Chapter 2

Discrete-time Signals and Systems

2.1 Discrete-time signals
Digital signals are discrete in both time (the independent variable) and amplitude (the
dependent variable). Signals that are disrete in time but continuous in amplitude are
referred to as discrete-time signals.
Discrete-time signals are data sequences. A sequence of data is denoted {x[n]} or simply
x [n] when the meaning is clear. The elements of the sequence are called samples. The
index n associated with each sample is an integer. If appropriate, the range of n will be
specified. Quite often, we are interested in identifying the sample where n = 0. This is
done by putting an arrow under that sample. For instance,
{x[n]} ={. . . ,0.35,1,1.5,0.6, 2,. . .}

The arrow is often omitted if it is clear from the context which sample is x[0]. Sample
values can either be real or complex. In the rest of this book, the terms discrete-time
signals and sequences are used interchangeably.
The time interval between samples is not explicitly shown. It can be assumed to be
normalized to 1 unit of time. So the corresponding normalized sampling frequency is 1
Hz. If the actual sampling interval is T seconds, then the sampling frequency is given by
fs = T1 .

20

Discrete-time Signals and Systems

2
1.8
1.6
1.4

Amplitude

1.2
1
0.8
0.6
0.4
0.2
0
3

Index (n)

Figure 2.1: The Unit Impulse Sequence

2.1.1 Some Elementary Sequences

There are some sequence that we shall encounter frequently. They are described here.
2.1.1.1 Unit Impulse Sequence
The unit impulse sequence is defined by
[n] =

1, n = 0
0, n =
6 0

(2.1)

This is depicted graphically in Figure 2.1. Note that while the continuous-time unit impulse function is a mathematical object that cannot be physically realized, the unit impulse sequence can easily be generated.
2.1.1.2 Unit Step Sequence
The unit step sequence is one that has an amplitude of zero for negative indices and an
amplitude of one for non-negative indices. It is shown in Figure 2.2.

Discrete-time signals

21

2
1.8
1.6
1.4

Amplitude

1.2
1
0.8
0.6
0.4
0.2
0
3

Index (n)

Figure 2.2: The Unit Step Sequence

2.1.1.3 Sinusoidal Sequences
A sinusoidal sequence has the form
x[n] = A cos (0 n + )

(2.2)

This function can also be decomposed into its in-phase xi [q] and quadrature xq [n] components.
x[n] = A cos cos 0 n A sin sin 0 n
= xi [n] + xq [n]

(2.3)
(2.4)

This is a common practice in communications signal processing.

2.1.1.4 Complex Exponential Sequences
Complex exponential sequences are essentially complex sinusoids.
x[n] = Aej(o n+)
= A cos(0 n + ) + jA sin(0 n + )

(2.5)
(2.6)

22

Discrete-time Signals and Systems

0.5
0.4
0.3
0.2

Amplitude

0.1
0
0.1
0.2
0.3
0.4
0.5

10

15
Index (n)

20

25

30

Figure 2.3: Uniformly distributed random sequence with amplitudes between -0.5 and
0.5.
2.1.1.5 Random Sequences
The sample values of a random sequence are randomly drawn from a certain probability
distribution. They are also called stochastic sequences. The two most common distributions are the Gaussian (normal) distribution and the uniform distribution. The zero-mean
Gaussian distribution is often used to model noise. Figure 2.3 and figure 2.4 show examples of uniformly distributed and Gaussian distributed random sequences respectively.

2.1.2 Types of Sequences

The discrete-time signals that we encounter can be classified in a number of ways. Some
basic classifications that are of interest to us are described below.
Real vs. Complex Signals
A sequence is considered complex at least one sample is complex-valued.

Discrete-time signals

23

2.5

1.5

Amplitude

0.5

0.5

1.5

10

15
Index (n)

20

25

30

Figure 2.4: Gaussian distributed random sequence with zero mean and unit variance.
Finite vs. Infinite Length Signals Finite length sequences are defined only for a range
of indices, say N1 to N2 . The length of this finite length sequence is given by |N2 N1 + 1|.
Causal Signals A sequence x[n] is a causal sequence if x[n] = 0 for n < 0.
Symmetric Signals First consider a real-valued sequence {x[n]}. Even symmetry implies that x[n] = x[n] and for odd symmetry x[n] = x[n] for all n.

Any real-valued sequence can be decomposed into odd and even parts so that
x[n] = xe [n] + xo [n]
where the even part is given by
xe [n] =

1
(x[n] + x[n])
2

xo [n] =

1
(x[n] x[n])
2

and the odd part is given by

24

Discrete-time Signals and Systems

A complex-valued sequence is conjugate symmetric if x[n] = x? [n]. The sequence has

conjugate anti-symmetry if x[n] = x? [n]. Analogous to real-valued sequences, any
complex-valued sequence can be decomposed into its conjugate symmetric and conjugate
anti-symmetric parts:
x[n] = xcs [n] + xca [n]
1
xcs [n] =
(x[n] + x? [n])
2
1
xca [n] =
(x[n] x? [n])
2

(2.7)
(2.8)
(2.9)

Periodic Signals A discrete-time sequence is periodic with a period of N samples if

x[n] = x[n + kN ]

(2.10)

for all integer values of k. Note that N has to be a positive integer. If a sequence is not
periodic, it is aperiodic or non-periodic.
We know that continuous-time sinusoids are periodic. For instance, the continuous-time
signal
x(t) = cos (0 t)
(2.11)
has a frequency of 0 radians per second or f0 = 0 /2 Hz. The period of this sinusoidal
signal is T = 1/f0 seconds.
Now consider a discrete-time sequence x [n] based on a sinusoid with angular frequency
0 :
x[n] = cos (0 n)
(2.12)
If this sequence is periodic with a period of N samples, then the following must be true:
cos 0 (n + N ) = cos 0 n

(2.13)

However, the left hand side can be expressed as

cos 0 (n + N ) = cos (0 n + 0 N )

(2.14)

and the cosine function is periodic with a period of 2 and therefore the right hand side
of (2.13) is given by
cos 0 n = cos (0 n + 2r)
(2.15)

Discrete-time signals

25

for integer values of r. Comparing (2.14) with (2.15), we have

0 N =2r
2f0 N =2r
r
f0
=
N

(2.16)

where 0 = 2f0 . Since both r and N are integers, a discrete-time sinusoidal sequence
is periodic if its frequency is a rational number. Otherwise, it is non-periodic.


Example 2.1. Is x[n] = cos 8 n periodic? If so, what is the period?
The sequence can be expressed as

x[n] = cos 2

1
16

 
n

So in this case, f0 = 1/16 is a rational number and the sinusoidal sequence is periodic
with a period N = 16.
Example 2.2. Determine the fundamental period of the following sequence:
x[n] = cos (1.1n) + sin (0.7n)
Solution:
For the cosine function, the angular frequency is
1 = 1.1 = 2 (0.55) = 2f1
Therefore,
55
100
11
=
20

f1 =

and the period is N1 = 20.

For the sine function, the angular frequency is
2 = 0.7 = 2 (0.35) = 2f2

26

Discrete-time Signals and Systems

where
35
100
7
=
20

f2 =

and the period is N2 = 20.

So the period of x[n] is 20.
It is interesting to note that for discrete-time sinusoidal sequences, a small change in
frequency can lead to a large change in period. For example, a certain sequence has
frequency f1 = 0.51 = 51/100. So its period is 100 samples. Another sinusoidal
sequence with frequency f2 = 0.5 = 50/100 has a period of only 2 samples since f2
can be simplified to 1/2. Thus a frequency difference of 0.01 can cause the period of the
two sinusoidal sequences to differ by 98 samples!
Another important point to note is that discrete-time sinusoidal sequences with frequencies separated by an integer multiple of 2 are identical.
Energy and Power Signals

The energy of a finite length sequence x[n] is defined as

E=

N2
X

|x[n]|2

(2.17)

|x[n]|2

(2.18)

n=N1

while that for an infinite sequence is

E=

n=

Note that the energy of an infinite length sequence may not be infinite. A signal with
finite energy is usually referred to as an energy signal.
Example 2.3. Find the energy of the infinite length sequence
 n
2 , n0
x[n] =
0,
n<0

Discrete-time signals

27

According to the definition, the energy is given by

E=

2n

n=0

To evaluate the finite sum, first consider

SN =

N
1
X
n=0

 n
X
1
n=0

an = 1 + a + a2 + + aN 1

(2.19)

Multiplying this equation by a, we have

aSN = a + a2 + + aN

(2.20)

and the difference between these two equations give

SN aSN = (1 a) SN = 1 aN
Hence if a 6= 1,
SN =

1 aN
1a

For a = 1, it is obvious that SN = N . For a < 1, the infinite sum is therefore

1
S =
1a

(2.21)

(2.22)

(2.23)

Making use of this equation, the energy of the signal is

1
4
E=
=
1 1/4
3

Equations (2.22) and (2.23) are very useful and we shall be making use of them later.
The average power of a periodic sequence with a period of N samples is defined as
N 1
1 X
Px =
|x[n]|2
N n=0

(2.24)

and for non-periodic sequences, it is defined in terms of the following limit if it exists:
K
X
1
Px = lim
|x[n]|2
K 2K + 1
n=K

A signal with finite average power is called a power signal.

(2.25)

28

Discrete-time Signals and Systems

Example 2.4. Find the average power of the unit step sequence u[n].
The unit step sequence is non-periodic, therefore the average power is

X
1
u2 [n]
P = lim
K 2K + 1
n=0
K +1
K 2K + 1
1
=
2

lim

Therefore the unit step sequence is a power signal. Note that its energy is infinite and so
it is not an energy signal.
Bounded Signals A sequence is bounded if every sample of the sequence has a magnitude which is less than or equal to a finite positive value. That is,
|x[n]| Bx <

(2.26)

Summable Signals A sequence is absolutely summable if the sum of the absolute value
of all its samples is finite.

X
|x[n]| <
(2.27)
n=

A sequence is square summable if the sum of the magnitude squared of all its samples is
finite.

X
|x[n]|2 <
(2.28)
n=

2.1.3 Some Basic Operations on Sequences

Scaling Scaling is the multiplication of a scalar constant with each and every sample
value of the sequence. This operation is depicted schematically in Figure 2.5.
Addition Addition of two sequences usually refers to point-by-point addition as shown
in Figure 2.6.

Discrete-time signals

29

x[n]

Ax[n]

Figure 2.5: Scalar Multiplication by A.

x[n]

y[n]=x[n]+w[n]

w[n]

Figure 2.6: Point-by-point Addition

Delay A unit delay shifts a sequence by one unit of time as shown in Figure 2.7. A
sequence x[n] is delayed by N samples to produce y[n] if y[n] = x[n N ].
Up/Down Sampling Down-sampling by a factor of L (a positive integer) is an operation by which only one every L-th sample of the original sequence is kept, with the rest
discarded. The schematic for a down-sampler is shown in Figure 2.8. The down-sampled
signal y[n] is given by
(2.29)

y[n] = x[nM ]

Up-sampling is the opposite operation to down-sampling. It increases the number of

samples of the original sequence by a certain factor L (a positive integer). This is done by
inserting L 1 zeros between each pair of original samples. Figure shows the schematic
x[n]

Figure 2.7: A Unit Delay

x[n1]

30

Discrete-time Signals and Systems

x[n]

y[n]

Figure 2.8: Down-sampling by a factor of L.

x[n]

y[n]

Figure 2.9: Up-sample by a factor of L.

diagram of an up-sampler and the up-sampled sequence y[n] is given by
 n
x L , n = 0, L, 2L, . . .
y[n] =
0,
otherwise

(2.30)

An interpolated signal can be obtained by passing the up-sampled signal through a lowpass filter with an appropriate bandwidth. This process will be discussed in more detail
in a later chapter.
Modulation

Given two sequences x[n] and w[n], and

y[n] = x[n] w[n]

(2.31)

Then we say that y[n] is w[n] modulated by x[n]. This is analogous to carrier modulation
in communication systems.
Correlation The correlation, or more precisely cross-correlation, between two finite
length data sequences x[n] and w[n] is defined by
N
1 X
r=
x[n]w[n]
N n=0

(2.32)

if each sequence is of length N . The correlation coefficient r is often used as a measure

of how similar the two sequences are. If they are very different, then the value of r is low.
The matched filter used in digital communication receivers for optimal detection is also
effectively a correlator between the incoming and the template signals.

Discrete-time Systems

31

2.2 Discrete-time Systems

A discrete-time system is one that processes a discrete-time input sequence to produce a
discrete-time output sequence. There are many different kinds of such systems.

2.2.1 Classification of Systems

Discrete-time systems, like continuous-time systems, can be classified in a variety of
ways.
Linearity A linear system is one which obeys the superposition principle. For a certain system, let the outputs corresponding to inputs x1 [n] and x2 [n] are y1 [n] and y2 [n]
respectively. Now if the input is given by
x[n] = Ax1 [n] + Bx2 [n]

(2.33)

where A and B are arbitrary constants, then the system is linear if its corresponding
output is
y[n] = Ay1 [n] + By2 [n]
(2.34)
Superposition is a very nice property which makes analysis much simpler. Although
many real systems are not entirely linear throughout its operating region (for instance, the
bipolar transistor), they can be considered approximately linear for certain input ranges.
Linearization is a very useful approach to analyzing nonlinear systems. Almost all the
discrete-time systems considered in this book are linear systems.
Example 2.5. Are the down-sampler and up-sampler linear systems?
Consider the down-sampler
y[n] = x[nM ]
For input x1 [n], the corresponding output is y1 [n] = x1 [nM ]. For input x2 [n], the output
is y2 [n] = x2 [nM ]. Let x[n] be a linear combination of these two input with arbitrary
constants A and B so that
x[n] = Ax1 [n] + Bx2 [n]
The output is given by
y[n] = Ax1 [nM ] + Bx2 [nM ]
= Ay1 [n] + By2 [n]

32

Discrete-time Signals and Systems

Therefore the down-sampler is a linear system.

Now consider the up-sampler
y[n] =

x[n/L], n = 0, L, 2L, . . .
0,
otherwise

Let y1 [n] and y2 [n] be the outputs for inputs x1 [n] and x2 [n] respectively. For
x[n] = Ax1 [n] + Bx2 [n]
then the output is


Ax1 [n/L] + Bx2 [n/L], n = 0, L, 2L, . . .

0,
otherwise
= Ay1 [n] + By2 [n]

y[n] =

Hence the up-sampler is also linear.

Shift Invariance A shift (or time) invariant system is one that does not change with
time. Let a system response to an input x[n] be y[n]. If the input is now shifted by n 0 (an
integer) samples,
x1 [n] = x[n n0 ]
(2.35)
then the system is shift invariant if its response to x1 [n] is
y1 [n] = y[n n0 ]

(2.36)

In the remainder of this book, we shall use the terms linear time-invariant (LTI) and linear
shift-invariant interchangeably.
Example 2.6. A system has input-output relationship given by
y[n] =

n
X

k=

Is this system shift-invariant?

x[k]

Discrete-time Systems

33

If the input is now x1 [n] = x[n n0 ], then the corresponding output is

y1 [n] =
=

n
X

k=
n
X

k=

nn
X0

x1 [k]
x[k n0 ]
x[k]

k=

= y[n n0 ]
Therefore the system is shift invariant.
Example 2.7. Is the down-sampler a shift invariant system?
Let M (a positive integer) be the down-sampling ratio. So for an input x[n] the output is
y[n] = x[nM ]. Now if x1 [n] is x[n] delayed by n0 samples, then
x1 [n] = x[n n0 ]
and the corresponding output is
y1 [n] = x[nM n0 ]
h
n0  i
M
= x n
M

If the system is shift invariant, one would expect the output to be

y[n n0 ] = x [(n n0 ) M ]
Since this is not the case, the down-sampler must be shift variant.
Causality The response of a causal system at any time depends only on the input at the
current and past instants, not on any future samples. In other words, the output sample
y[n0 ] for any n0 only depends on x[n] for n n0 .
Example 2.8. Determine if the following system is causal:
y[n] =

k=

(n k) u[n k]x[k]

34

Discrete-time Signals and Systems

Note that u[n k] = 0 for n < k because the unit step sequence is zero for negative
indices. In other words, for a certain n, u[n k] = 0 for k > n. So the output can be
written as
n
X
(n k) x[k]
y[n] =
k=

So y[n] depends on x[k] for k n and therefore the system is causal.

Stability There are two common criteria for system stability. They are exponential
stability and bounded-input bounded-output (BIBO) stability. The first criterion is more
stringent. It requires the response of the system to decay exponentially fast for a finite
duration input. The second one merely requires that the output be a bounded sequence if
the input is a bounded sequence.
Example 2.9. Determine if the system with the following input-output relationship is
BIBO stable.

X
(n k) u[n k]x[k]
y[n] =
k=

Consider input x[n] = [n]. Then

y[n] =

k=

(n k) u[n k][k]

= nu[n]
which is unbounded as it grows linearly with n. Therefore the system is not BIBO stable.
Example 2.10. Determine if the following system is BIBO stable. Note that this system
is an averager, taking the average of the past M samples.
M 1
1 X
y[n] =
x[n k]
M k=0

Discrete-time Systems

35

Let the input |x[n]| B for some finite value B. Consider the magnitude of the output


1

1 M
X


|y[n]| = 
x[n k]
M

1
M

k=0
M
1
X
k=0

|x[n k]|

1
(M B)

M
= B
Hence the output is bounded and the system is BIBO stable.
Lossy or Lossless For a passive system that does not generate any energy internally,
the output should have at most the same energy as the input. So

n=

|y[n]|

n=

|x[n]|2 <

(2.37)

A lossless system is one which the equality holds.

2.2.2 Linear Shift-Invariant Systems

An discrete-time LTI system, like its continuous-time counterpart, is completely characterized by its impulse response. In other words, the impulse response tells us everything
we need to know about an LTI system as far as signal processing is concerned. The
impulse response is simply the observed system output when the input is an impulse
sequence. For continuous-time systems, the impulse function is purely a mathematical
entity. However, for discrete-time systems, since we are dealing with sequences of numbers, the impulse sequence can realistically (and easily) be generated.
2.2.2.1 Linear Convolution
Let us consider a discrete-time LTI system with impulse reponse h[n] as shown in Figure
2.10. What would be the output y[n] of the system if the input x[n] is as shown in Figure
2.11?

36

Discrete-time Signals and Systems

0.8

0.7

0.6

h[n]

0.5

0.4

0.3

0.2

0.1

4
Index (n)

Figure 2.10: Impulse Response of the System Considered in Section 2.2.2.1

2.5

Input Signal x[n]

1.5

0.5

4
Index (n)

Figure 2.11: Input Signal Sequence

Discrete-time Systems

37

Since the system is linear and time invariant, we can make use of the superposition principle to compute the output. The input sequence is composed of three impulses. Mathematically, it can be expressed as
x[n] = [n] + 0.5[n 1] + 2[n 2]

(2.38)

Let
x1 [n] = [n]
x2 [n] = 0.5[n 1]
x3 [n] = 2[n 2]
and the system response to each of these inputs are respectively y 1 [n], y2 [n] and y3 [n].
The sample values of y1 [n] are given by
y1 [0] = h[0]x1 [0]=0.8
y1 [1] = h[1]x1 [0]=0.4
y1 [2] = h[2]x1 [0]=0.2
y1 [3] = h[3]x1 [0]=0.1
which is the same as the impulse response since x1 [n] is a unit impulse. Similarly,
y2 [1] = h[0]x2 [1]=0.4
y2 [2] = h[1]x2 [1]=0.2
y2 [3] = h[2]x2 [1]=0.1
y2 [4] = h[3]x2 [1]=0.05
and
y3 [2] = h[0]x3 [2]=1.6
y3 [3] = h[1]x3 [2]=0.8
y3 [4] = h[2]x3 [2]=0.4
y3 [5] = h[3]x3 [2]=0.2
The system output y[n] in response to input x[n] is therefore, through the superposition
principle, given by
y[n] = y1 [n] + y2 [n] + y3 [n]
= {0.8, 0.8, 2, 1, 0.45, 0.2}

38

Discrete-time Signals and Systems

Note that
y[0]
y[1]
y[2]
y[3]
y[4]
y[5]

=
=
=
=
=
=

h[0]x[0]
h[1]x[0] + h[0]x[1]
h[2]x[0] + h[1]x[1] + h[0]x[2]
h[3]x[0] + h[2]x[1] + h[1]x[2]
h[3]x[1] + h[2]x[2]
h[3]x[2]

In general, we have

or

y[n] = h[n]x[0] + h[n 1]x[1] + . . . + h[1]x[n 1] + h[0]x[n]

y[n] =

n
X

h[k]x[n k]

(2.40)

n
X

x[n]h[n k]

(2.41)

k=0

Alternatively,
y[n] =

(2.39)

k=0

Equations 2.40 and 2.41 are the linear convolution equations for finite length sequences.
If the length of h[n] is M and the length of x[n] is N , then the length of y[n] is N +M 1.
We can further generalize it for infinite length sequences:
y[n] =
=

k=

k=

h[k]x[n k]

(2.42)

x[k]h[n k]

(2.43)

These equations are analogous to the linear convolution equation for continuous-time
signals. Note that the linear convolution equation comes about as a result of the superposition principles and therefore applies only to LTI systems.
The convolution equation for discrete-time signals is also called the convolution sum. It
is denoted by ?. So equations 2.42 and 2.43 can be written as
y[n] = x[n] ? h[n]
The convolution sum is one of the most important fundamental equations in DSP.

(2.44)

Discrete-time Systems

39

2.2.2.2 Properties of Linear Convolution

The convolution sum has three important properties:
1. Commutative
x[n] ? y[n] = y[n] ? x[n]

(2.45)

(x[n] ? w[n]) ? y[n] = x[n] ? (w[n] ? y[n])

(2.46)

x[n] ? (w[n] + y[n]) = x[n] ? w[n] + x[n] ? y[n]

(2.47)

2. Associative
3. Distributive

2.2.2.3 Condition for Stability

Since the impulse response completely characterize an LTI system, we should be able to
draw conclusions regarding the stability of a system based on its impulse response. We
shall consider BIBO stability here.
Theorem 2.1. A discrete-time LTI system is BIBO stable if its impulse response is absolutely summable.
Proof. Let the input be bounded, i.e. |x[n]| < B < for some finite value B. The
magnitude of the output is given by



X



|y[n]| = 
h[k]x[n k]



k=

|h[k]| |x[n k|

k=

k=

|h[k]|

P
So the magnitude of y[n] is bounded if
k= |h[k]| is finite. In other words, the impulse
response must be absolutely summable.

40

Discrete-time Signals and Systems

2.2.2.4 Condition for Causality

Theorem 2.2. A discrete-time LTI system is causal if and only if its impulse response is
a causal sequence.
Proof. Consider an LTI system with impulse response h[k]. Two different inputs x 1 [n]
and x2 [n] are the same up to a certain point in time, that is x1 [n] = x2 [n] for n n0 for
some n0 . The outputs y1 [n] and y2 [n] at n = n0 are given by
y1 [n0 ] =

k=

h[k]x1 [n0 k]

1
X

h[k]x1 [n0 k] +

h[k]x2 [n0 k]

k=

h[k]x1 [n0 k]

h[k]x2 [n0 k]

k=0

and
y2 [n0 ] =

k=

1
X

k=

h[k]x2 [n0 k] +

k=0

Since x1 [n] = x2 [n] for n n0 , if the system is causal, then the outputs y1 [n] and y2 [n]
must be the same for n n0 . More specifically, y1 [n0 ] = y2 [n0 ]. Now,

X
k=0

h[k]x1 [n0 k] =

X
k=0

h[k]x2 [n0 k]

because x1 [n0 k] = x2 [n0 k] for non-negative values of k. Since x1 [n] may not be
equal to x2 [n] for n > n0 , we must have
1
X

k=

h[k]x1 [n0 k] =

which means that h[k] = 0 for k < 0.

1
X

k=

h[k]x2 [n0 k] = 0

Discrete-time Systems

41

2.2.3 FIR and IIR Systems

In DSP, finite impulse response (FIR) and infinite impulse response (IIR) systems and
usually referred to as FIR and IIR filters.
An causal FIR filter has impulse response
h[n] = {h0 , h1 , h2 , , hM , 0, 0, }
where h0 , h1 , h2 , , hM are called filter coefficients. The length of the impulse response
is M + 1 and it is called an M -th order filter.
Example 2.11. What is the impulse response of a causal LTI system with the following
input-output relationship?
y[n] = 2x[n] + x[n 1] x[n 3]
Impulse response is the systems output when the input is a unit impulse signal. Therefore,
h[n] = 2[n] + [n 1] [n 3]
= {2, 1, 0, 1}
This filter is an FIR filter.
The filtering equation is essentially the linear convolution equation. Since the impulse
response has a length of M + 1, the output can be computed as (assuming a causal filter):
y[n] =

M
X
k=0

h[k]x[n k]

(2.48)

The impulse response of an IIR system, however, has infinite length. So the filtering
equation for a causal filter is given by
y[n] =

X
k=0

h[k]x[n k]

(2.49)

This equation is not computable in practice due to the infinite limit. But there is a type
of IIR systems that admits recursive computation of y[n] and so it can be computed. This

42

Discrete-time Signals and Systems

type of IIR systems are the only IIR filters that are used in DSP. The general input-output
relationship of these IIR filters is
y[n] =

N
X
k=1

a[k]y[n k] +

M
X
k=0

b[k]x[n k]

(2.50)

where {a[k]} and {b[k]} are the filter coefficients. The present output y[n] is a linear
combination of previous N outputs and the present and previous M outputs. The order
of the filter is M 1 or N , whichever is larger.
Example 2.12. A causal LTI system has impulse response
h[n] = ah[n 1] + [n]
where a is a constant. What is the input-output equation for this system?
Since y[n] = h[n] for x[n] = [n], it is obvious from the impulse response that the
input-output equation for this system is
y[n] = ay[n 1] + x[n]
Example 2.13. Find the impulse response of the causal system with the following inputoutput equation:
y[n] = 0.25y[n 2] + x[n]
The impulse response of this system is given by
h[n] = 0.25h[n 2] + [n]
and so
h[0] = 0.25h[2] + [0]=1
h[1] = 0.25h[1] + [1]=0
h[2] = 0.25h[0] + [2] =(0.25)2
h[3] = 0.25h[1] + [3] =0
..
..
.
.
Hence
h[n] =

(0.5)n , n 0, and even

0,
otherwise

Summary

43

2.3 Summary
In this chapter, we have studied the fundamentals of discrete-time signals and systems
in the time domain. Discrete-time signals are essentially a sequence of numbers. Some
fundamental sequences, such as the unit impulse, unit step and sinusoidal (both real and
complex) sequences are examined because more complex sequences can be expressed as
a combination of some or all of these fundamental ones.
The most important discrete-time systems are the linear time invariant (LTI) systems. For
these systems, the superposition principle applies which leads to the linear convolution
sum. This convolution sum is the way by which we derive the output of the system given
an input. An LTI system is completely characterized by its impulse response. All the
properties of such a system is revealed by its impulse response, including causality and
BIBO stability.