1. Course Description
This course introduces signal processing techniques for processing speech
signals. Course content will be split into three major parts: speech signal
processing techniques and analysis methods, automatic speech recognition systems,
and current research trends in speech and language processing. Evaluation is done
based on two major components: traditional evaluation (homework + midterm) &
project
2. Required Textbooks
Lecture notes and papers provided in class
3. Reference Books
Rabiner and Schafer: Theory and Applications of Digital Speech Processing,
Prentice Hall, 2010
Huang, Acero, and Hon: Spoken Language Procesing, Prentice Hall, 2001
T. Quatieri: Discrete-time Speech Signal Processing, Prentice Hall, 2001
4. Teaching Methods
English teaching. Mix of lectures and discussions
5. Syllabus
Overview: Course overview, Review of basic DSP
Overview: Introduction to Speech Production & Acoustics Speech Models
Speech Analysis: Short-term Time Domain Processing, Short-time Fourier Transform
Speech Analysis: All-pole model: Linear Prediction
Speech Analysis: Homomorphic Signal Processing & Cepstral Analysis
ASR: Intro to Automatic Speech Recognition
ASR: Hidden Markov Model
ASR: Language Model
ASR: Robust front-end processing, Speaker Adaptation
ASR: Advanced issues in ASR
Exemplary Speech Technology (I): Voice Activity Detector & Diarization System
Exemplary Speech Technology (II): Speaker and Language Identity Recognition
Behavior Informatics from Speech and Language: Affective Computing
Behavior Informatics from Speech and Language: Mental Health Applications
6. Evaluation
Homework 15% (each 5%)
Midterm 25%
Project 60% (1st presentation 15%, final presentation 25%, participation 10%,
report 10%)
No late homework accepted
7. Course Website
(UNDER CONSTRUCTION)