam
Taught by
Jin Wang
