Lecture Notes (or) Lecture Material (or) Lecture PPTs on "Speech Signal Processing" by Dr. Rajesh Hegde, IIT Kanpur

Rajesh M. Hegde

Office:  ACES 203-B, Dept. of Electrical Engg., IIT Kanpur India

Labs       ACES 203, 204 : MiPS Lab,WSN Lab,  MiPS Lab (For access inside IIT-K)

Phone: +91-512-259-6248 (Off)

Fax:        +91-512-259-0063

Email:    rhegde@iitk.ac.in

Rajesh M Hegde is currently Associate Professor in the Dept. of Electrical Engineering at IIT Kanpur. His areas of interest include speech, audio, vision, multimodal information fusion, and wireless sensor networks. He also has considerable interests in pervasive and mobile computing. Prior to this, he was a researcher and lecturer  at the California Inst. of Telecommunications and Information Technology (University of California San Diego) and associated with the DSP Lab.

Research Interests:

Sensor Array and Multi channel Signal processing

Multi microphone Speech Processing and Distant Speech Recognition

ICT in Agricultural Applications and Emergency Response

Localization and Mapping in Wireless Sensor Networks

Pervasive Computing and Communication

Spatial Audio for Virtual and Augmented Reality

Multimodal Signal Processing and Information Fusion     

Speech Signal Processing and Recognition of Indian Languages 

Lecture	Topic
1/2	Overview of speech recognition, Modeling the speech production mechanism, Source-system model of speech, Physiological and Mathematical categorization of speech sounds
1/2	Overview of speech recognition, Modeling the speech production mechanism, Source-system model of speech, Physiological and Mathematical categorization of speech sounds
3/4	Discrete time processing of speech signals, Relevance of the DFT, the ZT, convolution, filter banks, and analytical pole-zero modeling in speech recognition
3/4	Discrete time processing of speech signals, Relevance of the DFT, the ZT, convolution, filter banks, and analytical pole-zero modeling in speech recognition
3/4	Discrete time processing of speech signals, Relevance of the DFT, the ZT, convolution, filter banks, and analytical pole-zero modeling in speech recognition
3/4	Discrete time processing of speech signals, Relevance of the DFT, the ZT, convolution, filter banks, and analytical pole-zero modeling in speech recognition
5/6	Short time Fourier Analysis and Spectral estimation models for Speech - DTFT, DFT, Filter banks
5/6	Short time Fourier Analysis and Spectral estimation models for Speech - DTFT, DFT, Filter banks
5/6	Short time Fourier Analysis and Spectral estimation models for Speech - DTFT, DFT, Filter banks
7/8	Pole zero modeling and All pole modeling of speech, LPC model for speech, Basics of Speech Coding
Q1	Quiz 1
7/8	Pole zero modeling and All pole modeling of speech, LPC model for speech, Basics of Speech Coding
7/8	Pole zero modeling and All pole modeling of speech, LPC model for speech, Basics of Speech Coding
7/8	Pole zero modeling and All pole modeling of speech, LPC model for speech, Basics of Speech Coding
7/8	Pole zero modeling and All pole modeling of speech, LPC model for speech, Basics of Speech Coding
7/8	Pole zero modeling and All pole modeling of speech, LPC model for speech, Basics of Speech Coding
11/12	Homomorphic speech signal deconvolution, real and complex cepstral analysis
11/12	Homomorphic speech signal deconvolution, real and complex cepstral analysis
13/14	Features for speech recognition: MFCC, RASTA-PLP, Issues in speech feature vector extraction, dynamic features, feature selection
13/14	Features for speech recognition: MFCC, RASTA-PLP, Issues in speech feature vector extraction, dynamic features, feature selection
13/14	Features for speech recognition: MFCC, RASTA-PLP, Issues in speech feature vector extraction, dynamic features, feature selection
13/14	Features for speech recognition: MFCC, RASTA-PLP, Issues in speech feature vector extraction, dynamic features, feature selection
15/16	Spectral and cepstral distances in speech recognition, Vector Quantization
15/16	Spectral and cepstral distances in speech recognition, Vector Quantization
15/16	Spectral and cepstral distances in speech recognition, Vector Quantization
Q2	Quiz 2
17/18	GMMs for speaker and Language Identification
17/18	GMMs for speaker and Language Identification
DTW	Dynamic Time Warping  (The slides will be emailed to all registered students)
DTW	Dynamic Time Warping
DTW	Dynamic Time Warping
HMM	Hidden Markov Models for isolated word and continuous speech recognition - Part I
HMM	Hidden Markov Models for isolated word and continuous speech recognition - Part I
HMM	Hidden Markov Models for isolated word and continuous speech recognition - Part II
HMM	Hidden Markov Models for isolated word and continuous speech recognition - Part II
SP***	Student Project Presentations in the evenings at the end of semester