JAWAHARLAL NEHRU TECHNOLOGICAL UNIVERSITY HYDERABAD

Course Objectives:

• 1. To understand the concept of speech processing.
• 2. To build speech based systems.
• 3. To analyze the performance of speech processing systems.

Course Outcomes

Upon the Successful Completion of the Course, the Students would be able to:

• 1. The mechanism of human speech production and perception.
• 2. Each component of speech recognition systems.
• 3. The importance of probabilistic modeling in speech recognition.
• 4. Build a speech recognition system.

UNIT - I

Introduction: Speech and Language, Speech analysis, Speech coding, speech production models, speech analysis and analysis-synthesis systems, Mechanisms and Models of the Human Auditory System

UNIT - II

Linear Predictive Coding (LPC) Analysis: Speech recognition, Graphical models, Hidden Markov models, Recognition and training algorithms, Language models

UNIT - III

Search Algorithms: Optimization, adaptation, Noise robustness, Digital Coding of Speech. Message Synthesis from Stored Human Speech Components. Phonetic Synthesis by Rule. Speech Synthesis from Textural or Conceptual Input

UNIT - IV

Introduction to Automatic Speech Recognition: Template Matching. Stochastic Modeling. Practical Techniques for Improving Speech Recognition and Performance, Automatic Speech Recognition for Large Vocabularies. Speaker Recognition and other Para-linguistic Technologies. Human Auditory System. Digital Coding of Speech.

UNIT - V

Phonetics Synthesis by Rule: Introduction Automatic Speech Recognition, Discriminative training for speech recognition, Speech recognition applications, Speech synthesis, voice conversion, Speaker recognition

TEXT BOOKS:

• 1. Digital Speech Processing: Synthesis, and Recognition by Sadaoki Furui, CRC Press.
• 2. Speech Synthesis and Recognition by Wendy Holmes, CRC Press.
• 3. Audio Signal Processing and Coding by Andreas Spanias, Ted Painter and Venkatraman Atti, Willey.