Professional Documents
Culture Documents
Fundamentals of
Digital Signal Processing Lecture 1:
Course Logistics & Introduction
Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #1 Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #2
• To establish the idea of using computing techniques • Discrete-index signals • Discrete & Fast Fourier
to alter the properties of a signal for desired effects, Transform
via understanding of • Linear systems
– Fundamentals of discrete-time, linear, shift-invariant signals • z-Transforms • Special Filtering Topics
and systems in • Sampling & – Inverse & Matched
• Representation: sampling and quantization; Filtering
Quantization
• Processing: filtering and transform techniques;
• Frequency Response • Time-Frequency Analysis
• Processing System Design: filter & processing
algorithm design. • IIR and FIR Filters • Non-linear Filtering
– Efficient computational algorithms and their implementation.
• Filter Design Techniques
• To gain preliminary experience in computational • Symbolic Processing*
processing of real signal and to relate the above
understanding to real world scenario
Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #3 Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #4
Prerequisites Instructor Info & Office Hours
Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #5 Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #6
Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #7 Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #8
Textbook What is DSP?
Input Output
• Oppenheim and Schafer, Discrete- Signal Signal
Analog-to- Digital-to-
Time Signal Processing, 2nd edition
(Prentice-Hall, 1999)
Digital Computer Analog
Conversion Conversion
• Supplemental texts:
– Hayes, Digital Signal Processing Digital
(Schaum’s Outlines Series), 1999 • Method to represent a quantity, a phenomenon or an event
– McClellan, Schafer, & Yoder, DSP • Why digital?
First Signal
• What is a signal?
– Both on reserve in library – something (as a sound, gesture, or object) that conveys notice or information;
– a detectable physical quantity (as a voltage, current, or magnetic field strength) by
• Some class notes which messages or information can be transmitted
– primarily for special topics & non- • What are we interested in?
linear techniques Processing
• What kind of processing we need and encounter almost everyday?
• Special effects?
Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #9 Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #10
Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #11 Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #12
Digital Processing of Analog Signals Discrete-time Signal
Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #13 Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #14
y
6 1.8
representation that assumes
6
6
Analog sinusoid, -2
Discrete
5 1.2 one out of a finite set of values 4 5sin(2πx)
-4
4 sinusoid
round[5sin(2πnT)]
4 0.6 2
• The finite set of output values -6
0 5 10 15 20 25 30 35 40
2
y
0.3 0.9 1.5 2.1
2
has an index of 6, or (110)2 in -2 6
Quantized
-2
y
quantization table is needed -2
-6
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
x
Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #15 Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #16
The Sampling Theorem Basic Interests & Issues in DSP
Sampled 1000 Hz and 7000 Hz Cosine Waves; fs = 8000 Hz
• Digital Filters: Filter design, noise analysis, structures
1
• Fourier Analysis: Spectrum estimation, FFT, cosine transform,
0.5
amplitude
cepstrum, short-time FT
0 • Signal Modeling and Analysis: Linear prediction, wavelets,
-0.5 chaos, fractals
-1 • Hardware and Software: Minicomputers, array processors, DSP
chips, workstations, PCs, MATLAB, real-time operating systems
0 0.2 0.4 0.6 0.8 1 1.2
time in ms • Applications: Speech, radar, image, video, data, ...
Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #17 Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #18
1997 Kilby
Medallists
Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #19 Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #20
State-of-the-Art Graphics, 1965 DSP in 1967
Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #21 Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #22
Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #23 Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #24
FIR Bandpass Digital Filter Parks and McClellan, 1972
Coefficients for FIR Bandpass Filter
0.4
0.2
-0.2
-0.4
0 20 40 60 80 100
sample index m
Frequency Response of FIR Equiripple Bandpass Filter
log magnitude in dB
-50
-100
0 0.5 1 1.5 2 2.5 3 3.5 4
frequency in kHz
Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #25 Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #26
N −1 kn
X[k] = ∑ x[n]WN k = 0,1,K, N −1
n=0
WN ≡ e − j (2π / N )
1 N −1
x[n] = ∑ X [k ]WN− kn , n = 0,1,K, N − 1
N k =0
• Exact representation of finite-length or periodic sequences
(x[n+N]=x[n]).
• X[k] and x[n] can be computed efficiently by the FFT. (Gauss
knew about it, Cooley and Tukey rediscovered it at just the right
time.)
Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #27 Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #28
Cooley and Tukey, 1965 Jim Cooley at Arden House, 1968
Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #29 Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #30
“Oppenheim and Schafer”, 1975 Jim Flanagan, Larry Rabiner, & Ron Schafer
1999 Kilby
Medallist
Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #31 Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #32
“Rabiner and Gold”, 1975 Linear Predictive Signal Modeling
Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #33 Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #34
Bishnu Atal
Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #35 Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #36
Spectrum Analysis with FFT & LPC DSP Applications in Speech Recognition
Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #37 Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #38
Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #39 Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #40
William Pratt, 1978
Multidimensional DSP Pioneers
Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #41 Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #42
• Common algorithms
– Linear filtering: FIR, IIR
– FFT, cosine transform, filterbanks
– Correlation
– Matrix calculations
• Common features
– Lots of multiply/accumulate operations
– Block processing is often appropriate
– Fixed-point arithmetic for economical solutions
– Real-time operation
Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #43 Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #44
The Rockland Digital Filter, 1971 Speak and Spell, 1978
For the price of a small house, you could have one of these.
Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #45 Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #46
TMS320-C31
Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #49 Summer 2004 ECE 4270 B. H. Juang Copyright 2004 Lecture #1, Slide #50