Chapter 2.

Discrete-Time Signals and Systems

Gao Xinbo
School of E.E., Xidian Univ.
Xbgao@ieee.org
http://see.xidian.edu.cn/teach/matlabdsp/

Main Contents
Important types of signals and their operations
Linear and shift-invariant system
Easier to analyze and implement
The convolution and difference equation
representations
Representations and implementation of signal
and systems using MATLAB
Discrete-time signals
Analog and discrete signals
analog signal
t represents any physical quantity, time in sec.
Discrete signal: discrete-time signal
N is integer valued, represents discrete
instances in times
) (t x
a
) (n x
} ), 1 ( ), 0 ( ), 1 ( , { )} ( { ) ( x x x n x n x = =
Discrete-time signal
In Matlab, a finite-duration sequence representation
requires two vectors, and each for x and n.
Example:




Question: whether or not an arbitrary infinite-duration
sequence can be represented in MATLAB?
} 7 , 3 , 4 , 1 , 0 , 1 , 1 , 2 { ) ( = n x
]; 7 , 3 , 4 , 1 , 0 , 1 , 1 , 2 [
]; 4 , 3 , 2 , 1 , 0 , 1 , 2 , 3 [
=
=
x
n
Types of sequences
Elementary sequence for analysis purposes
1. Unit sample sequence


Representation in MATLAB
{ } , 0 , 0 , 1 , 0 , 0 ,
0 , 0
0 , 1
) (
|
=

=
=
=
n
n
n o
2 0 1 2 1
0
0
0
, ,
, 0
, 1
) ( n n n n n n
n n
n n
n n s s s s

=
=
= o
Function [x,n]=impseq(n
0
,n
1
,n
2
)
A: n=[n1:n2];
x = zeros(1,n2-n1+1); x(n0-n1+1)=1;
B: n=[n1:n2]; x = [(n-n0)==0]; stem(n,x,ro);

-3 -2 -1 0 1 2 3
0
0.2
0.4
0.6
0.8
1
n
o

(
n
-
n
0
)
-3<n<3
n0=0

2. Unit step sequence



{ } , 1 , 1 , 1 , 0 , 0 ,
0 , 0
0 , 1
) (
|
=

<
>
=
n
n
n u
2 0 1 2 1
0
0
0
, ,
, 0
, 1
) ( n n n n n n
n n
n n
n n u s s s s

<
>
=
A: n=[n1:n2]; x=zeros(1,n2-n2+1); x(n0-n1+1:end)=1;
B: n=[n1:n2]; x=[(n-n0)>=0];
3. Real-valued exponential sequence




R a n a n x
n
e = ; , ) (
For Example:
10 0 , ) 9 . 0 ( ) ( s s = n n x
n
n=[0:10]; x=(0.9).^n; stem(n,x,ro)
0 1 2 3 4 5 6 7 8 9 10
0
0.2
0.4
0.6
0.8
1
4. Complex-valued exponential sequence




n e n x
n j
=
+
, ) (
) (
0
e o
Attenuation:
frequency in radians:

For Example: n=[0:10]; x=exp((2+3j)*n);
5. Sinusoidal sequence


n n n x + = ), cos( ) (
0
u e
Phase in radians

For Example:
n=[0:10]; x=3*cos(0.1*pi*n+pi/3)+2*sin(0.5*pi*n)
6. Random sequence
Rand(1,N)
Generate a length N random sequence whose
elements are uniformly distributed between [0,1]
Randn(1,N)
Generate a length N Gaussian random sequence
with mean 0 and variance 1. en [0,1]
7. Periodic sequence
A sequence x(n) is periodic if x(n)=x(n+N)
The smallest integer N is called the
fundamental period
For example
A: xtilde=[x,x,x,x]
B: xtilde=x*ones(1,P); xtilde=xtilde(:);
xtilde=xtilde; transposition
Operations on sequence
1. Signal addition
Sample-by-sample addition
{x1(n)}+{x2(n)}={x1(n)+x2(n)}

Function [y,n]=sigadd(x1,n1,x2,n2)
n=min(min(n1),min(n2)): max(max(n1),max(n2));
y1=zeros(1,length(n)); y2=y1;
y1(find((n>=min(n1)) & (n<=max(n1))==1))=x1;
y2(find((n>=min(n2)) & (n<=max(n2))==1))=x2;
Y=y1 + y2;
2. Signal multiplication
Sample-by-sample multiplication
Dot multiplication
{x1(n)}.{x2(n)}={x1(n) x2(n)}

Function [y,n]=sigmult(x1,n1,x2,n2)
n=min(min(n1),min(n2)) : max(max(n1),max(n2));
y1=zeros(1,length(n)); y2=y1;
y1(find((n>=min(n1)) & (n<=max(n1))==1))=x1;
y2(find((n>=min(n2)) & (n<=max(n2))==1))=x2;
Y=y1 .* y2;
3. Scaling
a{x(n)}={ax(n)}

5. folding
y(n)={x(n-k)}
m=n-k; y=x;

4. Shifting
y(n)={x(-n)}
y=fliplr(x); n=-fliplr(n);
6. Sample summation


ss = sum(x(n1:n2);

7. Sample production


sp = prod(x(n1:n2));

=
+ + =
2
1
) ( ) ( ) (
2 1
n
n n
n x n x n x
) ( ) ( ) (
2 1
2
1
n x n x n x
n
n n
=
H
=

8. Signal energy


se = sum(x .* conj(x)); or
se = sum(abs(x) .^ 2);

9. Signal power



+
=
+
=
= =
n n
x
n x n x n x
2 *
| ) ( | ) ( ) ( c

=
=
1
0
2
| ) ( |
1
N
n
x
n x
N
P
Examples
Ex020100 composite basic sequences
Ex020200 operation on sequences
Ex020300 complex sequence generation
Ex020400 even-odd decomposition
Some useful results
Unit sample synthesis
Any arbitrary sequence can be synthesized as a weighted
sum of delayed and scaled unit sample sequence.


Even and odd synthesis
Even (symmetric): x
e
(-n)=x
e
(n)
Odd (antisymmetric): x
o
(-n)=-x
o
(n)
Any arbitrary real-valued sequence can be decomposed
into its even and odd component: x

(n)=x
e
(n)+ x
o
(n)

+
=
=
k
k n k x n x ) ( ) ( ) ( o
)] ( ) ( [
2
1
) (
)] ( ) ( [
2
1
) (
n x n x n x
n x n x n x
o
e
=
+ =
Function [xe, x0, m] = evenodd(x,n)
If any(imag(x) ~= 0)
error(x is not a real sequence);
End
m = -fliplr(n);
m1 = min([m,n]); m2 = max([m,n]); m=m1:m2;
nm = n(1)-m(1); n1 = 1:length(n);
x1 = zeros(1, length(m)); x1(n1+nm) = x; x = x1;
xe = 0.5 * (x + flipflr(x));
xo = 0.5*(x - fliplr(x));
The geometric series
A one-side exponential sequence of the form {a
n
, n>=0},
where a is an arbitrary constant, is called a geometric
series.




Expression for the sum of any finite number of terms of
the series
1 | | ,
1
1
0
<

=
a f or
a
a
n
n
a
a
a
a
N
N
n
n

=
,
1
1
1
0
Correlations of sequences
It is a measure of the degree to which two sequences are
similar. Given two real-valued sequences x(n) and y(n) of
finite energy,
Crosscorrelation


Autocorrelation

+
=
=
n
y x
l n y n x l r ) ( ) ( ) (
,

+
=
=
n
x x
l n x n x l r ) ( ) ( ) (
,
The index l is called the shift or lag
parameter.
The special case: y(n)=x(n)
Discrete Systems
Mathematically, an operation T[.]
y(n) = T [ x(n)]
x(n): excitation, input signal
y(n): response, output signal
Classification
Linear systems
Nonlinear systems
Linear operation L[.]
Iff L[.] satisfies the principle of superposition


The output y(n) of a linear system to an arbitrary input x(n)


is called impulse response, and is denoted by h(n,k)
) ( ), ( , ,
)] ( [ )] ( [ )] ( ) ( [
2 1 2 1
2 2 1 1 2 2 1 1
n x n x a a
n x L a n x L a n x a n x a L

+ = +
] ) ( [ ) ( ) ( ) ( )] ( [ ) (

+
=
+
=
=
(

= =
n n
k n L k x k n k x L n x L n y o o
)] ( [ k n L o

+
=
=
n
k n h k x n y ) , ( ) ( ) (
h(n,k): the time-varying impulse response
Linear time-invariant (LTI) system
A linear system in which an input-output pair is invariant to a shift n in
time is called a linear times-invariant system
y(n) = L[x(n)] --- y(n-k) = L[x(n-k)]


The output of a LTI system is call a linear convolution sum



An LTI system is completely characterized in the time domain by the
impulse response h(n).
) ( )] ( [ ) , ( k n h k n L k n h = = o
) ( * ) ( ) ( ) ( )] ( [ ) ( n h n x k n k k x n x LTI n y
k

+
=
A
= = =
Properties of the LTI system
Stability
A system is said to be bounded-input bounded-output
(BIBO) stable if every bounded input produces a
bounded output.
Condition: absolutely summable


To avoid building harmful systems or to avoid burnout
or saturation in system operation

+
=
<
n
n h Stability BIBO | ) ( |
Properties of the LTI system
Causality
A system is said to be causal if the output at index n
0

depends only on the input up to and including the index n
0

The output does not depend on the future values of the
input
Condition: h(n) = 0, n < 0
Such a sequence is termed a causal sequence.

To make sure that systems can be built.
Convolution
Convolution can be evaluated in many different ways
If the sequences are mathematical functions, then we can
analytically evaluate x(n)*h(n) for all n to obtain a
functional form of y(n)
Graphical interpretation, folded-and-shifted version
Matlab implementation
Function [y,ny]=conv_m(x,nx,h,nh)
nyb = nx(1)+nh(1); nye = nx(length(x))+nh(length(h));
ny = [nyb:nye];
n = conv(x,h)
Function form of convolution


) ( ) 9 . 0 ( 1 ) ( * ) ( ) (
) ( ) 9 . 0 ( ) ( ), 10 ( ) ( ) (
9
0
k n u n h n x n y
n u n h n u n u n x
k n
k
n
= =
= =

Three different conditions under which u(n-k) can be

evaluated:
Case 1: n<0 % the nonzero values of x(n)and y(n) do not overlap.
Case 2: 0<=n<9 % partially overlaps
Case 3: n>=9 % completely overlaps
Folded-and-shifted
Signals x=[x(1),x(2),x(3),x(4),x(5)]
System Impulse Response: h=[h(1),h(2)h(3),h(4)]
y=conv(x,h)
y(1)=x(1)*h(1); y(2)=x(1)*h(2)+x(2)*h(1)
y(3)=x(1)*h(3)+x(2)*h(2)+x(3)*h(1);
x(1),x(2),x(3),x(4),x(5)
h(4),h(3),h(2),h(1)
Note that the resulting sequence y(n) has a longer length than both the x(n)
and h(n) sequence.
Sequence correlations revisited
The correlation can be computed using the conv function if
sequences are of finite duration.





Example 2.8
The meaning of the crosscorrelation
This approach can be used in applications like radar signal
processing in identifying and localizing targets.


+
=
+
=
+
=
+
=
= = =
= = =
k k
xx
k k
yx
n x n x n k x n x k n x n x n r
n x n y n k x n y k n x n y n r
) ( * ) ( )) ( ( ) ( ) ( ) ( ) (
) ( * ) ( )) ( ( ) ( ) ( ) ( ) (
Difference Equation
An LTI discrete system can also be described by a linear
constant coefficient difference equation of the form


If a
N
~= 0, then the difference equation is of order N
It describes a recursive approach for computing the
current output,given the input values and previously
computed output values.
n m n x b k n y a
M
m
m
N
k
k
=

= =
, ) ( ) (
0 0

= =
=
M
m
N
k
k m
k n y a m n x b n y
0 1
) ( ) ( ) (
Solution of difference equation
y(n) = y
H
(n) + y
P
(n)
Homogeneous part: y
H
(n)
Particular part: y
P
(n)
Analytical approach using Z-transform will
be discussed in Chapter 4
Numerical solution with Matlab
y = filter(b,a,x)
Example 2.9
Zero-input and Zero-state response
In DSP the difference equation is generally solved forward in
time from n=0. Therefore initial conditions on x(n) and y(n)
are necessary to determine the output for n>=0.


Subject to the initial conditions:

0 , ) ( ) ( ) (
0 1
> =

= =
n k n y a m n x b n y
M
m
N
k
k m
} 1 ); ( { } 1 ); ( { s s s s n M n x and n N n y
) ( ) ( ) ( n y n y n y
ZS ZI
+ =
Solution:
Zero-input and Zero-state response
y
ZI
(n): zero-input solution
A solution due to the initial conditions alone
y
ZS
(n): zero-state solution
A solution due to input x(n) alone
Digital filter
Discrete-time LTI systems are also called digital filter.
Classification
FIR filter & IIR filter
FIR filter
Finite-duration impulse response filter
Causal FIR filter

h(0)=b
0
,,h(M)=b
M
Nonrecursive or moving average (MA) filter
Difference equation coefficients, {b
m
} and {a
0
=1}
Implementation in Matlab: Conv(x,h); filter(b,1,x)

=
=
M
m
m
m n x b n y
0
) ( ) (
IIR filter
Infinite-duration impulse response filter
Difference equation

Recursive filter, in which the output y(n) is
recursively computed from its previously
computed values
Autoregressive (AR) filter

=
=
N
k
k
n x k n y a
0
) ( ) (
ARMA filter
Generalized IIR filter


It has two parts: MA part and AR part
Autoregressive moving average filter, ARMA
Solution
filter(b,a,x); %{b
m
}, {a
k
}
0 , ) ( ) ( ) (
0 1
> =

= =
n k n y a m n x b n y
M
m
N
k
k m
Reference and Assignment
Textbook: pp1 to pp35
Chinese reference book: pp1 to pp18

,20011

Exercises:
Textbook: p2.1b,c; p2.2b,d; 2.5
Textbook: P2.12b, 2.15, 2.17b, 2.8