ECE410: Voice Signal Processing
Summer 2005
|
Time: |
Tuesdays |
|
|
Place: |
Norden: on-site |
|
|
Instructors: |
Jeffrey N. Denenberg |
Douglas Lyon |
|
Phone: |
(203) 268-1021 |
(203) 641-6293 |
|
Fax: |
|
(203) 877-4187 |
|
Email: |
||
|
Home Page: |
||
|
Prerequisites: |
CS232, MA 172 or permission of instructor. |
|
|
Office Hours: |
The hour before class on Tuesdays |
|
|
Texts: |
Sanjit J. Mitra, |
Lyon and Rao, |
|
Software: |
MatLab 4.2c (or later), The Mathworks |
|
|
Phillips and Parr, Signals, Systems, and Transforms, Philips and Parr, Interactive Lecture Notes (Chapters 1-7), EE235 Smith, DSP-Guide, Zip File (8
Mbytes) Mitra, 2nd Edition, Lecture Notes, Index Hsu, Analog and Digital Communications, Lathi, Linear Systems and Signals, Soliman and Srinath, Continous and Discrete Signals
and Systems, Lathi, Linear Systems and Signals, Denenberg, Fourier Series Denenberg, Fourier Transform Denenberg, Sampling and Reconstruction Denenberg, The Discrete Fourier Transform (DFT) Denenberg, The z-Transform Denenberg, Filter Approximation Theory Denenberg, Linear Systems Denenberg, Introduction to Noise |
|
Course Description:
Overview of Digital Audio and its application Current state of streaming Audio on the Internet Digital Audio Processing Fundamentals. This course applies transform concepts and applied multi-media object-oriented programming. Students will apply the theories of Sampling, Spectra, Fast Fourier Transform Class, convolution and frequency space processing, compression and one-dimensional streaming. Students will apply the theories by creating programs that read processing and write audio streams. They are exposed to the elements of multi-media network delivery of data. They learn about a wide class of FFT algorithms and elementary sound synthesis. This course requires substantial programming effort and emphasis is place on good software engineering practices. Students will learn enough signal processing to take Image Processing, the follow-on course.
Schedule - Summer 2005 (Updated periodically so check regularly)
|
Student Learning Objective |
Outcome |
|
The students will learn the principles of
Digital Signal Processing. |
Demonstrate the ability to analyze the performance
of Discrete-Time systems |
|
The students will become proficient with
the usage of the Java language. |
Demonstrate the ability to utilize
Java in practical signal processing problems |
|
The students will recognize
interrelationships among signals and spectra. |
Perform spectral analysis |
|
The students will learn enough CS to do
basic multi-media programming. |
Write a program to display waveforms
using a software o-scope Perform constructive and destructive
synthesis |
|
The students will learn about Object
Oriented Design. |
Uses appropriate object-oriented
design patterns to solve problems |
|
The students will learn how to apply
their mathematical background to signal processing. |
Apply transform concepts in programming
situations |
Supplementary Materials:
· MatLab Tutorial by B. Aliane
· MatLab files for Phillips and Parr, "Signals, Systems and Transforms", 2rd Ed.: (zip archive - 20 KB) Thanks to Drs. Phillips/Parr and The Mathworks
· Index to UCLA 102 Practice Sets Thanks to Dr. Paganini
· Denenberg References In MS Word Format (732KB zip archive)
· Spectrogram - A spectrum analysis tool (257KB zip archive, freeware for non-commercial use only)
Class Performance Summary: As of Exam 1
Grading:
1/3 Homework
1/3 Mid-Term Examination
1/3 Final Exam
There will be no make-up examinations. If you are unable to take one of the mid-term examinations on the assigned date, let me know in advance to make alternative arrangements. If you are unable to take the final examination on the assigned date, then the instructor will record a course grade of incomplete which can be redeemed by taking the final examination when the course is next offered (usually 1 year later).
In case of a class cancellation, use U.S. mail, e-mail, fax, or hand delivery to send in any assignment by the Friday following the due date. The activity scheduled for a cancelled meeting date, whether exam or lecture, is automatically postponed to the next class meeting when it is eventually held. After a cancellation, an updated schedule will be issued.
Each reading assignment should be completed before the weekly meeting. Homework is due at the meeting following the week when it is assigned.
Assignments are due at the beginning of class.
Assignments handed in during class lose 5 points, after class 10 points.
Late submittals lose 10 points per day including weekends and holidays.
Missing an
exam results in a zero unless a written excuse is presented.
Homework
requirements:
Print out a listing of the program. Print
out the program input and output. You may need to do this
at various levels of detail. Hand in a labeled disk with a printout. Place the
disk in a #10 letter
envelope and staple the envelope to the printout.
Topics :
A. Exceptions (ch 10)
B. Nested Reference Data Types (ch 12)
C. Threads (ch 13)
D. Streams (ch 15)
- Overview
of Digital Audio and its application
Current state of streaming Audio on the Internet
Problems and solutions - Digital
Audio Processing Fundamentals
Sampling
Spectra - The
AudioFrame Class
internal data structures
the AudioStream
the AudioDataStream
doubleData
ulawData
the AudioPlayer - Building
GUI’s for Signal Processing
building the menu
intercepting menu event
intercepting keyboard events - Audio
File Formats/Coding
Audio file formats
u-law companding in the Sun AU files
decoding u-law
encoding u-law
reading
writing
playing
graphing - Constructive
Synthesis
Sine
Triangle
Square
Sawtooth - Spectral
Analysis
The Discrete Fourier Transform
The Inverse DFT
The Fast Fourier Transform Class
The Inverse FFT method
Fast Convolution using the FFT
Power Spectral Estimation
Filtering using FFT - Additive
vs. Subtractive Synthesis
- Frequency
shifting using the FFT
- Delay,
echo vs. reverb spectral impact
- OOD
and Multi-media Programming