Digital Signal Processing

Design
Lecture 1: Introduction to DSP Design

Dr. Hoang Tuan Anh

School of Electrical Engineering
International University
University, VNU-HCM

September 2010

Digital Signal Processing

1
 Course Outline–
Outline–DSP Design

1. Introduction to DSP Design

2. DSP Concepts
3 Fundamental of speech processing
3.
4. Fundamental of image processing
5. Discrete Linear Transforms: FFT, STFT
6. Digital Filters: FIR Filters
7. FIR Filter Design
8 Digital
8. Di it l Filters:
Filt IIR Filters
Filt
9. Applications of DSP in communications

Term Project

ƒ All registered students are required to do a term project.

ƒ Implemented using Matlab a signal processing system that
accomplishes a simple or even a complex task - e.g., pitch
detection, voiced unvoiced detection, edge detection, digit
recognition, etc.
ƒ Every student/team is also required to submit project report
and make a 15-20 minute Power Point presentation of their
project to the entire class. The presentation must include:
‰ a short description
p of the pproject
j and its objectives
j
‰ an explanation of the implemented algorithm and relevant
theory.
‰ a demonstration of the working program – i.e., results
obtained when running the program.
4

2
 Speech/Audio/Image Processing
Term Projects

1. Pitch detector – time domain, autocorrelation,

cepstrum, LPC, etc.
2. Formant analyzer/tracker
3. Speech coders including ADPCM, LDM, CELP,
Multipulse, etc.
4. N-channel spectral analyzer and synthesizer – phase
vocoder, channel vocoder
p
5. Speech endpoint
p detector
6. Simple speech recognizer – e.g. isolated digits,
speaker trained
7. Audio/music coder
5

Speech/Audio/Image Processing
Term Projects

8. Audio/music filters for voice

9. System to speed up and slow down speech by
arbitrary factors
10. Speaker verification system
11. Speech enhancement system (noise reduction, post
filtering, spectral flattening)
12. Photoshop Emulation in MATLAB
13. Image DeDe-blurring
blurring Methods
14. Object Recognition in Images

3
 Signals

A function of independent variables such as time,

distance, position, temperature, pressure, etc.
ƒ A signal carries information:

ƒ Examples: speech, audio, seismic, image and video

Signals

ƒ A signal can be a function of one, two or N

independent variables
‰ speech
p and audio are 1‐D signals
g ((functions of time))
‰ an image is a 2‐D signal (function of spatial
coordinates)
‰ video is a 3‐D signal (function of spatial coordinates and
time)

4
 Types of Signals

ƒ Analog Signals (Continuous-Time Signals)

‰ Signals that are continuous in both the dependant and
independent variable (e.g., amplitude and time).
‰ Most environmental signals are continuous-time signals.
ƒ Discrete Sequences (Discrete-Time Signals)
‰ Signals that are continuous in the dependant variable (e.g.,
amplitude) but discrete in the independent variable (e.g.,
time).
‰ They are typically associated with sampling of continuous-
time signals.

Types of Signals

ƒ Digital (Discrete‐Time, Discrete‐Amplitude Signal)

‰ Signals that are discrete in both the dependant and
independent variable (e.g., amplitude and time) are digital
signals.
‰ These are created by quantizing and sampling continuous-
time signals or as data signals (e.g., stock market price).

10

5
 Some Typical Signals

11

Compression of Multimedia Signals

12

6
 Compression of Multimedia Signals

13

Bandwidths for Speech and Audio Signals

14

7
 2‐Dimensional Signals

15

Matrix Image Representation

16

8
 What is DSP
DSP??

ƒ Changing or analyzing information that is measured as

discrete sequences of numbers
ƒ The representation, transformation, and manipulation
of signals and the information they contain

17

Processing Real Signals

ƒ Most of the signals in our environment are analog such as

sound, temperature and light
ƒ To processes these signals with a computer, we must:
‰ convert the analog signals into electrical signals, e.g., using a
transducer such as a microphone to convert sound into
electrical signal
‰ digitize these signals, or convert them from analog to digital,
using an ADC (Analog to Digital Converter)
ƒ Processed signal may need to be converted back to an
analog signal

DSP

18

9
 Typical DSP System Components

ƒ Sampler – every TS seconds; fS = 1/TS

ƒ Hold – retain current sample value, x(nTS)
ƒ A/D – analog-to-digital conversion
ƒ Digital Filter – process digitized samples
ƒ D/C – digital-to-continuous
di it l t ti transformation,
t f ti TS’
ƒ D/A – digital-to-analog conversion
ƒ Analog LPF – eliminate any remaining signal out-of
band (images)
19

Example

20

10
 Applications of DSP

ƒ Communication: encoding and decoding of digital

communication signals (e.g., speech, audio, image,
) detection, equalization,
video), q filtering,
g direction
finding, echo cancellation
ƒ Radar and Sonar: target detection, position and
velocity estimation, tracking
ƒ Biomedical Engineering: analysis of biomedical
signals, diagnosis, patient monitoring, preventive
h lth care, artificial
health tifi i l organs

21

Applications of DSP

ƒ Speech Processing: speech synthesis, speech

recognition, speaker recognition, speech enhancement,
g g understanding,
natural language g language
g g translation
ƒ Image Processing: image enhancement, image
segmentation and classification, image understanding
ƒ Video Processing: video enhancement, video
segmentation and classification, video understanding
ƒ Signal/data compressions

22

11
 DSP Functions

ƒ Common features of DSP applications

‰ they use a lot of multiplying and adding operations
‰ they
y deal with signals
g that come from the real world
‰ they require a certain response time
ƒ Key DSP operations
‰ filtering
‰ correlation
‰ discrete transformation

23

Filtering Example

ƒ Signals are usually a mix of “useful” information and

noise
ƒ How do we extract the useful information?
‰ filtering is one way
y[n] = x[n].h[n]

24

12
 Transform Example

ƒ Can you say which is “1” / ”#” by looking at them?

ƒ If not, go to “frequency” domain
‰ another way to look at signals
‰ done using transforms

25

Transform Example (cont.)

26

13
 Transform Equations

ƒ Discrete‐Time Fourier Transform (DTFT)

x – Time domain signal
X – Frequency domain representation of x

27

Correlation Example

ƒ Provides a measure of similarity between 2 signals

ƒ Typical application is locating a known signal
E.g.,
E g transmit a signal and see if you receive it back
and also at what time you receive it back

28

14
 Correlation Example (cont.)

ƒ Using radar, we
transmit the signal
shown below

ƒ We receive the
following (note the
noise!)

29

Correlation Example (cont.)

30

15
 Correlation Equations

ƒ Correlation
x – Transmitted signal
y – Received signal
rxy – Correlation coefficients

31

Color Image Coding

32

16
 Why do we need DSPs
DSPs??

ƒ Digital signal processors are microprocessors

specifically designed to handle digital signal
processingg tasks.
p
ƒ DSP operations require many calculations of the form:
A=BxC+D
‰ This simple equation involves a multiply and an add
operation
‰ The multiply instruction in a general purpose processor
is very
er slo
slow compared with
ith the add instruction
instr ction
‰ Motorola 68000 microprocessor uses
• 10 clock cycles for add
• 74 clock cycles for multiply
33

Why do we need DSPs

DSPs??

ƒ Digital signal processors can perform the multiply and

the add operation in just one clock cycle
‰ most DSPs have a specialized
p instruction that causes
them to multiply, add and save the result in a single
cycle
‰ this instruction is called a MAC (Multiply, Add, and
Accumulate)
‰ DSPs aim to minimize cost, power, memory use, and
development time

34

17
 Digital Signal Processor Architectures

Von Neuman
ƒ Von Neuman machines store program and data in the
same memory area with a single bus
ƒ an instruction contains the operation command and the
address of data to be operated on (operand)
ƒ most of the general‐purpose microprocessors such as
Motorola 68000 and Intel 80x86 use this architecture
ƒ it is simple
p in hardware implementation,
p , but the data
and program are required to share a single bus

35

Digital Signal Processor Architectures

Harvard architecture
ƒ the only difference in Harvard architecture is that
program and data memories are separated and use
physically separate transmission paths
ƒ enables the machine to transfer instructions and data
simultaneously‐‐ enhances performance
ƒ the Harvard architecture is more commonly used in
specialized microprocessors for real‐time and
embedded applications
ƒ however, not many DSP chips use the Harvard
architecture because of the cost

36

18
 Digital Signal Processor Architectures

Modified Harvard architecture

ƒ cost penalty with the Harvard architecture, which needs
twice as many address and data pins on the chip
ƒ to balance cost and performance, modified Harvard
architecture is used in most DSPs
ƒ uses single data and address bus externally but
internally there are two separate busses for program
and data
ƒ the separation of program and data information is done
by timing (multiplexing)
ƒ for one clock cycle, program information flows on the
pins, and in second cycle data follows on the same pins
37

Moore’s Law Growth in Computing Resources

38

19
 MATLAB

MATLAB is an interactive, matrix‐based system for

scientific and engineering numeric computation and
visualization
‰ strength: complex numerical problems can be solved
easily with a programming language similar to C
‰ can be easily extended to create new commands and
functions
‰ ideal software tool for studying digital signal processing
‰ graphing
hi capability
bilit makes
k it possible
ibl to
t view
i results
lt off
processing and provide insight into complicated
operations

39

MATLAB

ƒ Matlab programming is vector‐based

‰ should rarely need to use loops
‰ can do most operations
p on vectors or matrices
e.g., in C: in Matlab:
for i = 1:10 c = a+b
c(i) = a(i) + b(i) d = a.*b
d(i) = a(i)*b(i)
end

40

20