CHAPTER I
INTRODUCTION TO DIGITAL SIGNAL
PROCESSING
Abdelhamid DAAMOUCHE,
University M’hamed Bougara,
Boumerdes
1
Main features
• -Digital signal processing techniques have
reached high performance, which cannot be
achieved by using analog signal processing,
• -Examples are linear phase Finite Impulse
Response (FIR) filters and adaptive filters,
• -Analog systems use resistors, capacitors, and
inductors.
• -These components are very sensitive to
temperature, voltage changes, and mechanical
vibrations.
2
Main features
• -Using digital signal processors (DSPs), it is
easy to change applications and insert any
needed modifications,
• -Using DSPs reduces noise, cost, and power
consumption,
• Drawbacks:
• -High frequency signals cannot be processed
with DSPs because ADC cannot work fast
enough
3
Main features
• - Real-time processing cannot be reached.
• -Quantization errors due to the finite word-
length signal representation,
4
General purpose DSP system
• -Many applications require large number of
mathematical operations,
• -DSPs are built to speed up the arithmetic
opearations,
• -Most of digital signal processing algorithms
such as filtering, convolution, and fast Fourier
transform, require multiply-and-accumulate
opeartions,
5
General purpose DSP system
• -Usually conversion from analog-to-digital (ADC) and
digital-to-analog (DAC) are required,
• -Fig 1.2 shows a general purpose block
diagram of a DSP system,
6
General purpose DSP system
• -The first anti-aliasing filter is usually designed
according to the input signal characteristics,
• -It eliminates the high frequency components
to allow an efficient ADC process and prevent
aliasing,
• -The DS processor performs the required
arithmetic operations,
7
General purpose DSP system
• -The DAC delivers a stair-case signal, hence a
reconstruction filter is needed to smooth out
the output signal .
• Typical DSP tasks:
• -Convolution,
DSPs are designed to
• -DFT, perform multiplication and
addition in hardware and in
• -FIR, one cycle
• -IIR.
8
Review on sequences
• -A discrete-time signal is a complex-valued
sequence. It is denoted by
x[n], where n is an integer,
Basic sequences
Used as building blocks to construct and
represent other signals.
9
Review on sequences
• The unit impulse sequence (Delta function)
• It is given by
1, n 0
[ n]
0, n 0
10
Review on sequences
• The unit step sequence
• It is defined by
1, n 0
u[n]
0, n 0
11
Review on sequences
• Any sequence can be expressed as
x[n] k
x[k ] [n k ]
12
Review on sequences
• Using the unit impulse and the unit step
sequences, we can derive some common
expressions used in the analysis of systems:
u[ n] [n] [ n 1] [ n 2]
k 0
[n k ]
• Making a change of variable m=n-k, we get
n
u[n] m
[ m]
13
Review on sequences
• The above expression is known as the running
sum of the unit impulse function.
• Also, we have:
u[n ] [n] [n 1] [n 2]
u [ n 1)
[n ] u[n 1]
[n ] u[n ] u[n 1]
14
Discrete-time systems
• A discrete-time system is defined as a
transformation that maps an input sequence
with values x[n] into an output sequence with
values y[n]. That is y[n]=T{x[n]}
15
Discrete-time systems
• Examples:
• 1) Accumulator
n
y[n] m
x[m]
• 2) Squarer
2
y[ n ] x[ n ]
• 3)Delayer
y[n] x[n n0 ]
16
Linear Time-Invariant (LTI) Systems
• The two most important properties of systems
are the linearity and time-invariance. When
both properties are satisfied by a system, they
produce a very well known representation of
systems, it is called convolution.
• -In general, we have
x[n] k
x[k] [n k]
17
Linear Time-Invariant (LTI) Systems
• Then, y[n] T x[n]
T k
x[k ] [n k ]
• Since the system is linear, we can write
• y[n] x[k ]T [n k ]
k
• [n k] T{.} hk [n ]
18
Linear Time-Invariant (LTI) Systems
• Then
y[n] k
x[k ]hk [n]
• If, in addition to linearity, the system is TI, we
get
y[n] k
x[k ]h[n k ]
• It is called the convolution sum. Where h[n] is
known as the impulse response of the system.
19
The discrete-time Fourier transform
• -The Fourier transform of a signal is an
alternative representation of the data in the
signal.
• -The discrete-time Fourier transform maps
infinite sequences into the space of 2 -
periodic function.
20
The discrete-time Fourier transform
• -The DTFT of a nonperiodic sequence x[n] is
given by
j j n
X (e ) n
x[n]e
• X (e j
) is periodic with period 2 .
• The inverse DTFT is
x[n] 21 X (e j )e j n d
21
The discrete-time Fourier transform
• Example: Determine the DTFT of the sequence
1; M n M
x[n]
0; otherwise
• Answer: X (e j
)
M
x [ n ]e j n
n M
1 j n M j n
n M
e n 0
e
M j n M j n
n 1
e n 0
e
j j (M 1) j (M 1)
e e 1 e
j n j n
1 e 1 e
sin (M 1
2 )
sin
2
22
The discrete-time Fourier transform
• In general X (e j ) is a complex-valued function.
• -Let’s find the DTFT of x [n] , the symbol
• stands for conjugation.
Y (e j ) n
x [n]e j n
j n
( n
x[n]e )
j
X (e )
23
The discrete-time Fourier transform
• -Let’s find the DTFT of x [ n] :
j j n
Y (e ) n
x [ n]e
j m
m
x [m]e
j m
( m
x[m]e )
X (e j )
24
The discrete-time Fourier transform
• -If x[n] is real, then
j j
x[n] x [n] X (e ) X (e )
• Or
• Re{X(exp(jw)}+Im{X(exp(jw)}=
• Re{X(exp(-jw)}-Im{X(exp(-jw)}
j j
Re[ X (e )] Re[ X (e )]; even
j j
Im[ X (e )] Im[ X (e )]; odd
25
The discrete-time Fourier transform
• Also, 2
j 2 j 2 j
X (e ) Re [ X (e )] Im [ X (e )]
2 j 2 j
Re [ X (e )] Im [ X (e )]
• Then, the modulus of a real sequence is always
an even function in frequency (w).
• This property is always encountered in the design
of digital filters
• -The phase is an odd function in frequency (w).
26
The discrete-time Fourier transform
• -Knowing that the DTFT of x[-n] is X (e j ) , we
can show that:
• -If x[n] is real and even, then X(ej
) is also real
and even,
• - If x[n] is real and odd, then X(ej ) is odd and
purely imaginery.
27
Conclusion
• Throughout this chapter, we have reviewed
the necessary tools that will be exploited in
the upcoming chapters of this course.
• References
• -Oppenheim et al. « Discrete-time signal
processing » Second edition.
• -Proakis et al. « Digital signal processing:
Algorithms and Applications », Fourth edition.
28
