Module 1

Introduction

Signal, System and Signal Processing

A signal is defined as any physical quantity that varies with time, space, or any other
independent variable or variables. Mathematically, we describe a signal as a function of one or
more independent variables. For example, the functions
𝑠1 (𝑡) = 5𝑡
𝑠2 (𝑡) = 20𝑡 2
𝑠(𝑥, 𝑦) = 3𝑥 + 2𝑥𝑦 + 10𝑦 2
For example, a speech signal (see Fig. 1.1) cannot be described functionally by
expressions such as (1.1). In general, a segment of speech may be represented to a high degree
of accuracy as a sum of several sinusoids of different amplitudes and frequencies, that is, as
𝑁

∑ 𝐴𝑖 (𝑡) sin[2𝜋𝐹𝑖 (𝑡)𝑡 + 𝜃𝑖 (𝑡)]

𝑖=1

where {𝐴𝑖 (𝑡)}, {𝐹𝑖 (𝑡)}, 𝑎𝑛𝑑 {𝜃𝑖 (𝑡)} are the sets of (possibly time-varying) amplitudes,
frequencies, and phases, respectively, of the sinusoids.
Thus signal generation is usually associated with a system that responds to a stimulus
or force. In a speech signal, the system consists of the vocal cords and the vocal tract, also
called the vocal cavity. The stimulus in combination with the system is called a signal source.
A system may also be defined as a physical device that performs an operation on a
signal. For example, a filter used to reduce the noise and interference corrupting a desired
information-bearing signal is called a system.
If the operation is linear, the system is called linear. If the operation on the signal is
nonlinear, the system is said to be nonlinear, and so forth. Such operations are usually referred
to as signal processing.
In digital processing of signals on a digital computer, the operations performed on a
signal consist of a number of mathematical operations as specified by a software program. In
this case, the program represents an implementation of the system in software.

Dept. of ECE, CMRIT

Basic Elements of a Digital Signal Processing System
1. Most of the signals encountered in science and engineering are analog in nature.
2. These signals are functions of a continuous variable, such as time or space, and usually
take on values in a continuous range.
3. Such signals may be processed directly by appropriate analog systems (such as filters,
frequency analyzers, or frequency multipliers) for the purpose of changing their
characteristics or extracting some desired information.
4. In such a case we say that the signal has been processed directly in its analog form, as
illustrated in Fig. 1.1. Both the input signal and the output signal are in analog form.

Fig 1.1: Analog signal processing

Digital signal processing provides an alternative method for processing the analog signal, as
illustrated in Fig. 1.2.

Fig 1.2: Block diagram of a digital signal processing system

1. To perform the processing digitally, there is a need for an interface between the analog
signal and the digital processor. This interface is called an analog-to-digital (A/D)
converter. The output of the A/D converter is a digital signal that is appropriate as an
input to the digital processor.
2. The digital signal processor may be a large programmable digital computer or a small
microprocessor programmed to perform the desired operations on the input signal.
3. Programmable machines provide the flexibility to change the signal processing
operations through a change in the software, whereas hardwired machines are difficult
to reconfigure.
4. In applications where the digital output from the digital signal processor is to be given
to the user in analog form, such as in speech communications, we must provide another
interface from the digital domain to the analog domain. Such an interface is called a
digital-to-analog (D/A) converter.

Dept. of ECE, CMRIT

Advantages of Digital over Analog Signal Processing
1. A digital programmable system allows flexibility in reconfiguring the digital signal
processing operations simply by changing the program. Reconfiguration of an analog
system usually implies a redesign of the hardware followed by testing and verification
to see that it operates properly.
2. Tolerances in analog circuit components make it extremely difficult for the system
designer to control the accuracy of an analog signal processing system. On the other
hand, a digital system provides much better control of accuracy requirements in terms
of word length, floating-point versus fixed-point arithmetic, and similar factors.
3. Digital signals are easily stored on magnetic media (tape or disk) without deterioration
or loss of signal fidelity beyond that introduced in the A/D conversion.
4. It is usually very difficult to perform precise mathematical operations on signals in
analog form but these same operations can be routinely implemented on a digital
computer using software.
5. In some cases a digital implementation of the signal processing system is cheaper than
its analog counterpart. The lower cost may be due to the fact that the digital hardware
is cheaper, or perhaps it is a result of the flexibility for modifications provided by the
digital implementation.

Classification of Signals
The methods we use in processing a signal or in analyzing the response of a system to a signal
depend heavily on the characteristic attributes of the specific signal.
i. Multichannel and Multidimensional Signals
The value of the function (i.e., the dependent variable) can be a real-valued scalar
quantity, a complex-valued quantity, or perhaps a vector. For example, the signal
𝑠1 (𝑡) = 𝐴 sin(3𝜋𝑡) ------> real-valued signal

𝑠2 (𝑡) = 𝐴𝑒 𝑗(3𝜋𝑡) = 𝐴 cos(3𝜋𝑡) + 𝑗𝐴 sin(3𝜋𝑡) ------> complex-valued signal

If 𝑠𝑘 (𝑡), k = 1, 2, 3, denotes the electrical signal from the kth sensor as a function of
time, the set of p = 3 signals can be represented by a vector 𝑺𝟑 (𝑡) where
𝑠1 (𝑡)
𝑺𝟑 (𝑡) = [𝑠2 (𝑡)]
𝑠3 (𝑡)
We refer to such a vector of signals as a multichannel signal. In electrocardiography,
for example, 3-lead and 12-lead electrocardiograms (ECG) are often used in practice, which
result in 3-channel and 12-channel signals.
A signal is called M-dimensional if its value is a function of M independent variables.
The color TV picture is a three-channel, three-dimensional signal, which can be represented by
the vector

Dept. of ECE, CMRIT

 𝐼1 (𝑥, 𝑦, 𝑡)
𝑰(𝑥, 𝑦, 𝑡) = [𝐼2 (𝑥, 𝑦, 𝑡)]
𝐼2 (𝑥, 𝑦, 𝑡)

ii. Continuous-Time Versus Discrete-Time Signals

Continuous Time signals or Analog signals are defined for every value of time and they
take on values in the continuous interval (𝑎, 𝑏), where a can be −∞ and b can be ∞.
Ex: Speech waveform

Fig 1.3: Example of a speech signal.

𝑥1 (𝑡) = cos(𝜋𝑡) , 𝑥2 (𝑡) = 𝑒 −|𝑡| , −∞ < 𝑡 < ∞ are examples of analog signals.

Discrete-time signals are defined only at certain specific values of time. These time
instants need not be equidistant, but in practice they are usually taken at equally spaced intervals
for computational convenience and mathematical tractability.
The signal 𝑥(𝑡𝑛 ) =𝑒 −𝑡𝑛 , 𝑛 = 0, ±1, ±2, . .. is an example of a discrete-time signal. The
index 𝑛 of the discrete time instants as the independent variable, the signal value becomes a
function of an integer variable (i.e., a sequence of numbers). If the time instants 𝑡𝑛 are equally
spaced (i.e., 𝑡𝑛 = 𝑛𝑇 ), the notation 𝑥(𝑛𝑇 ) is also used. For example, the sequence
0.8𝑛 , 𝑖𝑓 𝑛 ≥ 0
𝑥(𝑛) = {
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒

Fig 1.4: Graphical representation of the discrete time signal 𝒙(𝒏) = 𝟎. 𝟖𝒏 for 𝒏 ≥ 𝟎 and
𝒙(𝒏) = 𝟎 for 𝒏 < 𝟎.

Dept. of ECE, CMRIT

 iii. Continuous-Valued Versus Discrete-Valued Signals
The values of a continuous-time or discrete-time signal can be continuous or
discrete. If a signal takes on all possible values on a finite or an infinite range, it is said
to be a continuous-valued signal.
Alternatively, if the signal takes on values from a finite set of possible values, it
is said to be a discrete-valued signal.
A discrete-time signal having a set of discrete values possible values is called a
digital signal.

Fig 1.5: Digital signal with four different amplitude values.

The process of converting a continuous-valued signal into a discrete-valued

signal, called quantization, is basically an approximation process. Thus the signal value
8.58 will be approximated by the value 8 if the quantization process is performed by
truncation or by 9 if the quantization process is performed by rounding to the nearest
integer.

iv. Deterministic Versus Random Signals

Any signal that can be uniquely described by an explicit mathematical
expression, a table of data, or a well-defined rule is called deterministic. This term is
used to emphasize the fact that all past, present, and future values of the signal are
known precisely, without any uncertainty.
In many practical applications, however, there are signals that either cannot be
described to any reasonable degree of accuracy by explicit mathematical formulas, or
such a description is too complicated to be of any practical use. The lack of such a
relationship implies that such signals evolve in time in an unpredictable manner. We
refer to these signals as random.

Dept. of ECE, CMRIT

The Concept of Frequency in Continuous-Time and Discrete-Time Signals
Frequency is closely related to a specific type of periodic motion called harmonic
oscillation, which is described by sinusoidal functions. The concept of frequency is directly
related to the concept of time. Actually, it has the dimension of inverse time.
Continuous-Time Sinusoidal Signals
A simple harmonic oscillation is mathematically described by the following continuous
time sinusoidal signal:
𝑥𝑎 (𝑡) = Acos(Ω𝑡 + 𝜃), −∞ < 𝑡 < ∞ (1)
The subscript a used with x(t) denotes an analog signal. This signal is completely
characterized by three parameters: A is the amplitude of the sinusoid, Ω is the frequency in
radians per second (rad/s), and θ is the phase in radians.
Ω = 2𝜋𝐹
𝑥𝑎 (𝑡) = Acos(2𝜋𝐹𝑡 + 𝜃), −∞ < 𝑡 < ∞

Fig 1.5: an analog sinusoidal signal

Notes:
1. For every fixed value of the frequency F, 𝑥𝑎 (𝑡) is periodic.
𝑥𝑎 (𝑡 + 𝑇𝑝 ) = 𝑥𝑎 (𝑡) for all 𝑡
1
where 𝑇𝑝 = 𝐹 is the fundamental period of the sinusoidal signal
2. Continuous-time sinusoidal signals with distinct (different) frequencies are themselves
distinct.
3. Increasing the frequency F results in an increase in the rate of oscillation of the signal,
in the sense that more periods are included in a given time interval.
Note that a sinusoidal signal can be obtained by adding two equal-amplitude complex-
conjugate exponential signals, sometimes called phase. As time progresses the phasors rotate
in opposite directions with angular frequencies ±𝛺 radians per second. Since a positive
frequency corresponds to counterclockwise uniform angular motion, a negative frequency
simply corresponds to clockwise angular motion.

Dept. of ECE, CMRIT

 𝐴 𝑗 𝐴
𝑥𝑎 (𝑡) = Acos(Ω𝑡 + 𝜃) = 𝑒 (Ω𝑡 + 𝜃) + 𝑒 −𝑗 (Ω𝑡 + 𝜃)
2 2

Fig 1.6: Representation of a cosine function by a pair of complex-conjugate exponentials

(phasors)

Discrete-Time Sinusoidal Signals

A discrete-time sinusoidal signal may be expressed as
𝑥(𝑛) = Acos(𝜔𝑛 + 𝜃) −∞ < 𝑛 < ∞ (1)
where n is an integer variable, called the sample number, A is the amplitude of the sinusoid, ω
is the frequency in radians per sample, and θ is the phase in radians
𝜔 = 2𝜋𝑓
𝑥(𝑛) = Acos(2𝜋𝑓𝑛 + 𝜃) −∞ < 𝑛 < ∞
The frequency f has dimensions of cycles per sample. Figure 1.7 shows a sinusoid with
𝜋 1
frequency 𝜔 = 6 radian per sample (𝑓 = 12 cycles per sample) and phase θ = π/3

Fig 1.7: Example of a discrete-time sinusoidal signal (ω = π/6 and θ = π/3)

Dept. of ECE, CMRIT

In contrast to continuous-time sinusoids, the discrete-time sinusoids are characterized by the
following properties
• A discrete-time sinusoid is periodic only if its frequency f is a rational number.
By definition, a discrete-time signal x(n) is periodic with period N (N > 0) if and
only if
𝑥(𝑛 + 𝑁) = 𝑥(𝑛) for all 𝑛 (1)
The smallest value of N for which (1) is true is called the fundamental period.
For a sinusoid with frequency 𝑓0 to be periodic, we should have
cos [2𝜋𝑓0 (𝑛 + 𝑁) + 𝜃] = Acos(2𝜋𝑓0 𝑛 + 𝜃)
This relation is true if and only if there exists an integer k such that
2𝜋𝑓0 𝑁 = 2𝜋𝑘
𝑘
𝑓0 =
𝑁
According to (2), a discrete-time sinusoidal signal is periodic only if its frequency 𝑓0 can
be expressed as the ratio of two integers (i.e., 𝑓0 is rational).

Dept. of ECE, CMRIT

 Discrete-Time Signals and Systems
A discrete time signal 𝑥(𝑛) is function of an independent variable that is an integer. ls
It is graphically represented as in Fig. 1.1. It is important to note that a discrete-time signal is
not defined at instants between two successive samples . Also, it is incorrect to think that 𝑥(𝑛)
is equal to zero if 𝑛 is not an integer. Simply, the signal is not defined for non integer values of
𝑛.
𝑥(𝑛) was obtained from sampling an analog signal xa(t), then 𝑥(𝑛) ≡ 𝑥𝑎 (𝑛𝑇), where
𝑇 is the sampling period (i.e., the time between successive samples). Besides the graphical
representation of a discrete-time signal or sequence as illustrated in Fig. 1.1, there are some
alternative representations that are often more convenient to use. These are:
1. Functional representation, such as
1, 𝑓𝑜𝑟 𝑛 = 1,3
𝑥(𝑛) = { 4, 𝑓𝑜𝑟 𝑛 = 2
0, 𝑒𝑙𝑠𝑒𝑤ℎ𝑒𝑟𝑒

2. Tabular representation, such as

𝑛 |. . . ______ − 2___ − 1 0_ 1 ____2_____3_____4______5_
𝑥(𝑛) |. . . _________0____ _0____0____1_____4____ 1_____0______0_

3. Sequence representation
An infinite-duration signal or sequence with the time origin (n = 0) indicated by the
symbol ↑ is represented as
𝑥(𝑛) = {. . . 0, 0, 1, 4, 1, 0, 0,. . . }
↑
A sequence 𝑥(𝑛), which is zero for 𝑛 < 0, can be represented as
𝑥(𝑛) = { 0, 1, 4, 1, 0, 0,. . . }
↑

Figure 1.1 Graphical representation of a discrete-time signal

Dept. of ECE, CMRIT

Classification of Discrete-Time Signals

Symmetric (even) and antisymmetric (odd) signals.

A real-valued signal 𝑥(𝑛) is called symmetric (even) if
𝑥(−𝑛) = 𝑥(𝑛)
On the other hand, a signal 𝑥(𝑛) is called antisymmetric (odd) if
𝑥(−𝑛) = −𝑥(𝑛)
We note that if x(n) is odd, then 𝑥(0) = 0.

Fig: Example of even (a) and odd (b) signals

The even signal component 𝑥𝑒 (𝑛) is formed by adding 𝑥(𝑛) to 𝑥(−𝑛) and dividing by 2
1
𝑥𝑒 (𝑛) = [𝑥(𝑛) + 𝑥(−𝑛)]
2
Similarly, we form an odd signal component 𝑥𝑜 (𝑛) according to the relation
1
𝑥𝑜 (𝑛) = [𝑥(𝑛) − 𝑥(−𝑛)]
2
Thus any arbitrary signal 𝑥(𝑛) can be expressed as
𝑥(𝑛) = 𝑥𝑒 (𝑛) + 𝑥0 (𝑛)

Dept. of ECE, CMRIT

Ex: Show that any signal can be decomposed into an even and an odd component. Is the
decomposition unique? Illustrate your arguments using the signal
𝑥(𝑛) = {2, 3, 4 , 5, 6}
↑
Sol:
𝑥(−𝑛) = {6, 5, 4 , 3, 2}
↑
Even Component,
1
𝑥𝑒 (𝑛) = [𝑥(𝑛) + 𝑥(−𝑛)] = {4,4,4,4,4}
2
↑
Odd Component,
1
𝑥𝑜 (𝑛) = [𝑥(𝑛) − 𝑥(−𝑛)] = {−2, −1,0,1,2}
2
↑
We can verify that 𝑥𝑒 (𝑛) = 𝑥𝑒 (−𝑛) and 𝑥𝑜 (𝑛) = −𝑥𝑜 (−𝑛)

Discrete-Time Systems
A discrete-time system is a device or algorithm that operates on a discrete-time signal, called
the input or excitation, according to some well-defined rule, to produce another discrete-time
signal called the output or response of the system.

Fig: Block diagram representation of a discrete-time system.

We express the general relationship between 𝑥(𝑛) and 𝑦(𝑛) as

𝑦(𝑛) = 𝒯[𝑥(𝑛)]
where the symbol 𝒯 denote the transformation (also called an operator) or processing
performed by the system on 𝑥(𝑛) to produce 𝑦(𝑛).

Ex 1: Determine the response of the following systems to the input signal

|𝑛|, −3 ≤ 𝑛 ≤ 3
𝑥(𝑛) = {
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒

Dept. of ECE, CMRIT

 a) 𝑦(𝑛) = 𝑥(𝑛) (identity system)
b) 𝑦(𝑛) = 𝑥(𝑛 − 1) (unit delay system)
c) 𝑦(𝑛) = 𝑥(𝑛 + 1) (unit advance system)
1
d) 𝑦(𝑛) = 3 [𝑥(𝑛 + 1) + 𝑥(𝑛) + 𝑥(𝑛 − 1)] (moving average filter)
e) 𝑦(𝑛) = ∑𝑛𝑘=−∞ 𝑥(𝑘) = 𝑥(𝑛) + 𝑥(𝑛 − 1) + ⋯ (accumulator)
Sol: First, we determine explicitly the sample values of the input signal
𝑥(𝑛) = {. . . 0, 3, 2, 1, 0, 1, 2, 3, 0. . . }
↑
Next, we determine the output of each system using its input–output relationship.
a) In this case the output is exactly the same as the input signal. Such a system is known
as the identity system.
b) This system simply delays the input by one sample. Thus its output is given by
𝑥(𝑛) = {. . . 0, 3, 2, 1, 0, 1, 2, 3, 0. . . }
↑
c) In this case the system “advances” the input one sample into the future. For example,
the value of the output at time 𝑛 = 0 is 𝑦(0) = 𝑥(1). The response of this system to
the given input is
𝑥(𝑛) = {. . . 0, 3, 2, 1, 0, 1, 2, 3, 0. . . }
↑
d) The output of this system at any time is the mean value of the present, the immediate
past, and the immediate future samples. For example, the output at time 𝑛 = 0 is
1 2
𝑦(0) = [𝑥(1) + 𝑥(0) + 𝑥(−1)] =
3 3
Repeating this computation for every value of n, we obtain the output signal
5 2 5
𝑦(𝑛) = {. . .0, 1, , 2, 1, , 1,2, , 1,0, . . . }
3 3 3
↑
e) This system is basically an accumulator that computes the running sum of all the past
input values up to present time. The response of this system to the given input is
𝑦(𝑛) = {. . .0, 3, 5, 6, 6, 7, 9,12,0, . . . }
↑
By a simple algebraic manipulation the input–output relation of the accumulator can be written
as
𝑛 𝑛−1

𝑦(𝑛) = ∑ 𝑥(𝑘) = ∑ 𝑥(𝑘) + 𝑥(𝑛)

𝑘=−∞ 𝑘=−∞

𝑦(𝑛) = 𝑦(𝑛 − 1) + 𝑥(𝑛)

which justifies the term accumulator

Ex2: The accumulator described by (2.30) is excited by the sequence 𝑥(𝑛) = 𝑛𝑢(𝑛).
Determine its output under the condition that:

Dept. of ECE, CMRIT

 a) It is initially relaxed [i.e., 𝑦(−1) = 0].
b) Initially, 𝑦(−1) = 1.
Sol: The output of the system is defined as
𝑛 −1 𝑛

𝑦(𝑛) = ∑ 𝑥(𝑘) = ∑ 𝑥(𝑘) + ∑ 𝑥(𝑘)

𝑘=−∞ 𝑘=−∞ 𝑘=0
𝑛

= 𝑦(−1) + ∑ 𝑥(𝑘)
𝑘=0

𝑛(𝑛 + 1)
= 𝑦(−1) +
2
a) It is initially relaxed [i.e., 𝑦(−1) = 0] and hence
𝑛(𝑛 + 1)
𝑦(𝑛) = , 𝑛≥0
2
b) On the other hand, if the initial condition is 𝑦(−1) = 1, then
𝑛(𝑛 + 1) 𝑛2 + 𝑛 + 2
𝑦(𝑛) = 1 + = , 𝑛≥0
2 2

Block Diagram Representation of Discrete-Time Systems

• An adder

• A constant multiplier

• A signal multiplier

• A unit delay element.

𝑦(𝑛) = 𝑥(𝑛 − 1)

Dept. of ECE, CMRIT

 • A unit advance element.
𝑦(𝑛) = 𝑥(𝑛 + 1)

Ex: Using basic building blocks introduced above, sketch the block diagram representation of
the discrete-time system described by the input–output relation
1 1 1
𝑦(𝑛) = 𝑦(𝑛 − 1) + 𝑥(𝑛) + 𝑥(𝑛 − 1)
4 2 2
where x(n) is the input and y(n) is the output of the system.
Sol:
1 1
𝑦(𝑛) = 𝑦(𝑛 − 1) + [𝑥(𝑛) + 𝑥(𝑛 − 1)
4 2

Fig: Block diagram realizations of the system

Classification of Discrete-Time Systems

i. Static versus dynamic systems. A discrete-time system is called static or memoryless
if its output at any instant n depends at most on the input sample at the same time, but
not on past or future samples of the input.
If the output of a system at time n is completely determined by the input samples
in the interval from 𝑛 − 𝑁 to 𝑛(𝑁 ≥ 0), the system is said to have memory of duration
𝑁.
If 𝑁 = 0, the system is static. If 0 < 𝑁 < ∞, the system is said to have finite
memory, whereas if 𝑁 = ∞, the system is said to have infinite memory.

Dept. of ECE, CMRIT

 Examples of static or memoryless systems
i. 𝑦(𝑛) = 𝑎𝑥(𝑛)
ii. 𝑦(𝑛) = 𝑛𝑥(𝑛) + 𝑏𝑥 3 (𝑛)
Examples of dynamic systems or systems with memory
i. 𝑦(𝑛) = 𝑥(𝑛) + 3𝑥(𝑛 − 1)
ii. 𝑦(𝑛) = ∑𝑛𝑘=0 𝑥(𝑛 − 𝑘)
iii. 𝑦(𝑛) = ∑∞𝑘=0 𝑥(𝑛 − 𝑘)

ii. Time-invariant versus time-variant systems.

A relaxed system 𝒯 is time invariant or shift invariant if and only if
𝒯
𝑥(𝑛) → 𝑦(𝑛)
implies that
𝒯
𝑥(𝑛 − 𝑘) → 𝑦(𝑛 − 𝑘)
for every input signal 𝑥(𝑛) and every time shift 𝑘.
Next we delay the input sequence by some amount k and recompute the output. In general, we
can write the output as
𝑦(𝑛, 𝑘) = 𝒯[𝑥(𝑛 − 𝑘)]
Now if this output 𝑦(𝑛, 𝑘) = 𝑦(𝑛 − 𝑘), for all possible values of k, the system is time
invariant.
If the output 𝑦(𝑛, 𝑘) ≠ 𝑦(𝑛 − 𝑘), even for one value of k, the system is time variant.

Ex: Determine if the systems shown in Fig are time invariant or time variant.

Dept. of ECE, CMRIT

Sol: a) This system is described by the input–output equations
𝑦(𝑛) = 𝒯[𝑥(𝑛)] = 𝑥(𝑛) − 𝑥(𝑛 − 1) (1)
If the input is delayed by k units in time and applied to the system, then the output is given by
𝑦(𝑛, 𝑘) = 𝒯[𝑥(𝑛 − 𝑘)] = 𝑥(𝑛 − 𝑘) − 𝑥(𝑛 − 𝑘 − 1) (2)
If we delay 𝑦(𝑛) by 𝑘 units in time. Substitute 𝑛 by 𝑛 − 𝑘, we get in (1)
𝑦(𝑛 − 𝑘) = 𝑥(𝑛 − 𝑘) − 𝑥(𝑛 − 𝑘 − 1) (3)
From (1) and (3), 𝒚(𝒏, 𝒌) = 𝒚(𝒏 − 𝒌)
Therefore, the system is time invariant

b) This system is described by the input–output equations

𝑦(𝑛) = 𝒯[𝑥(𝑛)] = 𝑛𝑥(𝑛) (1)
If the input is delayed by k units in time and applied to the system, then the output is given by
𝑦(𝑛, 𝑘) = 𝒯[𝑥(𝑛 − 𝑘)] = 𝑛𝑥(𝑛 − 𝑘) (2)
If we delay 𝑦(𝑛) by 𝑘 units in time. Substitute 𝑛 by 𝑛 − 𝑘, we get in (1)
𝑦(𝑛 − 𝑘) = (𝑛 − 𝑘)𝑥(𝑛 − 𝑘)
= 𝑛𝑥(𝑛 − 𝑘) − 𝑘𝑥(𝑛 − 𝑘) (3)
From (1) and (3), 𝒚(𝒏, 𝒌) ≠ 𝒚(𝒏 − 𝒌)
Therefore, the system is time variant

c) This system is described by the input–output equations

𝑦(𝑛) = 𝒯[𝑥(𝑛)] = 𝑥(−𝑛) (1)
If the input is delayed by k units in time and applied to the system, then the output is given by
𝑦(𝑛, 𝑘) = 𝒯[𝑥(𝑛 − 𝑘)] = 𝑥(−𝑛 − 𝑘) (2)
If we delay 𝑦(𝑛) by 𝑘 units in time. Substitute 𝑛 by 𝑛 − 𝑘, we get in (1)

Dept. of ECE, CMRIT

 𝑦(𝑛 − 𝑘) = 𝑥(−(𝑛 − 𝑘))
= 𝑥(−𝑛 + 𝑘) (3)
From (1) and (3), 𝒚(𝒏, 𝒌) ≠ 𝒚(𝒏 − 𝒌)
Therefore, the system is time variant
d) This system is described by the input–output equations
𝑦(𝑛) = 𝒯[𝑥(𝑛)] = 𝑥(𝑛) cos(𝜔0 𝑛) (1)
If the input is delayed by k units in time and applied to the system, then the output is given by
𝑦(𝑛, 𝑘) = 𝒯[𝑥(𝑛 − 𝑘)] = 𝑥(𝑛 − 𝑘) cos(𝜔0 𝑛) (2)
If we delay 𝑦(𝑛) by 𝑘 units in time. Substitute 𝑛 by 𝑛 − 𝑘, we get in (1)
𝑦(𝑛 − 𝑘) = 𝑥(𝑛 − 𝑘) cos(𝜔0 (𝑛 − 𝑘))
= 𝑥(𝑛 − 𝑘) cos(𝜔0 𝑛 − 𝜔0 𝑘) (3)
From (1) and (3), 𝒚(𝒏, 𝒌) ≠ 𝒚(𝒏 − 𝒌)
Therefore, the system is time variant

iii. Linear versus nonlinear systems

A system is linear if and only if
𝒯[𝑎1 𝑥1 (𝑛) + 𝑎2 𝑥2 (𝑛)] = 𝑎1 𝒯[𝑥1 (𝑛)] + 𝑎2 𝒯[𝑥2 (𝑛)]
for any arbitrary input sequences 𝑥1 (𝑛) and 𝑥2 (𝑛), and any arbitrary constants 𝑎1
and 𝑎2

Fig: Graphical representation of the superposition principle. 𝓣 is linear if and only if

𝒚(𝒏) = 𝒚′ (𝒏).

If a relaxed system does not satisfy the superposition principle as given by the definition above,
it is called nonlinear

Dept. of ECE, CMRIT

Ex: Determine if the systems described by the following input–output equations are linear or
nonlinear.
a) 𝑦(𝑛) = 𝑛𝑥(𝑛)
b) 𝑦(𝑛) = 𝑥(𝑛2 )
c) 𝑦(𝑛) = 𝑥 2 (𝑛)
d) 𝑦(𝑛) = 𝐴𝑥(𝑛) + 𝐵
e) 𝑦(𝑛) = 𝑒 𝑥(𝑛)
Sol: (a) For two input sequences 𝑥1 (𝑛) and 𝑥2 (𝑛), the corresponding outputs are
𝑦1 (𝑛) = 𝑛𝑥1 (𝑛) (1)

𝑦2 (𝑛) = 𝑛𝑥2 (𝑛) (2)

A linear combination of the two input sequences results in the output
𝑦3 (𝑛) = 𝒯[𝑎1 𝑥1 (𝑛) + 𝑎2 𝑥2 (𝑛)] = 𝑛[𝑎1 𝑥1 (𝑛) + 𝑎2 𝑥2 (𝑛)]
= 𝑎1 𝑛𝑥1 (𝑛) + 𝑎2 𝑛𝑥2 (𝑛) (3)
On the other hand, a linear combination of the two outputs in (1) and (2) results in the
output
𝑎1 𝑦1 (𝑛) + 𝑎2 𝑦2 (𝑛) = 𝑎1 𝑛𝑥1 (𝑛) + 𝑎2 𝑛𝑥2 (𝑛) (4)
Since the right-hand sides of (3) and (4) are identical, the system is linear.
(b) For two input sequences 𝑥1 (𝑛) and 𝑥2 (𝑛), the corresponding outputs are
𝑦1 (𝑛) = 𝑥1 (𝑛2 ) (1)

𝑦2 (𝑛) = 𝑥2 (𝑛2 ) (2)

A linear combination of the two input sequences results in the output
𝑦3 (𝑛) = 𝒯[𝑎1 𝑥1 (𝑛) + 𝑎2 𝑥2 (𝑛)] = 𝑎1 𝑥1 (𝑛2 ) + 𝑎2 𝑥2 (𝑛2 ) (3)

On the other hand, a linear combination of the two outputs in (1) and (2) results in the
output
𝑎1 𝑦1 (𝑛) + 𝑎2 𝑦2 (𝑛) = 𝑎1 𝑥1 (𝑛2 ) + 𝑎2 𝑥2 (𝑛2 ) (4)
Since the right-hand sides of (3) and (4) are identical, the system is linear.

(c) For two input sequences 𝑥1 (𝑛) and 𝑥2 (𝑛), the corresponding outputs are
𝑦1 (𝑛) = 𝑥1 2 (𝑛) (1)

𝑦2 (𝑛) = 𝑥2 2 (𝑛) (2)

A linear combination of the two input sequences results in the output
𝑦3 (𝑛) = 𝒯[𝑎1 𝑥1 (𝑛) + 𝑎2 𝑥2 (𝑛)] = [𝑎1 𝑥1 (𝑛) + 𝑎2 𝑥2 (𝑛)]2 (3)
= 𝑎1 2 𝑥1 2 (𝑛) + 𝑎2 2 𝑥2 2 (𝑛) + 2𝑎1 𝑎2 𝑥1 (𝑛)𝑥2 (𝑛)

On the other hand, a linear combination of the two outputs in (1) and (2) results in the
output
𝑎1 𝑦1 (𝑛) + 𝑎2 𝑦2 (𝑛) = 𝑎1 𝑥1 2 (𝑛) + 𝑎2 𝑥2 2 (𝑛) (4)
Since the right-hand sides of (3) and (4) are not identical, the system is non linear.

Dept. of ECE, CMRIT

 (d) For two input sequences 𝑥1 (𝑛) and 𝑥2 (𝑛), the corresponding outputs are
𝑦1 (𝑛) = 𝐴𝑥1 (𝑛) + 𝐵 (1)

𝑦2 (𝑛) = 𝐴𝑥2 (𝑛) + 𝐵 (2)

A linear combination of the two input sequences results in the output
𝑦3 (𝑛) = 𝒯[𝑎1 𝑥1 (𝑛) + 𝑎2 𝑥2 (𝑛)] = 𝐴[𝑎1 𝑥1 (𝑛) + 𝑎2 𝑥2 (𝑛)] + 𝐵
= 𝑎1 𝐴𝑥1 (𝑛) + 𝑎2 𝐴𝑥2 (𝑛) + 𝐵 (3)
On the other hand, a linear combination of the two outputs in (1) and (2) results in the output
𝑎1 𝑦1 (𝑛) + 𝑎2 𝑦2 (𝑛) = 𝑎1 𝐴𝑥1 (𝑛) + 𝐵 + 𝑎2 𝐴𝑥2 (𝑛) + 𝐵 (4)
𝑦3 (𝑛) ≠ 𝑎1 𝑦1 (𝑛) + 𝑎2 𝑦2 (𝑛)
the system is non linear.
(e) For two input sequences 𝑥1 (𝑛) and 𝑥2 (𝑛), the corresponding outputs are

𝑦1 (𝑛) = 𝑒 𝑥1(𝑛) (1)

𝑦2 (𝑛) = 𝑒 𝑥2(𝑛) (2)

A linear combination of the two input sequences results in the output

𝑦3 (𝑛) = 𝒯[𝑎1 𝑥1 (𝑛) + 𝑎2 𝑥2 (𝑛)] = 𝑒 [𝑎1 𝑥1(𝑛)+𝑎2𝑥2(𝑛)]

= 𝑒 𝑎1 𝑥1(𝑛) + 𝑒 𝑎2 𝑥2(𝑛) (3)

On the other hand, a linear combination of the two outputs in (1) and (2) results in the output

𝑎1 𝑦1 (𝑛) + 𝑎2 𝑦2 (𝑛) = 𝑎1 𝑒 𝑥1(𝑛) + 𝑎2 𝑒 𝑥2 (𝑛) (4)

𝑦3 (𝑛) ≠ 𝑎1 𝑦1 (𝑛) + 𝑎2 𝑦2 (𝑛)
the system is non linear.

iv. Causal versus noncausal systems

A system is said to be causal if the output of the system at any time 𝑛 [i.e., 𝑦(𝑛)]
depends only on present and past inputs [i.e., 𝑥(𝑛), 𝑥(𝑛 − 1), 𝑥(𝑛 − 2), . . .], but does not
depend on future inputs [i.e., 𝑥(𝑛 + 1), 𝑥(𝑛 + 2), . . .]. In mathematical terms, the output
of a causal system satisfies an equation of the form
𝑦(𝑛) = 𝐹[𝑥(𝑛), 𝑥(𝑛 − 1), 𝑥(𝑛 − 2), . . . ]
where 𝐹[·] is some arbitrary function.
If a system does not satisfy this definition, it is called noncausal. Such a system
has an output that depends not only on present and past inputs but also on future inputs.

Dept. of ECE, CMRIT

Ex: Determine if the systems described by the following input–output equations are causal or
noncausal.
(a) 𝒚(𝒏) = 𝒙(𝒏) − 𝒙(𝒏 − 𝟏) -----> output of the system at any time 𝑛 [i.e., 𝑦(𝑛)] depends
only on present {𝑥(𝑛)} and past input {𝑥(𝑛 − 1)}
Hence the given system is CAUSAL
(b) 𝒚(𝒏) = ∑𝒏𝒌=−∞ 𝒙(𝒌)-----> output of the system at any time 𝑛 [i.e., 𝑦(𝑛)] depends only on
present {𝑥(𝑛)} and past input {𝑥(𝑛 − 1), 𝑥(𝑛 − 1), 𝑥(𝑛 − 2),. . . . . }
Hence the given system is CAUSAL
(c) 𝒚(𝒏) = 𝒂𝒙(𝒏) -----> output of the system at any time 𝑛 [i.e., 𝑦(𝑛)] depends only on
present {𝑥(𝑛)}
Hence the given system is CAUSAL
(d) 𝑦(𝒏) = 𝒙(𝒏) + 𝟑𝒙(𝒏 + 𝟒) -----> output of the system at any time 𝑛 [i.e., 𝑦(𝑛)] depends
only on present {𝑥(𝑛)} and the future input {𝑥(𝑛 + 4)}
Hence the given system is ANTI-CAUSAL
(e) 𝒚(𝒏) = 𝒙(𝒏𝟐 ) -----> output of the system at any time 𝑛 [i.e., 𝑦(𝑛)] depends only on present
{𝑥(𝑛)} , past inputs and also on the future input.
At time index, 𝑛 = 1, 𝑦(1) = 𝑥(1) (present input)
At time index, 𝑛 = 2, 𝑦(2) = 𝑥(4) (future input)
Hence the given system is ANTI-CAUSAL
(f) 𝒚(𝒏) = 𝒙(𝟐𝒏) -----> output of the system at any time 𝑛 [i.e., 𝑦(𝑛)] depends only on present
{𝑥(𝑛)} , past inputs and also on the future input.
At time index, 𝑛 = 1, 𝑦(1) = 𝑥(1) (present input)
At time index, 𝑛 = 2, 𝑦(2) = 𝑥(4) (future input)
Hence the given system is ANTI-CAUSAL
(g) 𝒚(𝒏) = 𝒙(−𝒏) -----> output of the system at any time 𝑛 [i.e., 𝑦(𝑛)] depends only on
present {𝑥(𝑛)} , past inputs and also on the future input.
At time index, 𝑛 = 1, 𝑦(1) = 𝑥(−1) (past input)
At time index, 𝑛 = −1, 𝑦(−1) = 𝑥(1) (future input)
Hence the given system is ANTI-CAUSAL

Dept. of ECE, CMRIT

 iv. Stable versus unstable systems.
An arbitrary relaxed system is said to be bounded input–bounded output (BIBO)
stable if and only if every bounded input produces a bounded output.
The condition that the input sequence 𝑥(𝑛) and the output sequence 𝑦(𝑛) are
bounded is translated mathematically to mean that there exist some finite numbers, say
𝑀𝑥 and 𝑀𝑦 , such that

|𝑥(𝑛)| ≤ 𝑀𝑥 < ∞, |𝑦(𝑛)| ≤ 𝑀𝑦 < ∞,

for all 𝑛. If, for some bounded input sequence 𝑥(𝑛), the output is unbounded (infinite),
the system is classified as unstable.

Ex: Consider the nonlinear system described by the input–output equation

𝑦(𝑛) = 𝑦 2 (𝑛 − 1) + 𝑥(𝑛)
Check if the given DT system is stable or not?
Sol: As an input sequence we select the bounded signal
𝑥(𝑛) = 𝐶𝛿(𝑛) where 𝐶 is a constant
Assume 𝑦(−1) = 0, the output sequence is
𝑦(0) = 𝑦 2 (−1) + 𝑥(0) = 0 + 𝐶 = 𝐶
𝑦(1) = 𝑦 2 (0) + 𝑥(1) = 𝐶 2 + 0 = 𝐶 2
𝑦(2) = 𝑦 2 (1) + 𝑥(2) = 𝐶 4 + 0 = 𝐶 4
.
.
𝑦(𝑛) = 𝑦 2 (𝑛 − 1) + 𝑥(𝑛) = 𝐶 2𝑛
Clearly, the output is unbounded when 1 < |𝐶| < ∞
Therefore, the system is BIBO unstable, since a bounded input sequence has resulted in an
unbounded output.

Dept. of ECE, CMRIT

Resolution of a Discrete-Time Signal into Impulses
Suppose we have an arbitrary signal x(n) that we wish to resolve into a sum of unit sample
sequences. Select the elementary signals 𝑥𝑘 (𝑛)
𝑥𝑘 (𝑛) = 𝛿(𝑛 − 𝑘)
where 𝑘 represents the delay of the unit sample sequence.

Fig: Multiplication of a signal 𝒙(𝒏) with a shifted unit sample sequence

In other words, each multiplication of the signal 𝑥(𝑛) by a unit impulse at some delay 𝑘,
[i.e., 𝛿(𝑛 − 𝑘)], in essence picks out the single value 𝑥(𝑘) of the signal 𝑥(𝑛) at the delay
where the unit impulse is nonzero.
𝑥(𝑛)𝛿(𝑛 − 𝑘) = 𝑥(𝑘)𝛿(𝑛 − 𝑘)
Consequently, if we repeat this multiplication over all possible delays, −∞ < 𝑘 < ∞, and sum
all the product sequences, the result will be a sequence equal to the sequence 𝑥(𝑛), that is,
∞

𝑥(𝑛) = ∑ 𝑥(𝑘) 𝑥(𝑘)𝛿(𝑛 − 𝑘)

𝑘=−∞

Thus the right-hand side of (3.10) gives the resolution or decomposition of any arbitrary signal
x(n) into a weighted (scaled) sum of shifted unit sample sequences.

Dept. of ECE, CMRIT

Ex: Consider the special case of a finite-duration sequence given as
𝑥(𝑛) = {2, 4, 0, 3}
↑
Resolve the sequence 𝑥(𝑛) into a sum of weighted impulse sequences

Sol: Since the sequence 𝑥(𝑛) is nonzero for the time instants 𝑛 = −1, 0, 2, we need
three impulses at delays 𝑘 = −1, 0, Following we find that
𝑥(𝑛) = 2𝛿(𝑛 + 1) + 4𝛿(𝑛) + 3𝛿(𝑛 − 1)

Response of LTI Systems to Arbitrary Inputs: The Convolution Sum

First, we denote the response 𝑦(𝑛, 𝑘) of the system to the input unit sample sequence at
𝑛 = 𝑘 by the special symbol ℎ(𝑛, 𝑘), −∞ < 𝑘 < ∞.
𝑦(𝑛, 𝑘) ≡ ℎ(𝑛, 𝑘) = 𝒯 [𝛿(𝑛 − 𝑘)]
If the impulse at the input is scaled by an amount 𝑐𝑘 ≡ 𝑥(𝑘), the response of the system
is the correspondingly scaled output
𝑐𝑘 ℎ(𝑛, 𝑘) = 𝑥(𝑘)ℎ(𝑛, 𝑘)
Finally, if the input is the arbitrary signal 𝑥(𝑛) that is expressed as a sum of weighted
impulses, that is,
∞

𝑥(𝑛) = ∑ 𝑥(𝑘)𝛿(𝑛 − 𝑘)
𝑘=−∞

then the response of the system to x(n) is the corresponding sum of weighted outputs, that is
∞

𝑦(𝑛) = 𝒯[𝑥(𝑛)] = 𝒯 [ ∑ 𝑥(𝑘)𝛿(𝑛 − 𝑘)]

𝑘=−∞
∞

= ∑ 𝑥(𝑘)𝒯[𝛿(𝑛 − 𝑘)]
𝑘=−∞
∞

= ∑ 𝑥(𝑘)ℎ(𝑛, 𝑘)
𝑘=−∞

This expression is a function of both 𝑥(𝑛) and the responses ℎ(𝑛, 𝑘) of the system to the unit
impulses 𝛿(𝑛 − 𝑘) for −∞ < 𝑘 < ∞.
If the response of the LTI system to the unit sample sequence 𝛿(𝑛) is denoted as ℎ(𝑛), that is
ℎ(𝑛) = 𝒯 [𝛿(𝑛)]

Dept. of ECE, CMRIT

then by the time-invariance property, the response of the system to the delayed unit sample
sequence 𝛿(𝑛 − 𝑘) is
ℎ(𝑛 − 𝑘) = 𝒯 [𝛿(𝑛 − 𝑘)]
Consequently, the formula in (1) reduces to
∞

𝑦(𝑛) = 𝑥(𝑛) ∗ ℎ(𝑛) = ∑ 𝑥(𝑘)ℎ(𝑛 − 𝑘)

𝑘=−∞

The formula in (3.17) that gives the response 𝑦(𝑛) of the LTI system as a function of the input
signal 𝑥(𝑛) and the unit sample (impulse) response ℎ(𝑛) is called a convolution sum.
Note: Now we observe that the relaxed LTI system is completely characterized by a single
function ℎ(𝑛), namely, its response to the unit sample sequence 𝛿(𝑛).
Suppose that we wish to compute the output of the system at some time instant, 𝑛0 . The
response at 𝑛 = 𝑛0 is
∞

𝑦(𝑛0 ) = ∑ 𝑥(𝑘)ℎ(𝑛0 − 𝑘)
𝑘=−∞

To summarize, the process of computing the convolution between x(k) and h(k) involves the
following four steps.
1. Folding. Fold ℎ(𝑘) about 𝑘 = 0 to obtain ℎ(−𝑘).
2. Shifting. Shift ℎ(−𝑘) by 𝑛0 to the right (left) if 𝑛0 is positive (negative), to obtain
ℎ(𝑛0 − 𝑘).
3. Multiplication. Multiply 𝑥(𝑘) by ℎ(𝑛0 − 𝑘) to obtain the product sequence
𝑣𝑛0 (𝑘) ≡ 𝑥(𝑘)ℎ(𝑛0 − 𝑘) .
4. Summation. Sum all the values of the product sequence 𝑣𝑛0 (𝑘) to obtain the value of
the output at time 𝑛 = 𝑛0 .

Ex: The impulse response of a linear time-invariant system is

ℎ(𝑛) = {1, 2, 1, −1}
↑
Determine the response of the system to the input signal
𝑥(𝑛) = {1, 2, 3, 1}
↑
Sol: We shall compute the convolution according to the formula, but we shall use graphs of the
sequences to aid us in the computation. We illustrate the input signal sequence 𝑥(𝑘) and the
impulse response ℎ(𝑘) of the system, using k as the time index in order to be consistent.
∞

𝑦(𝑛) = ∑ 𝑥(𝑘)ℎ(𝑛 − 𝑘) (1)

𝑘=−∞

Dept. of ECE, CMRIT

Now we can compute the output at 𝑛 = 0, in (1)
∞

𝑦(0) = ∑ 𝑥(𝑘)ℎ(−𝑘)
𝑘=−∞

Since the shift 𝑛 = 0, we use ℎ(−𝑘) directly without shifting it. The product sequence
𝑣0 (𝑘) ≡ 𝑥(𝑘)ℎ(−𝑘)

Finally, the sum of all the terms in the product sequence yields
∞

𝑦(0) = ∑ 𝑣0 (𝑘) = 2 + 2 = 4
𝑘=−∞

We continue the computation by evaluating the response of the system at n = 1.

∞

𝑦(1) = ∑ 𝑥(𝑘)ℎ(1 − 𝑘)
𝑘=−∞

The sequence ℎ(1 − 𝑘) is simply the folded sequence ℎ(−𝑘) shifted to the right by one unit
in time. This sequence is illustrated in Fig. The product sequence
Dept. of ECE, CMRIT
 𝑣1 (𝑘) ≡ 𝑥(𝑘)ℎ(1 − 𝑘)

𝑦(1) = ∑ 𝑣1 (𝑘) = 8
𝑘=−∞

In a similar manner, we obtain 𝑦(2) by shifting ℎ(−𝑘) two units to the right, forming the
product sequence

𝑣2 (𝑘) ≡ 𝑥(𝑘)ℎ(2 − 𝑘)
and then summing all the terms in the product sequence obtaining

𝑦(2) = ∑ 𝑣2 (𝑘) = 8
𝑘=−∞

By shifting ℎ(−𝑘) farther to the right, multiplying the corresponding sequence, and summing
over all the values of the resulting product sequences, we obtain

𝑦(3) = 3,
𝑦(4) = −2,

Dept. of ECE, CMRIT

 𝑦(5) = −1.
For 𝑛 > 5, we find that 𝑦(𝑛) = 0 because the product sequences contain all zeros.
Thus we have obtained the response y(n) for n > 0.
Next we wish to evaluate 𝑦(𝑛) for 𝑛 < 0. We begin with 𝑛 = −1. Then
∞

𝑦(−1) = ∑ 𝑥(𝑘)ℎ(−1 − 𝑘)
𝑘=−∞

Now the sequence ℎ(−1 − 𝑘) is simply the folded sequence ℎ(−𝑘) shifted one time unit to
the

𝑣−1 (𝑘) ≡ 𝑥(𝑘)ℎ(−1 − 𝑘)

∞

𝑦(−1) = ∑ 𝑣−1 (𝑘) = 1

𝑘=−∞

Dept. of ECE, CMRIT

 𝑦(−1) = 1
It is clear that any further shifts of ℎ(−1 − 𝑘) to the left always result in an all-zero product
sequence, and hence
𝑦(𝑛) = 0 for 𝑛 ≤ −2
Now we have the entire response of the system for −∞ < 𝑛 < ∞, which we summarize as

𝒚(𝒏) = {. . . , 𝟎, 𝟎, 𝟏, 𝟒 , 𝟖, 𝟖, 𝟑, −𝟐, −𝟏, 𝟎, 𝟎, . . . }

↑

Ex: Determine the output 𝑦(𝑛) of a relaxed linear time-invariant system with impulse response
ℎ(𝑛) = 𝑎𝑛 𝑢(𝑛), |𝑎| < 1
when the input is a unit step sequence, that is 𝑥(𝑛) = 𝑢(𝑛)

Sol: In this case both ℎ(𝑛) and 𝑥(𝑛) are infinite-duration sequences.
∞

𝑦(𝑛) = 𝑥(𝑛) ∗ ℎ(𝑛) = ∑ 𝑥(𝑘)ℎ(𝑛 − 𝑘)

𝑘=−∞

This can be equivalently computed through

Dept. of ECE, CMRIT

 ∞

𝑦(𝑛) = ℎ(𝑛) ∗ 𝑥(𝑛) = ∑ ℎ(𝑘)𝑥(𝑛 − 𝑘) (1)

𝑘=−∞

The sequences ℎ(𝑘), 𝑥(𝑘), and 𝑥(−𝑘) are shown in Fig

𝑎𝑘 , 𝑘 ≥ 0
ℎ(𝑘) = {
0, 𝑘 < 0
1, 𝑘 ≥ 0
𝑥(𝑘) = 𝑢(𝑘) = {
0, 𝑘 < 0

Dept. of ECE, CMRIT

The product sequences 𝑣0 (𝑘), 𝑣1 (𝑘) and 𝑣2 (𝑘) corresponding to 𝑥(−𝑘) ℎ(𝑘), 𝑥(1 − 𝑘)ℎ(𝑘),
and 𝑥(2 − 𝑘)ℎ(𝑘) are illustrated in Fig.
𝑦(0) = 1
𝑦(1) = 1 + 𝑎
𝑦(2) = 1 + 𝑎 + 𝑎2
Clearly, for 𝑛 > 0, the output is
𝑦(𝑛) = 1 + 𝑎 + 𝑎2 +··· + 𝑎𝑛
1 − 𝑎𝑛+1
𝑦(𝑛) = , 𝑛≥0
1−𝑎
On the other hand, for 𝑛 < 0, the product sequences consist of all zeros. Hence
𝑦(𝑛) = 0, 𝑛< 0
Since |𝑎| < 1, the final value of the output as n approaches infinity is
1
𝑦(∞) = lim 𝑦(𝑛) =
𝑛→∞ 1−𝑎

Geometric Series
If 𝛽 is a complex number, then the following relationships hold
𝑙 𝛽 𝑘 − 𝛽 𝑙+1
, 𝛽≠1
∑ 𝛽𝑛 = { 1 − 𝛽
𝑛=𝑘 𝑙 − 𝑘 + 1, 𝛽=1

𝑀−1 1 − 𝛽𝑀
, 𝛽≠1
∑ 𝛽𝑛 = { 1 − 𝛽
𝑛=0 𝑀, 𝛽=1
∞
1
∑ 𝛽𝑛 = , |𝛽| < 1
1−𝛽
𝑛=0
∞
𝛽𝑘
𝑛
∑𝛽 = , |𝛽| < 1
1−𝛽
𝑛=𝑘
∞
𝛽
∑ 𝑛𝛽 𝑛 = , |𝛽| < 1
(1 − 𝛽)2
𝑛=𝑘

Ex 2: The output 𝑦[𝑛] of a moving average system is related to input 𝑥[𝑛] as

3
1
𝑦[𝑛] = ∑ 𝑥[𝑛 − 𝑘]
4
𝑘=0
The impulse response of the system is obtained by setting 𝑥[𝑛] = 𝛿[𝑛], which yields
1
ℎ[𝑛] = (𝑢[𝑛] − 𝑢[𝑛 − 4])
4
Determine the output of the system when the input is rectangular pulse defined as
𝑥[𝑛] = 𝑢[𝑛] − 𝑢[𝑛 − 10]

Dept. of ECE, CMRIT

Sol:

First, we obtain the graph of 𝑥[𝑘] and ℎ[𝑛 − 𝑘], treating 𝑛 as constant and 𝑘 as the independent
variable.
∞

𝑦[𝑛] = 𝑥[𝑛] ∗ ℎ[𝑛] = ∑ 𝑥[𝑘]ℎ[𝑛 − 𝑘]

𝑘=−∞
1⁄ , 0 ≤ 𝑘 ≤ 3
ℎ[𝑘] = { 4
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒

1⁄ , 0 ≤ 𝑛 − 𝑘 ≤ 3
ℎ[𝑛 − 𝑘] = { 4
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
𝑛 − 𝑘 ≥ 0 and 𝑛 − 𝑘 ≤ 3
𝑛 ≥ 𝑘 and 𝑛 − 3 ≤ 𝑘
1⁄ , 𝑛 − 3 ≤ 𝑘 ≤ 𝑛
ℎ[𝑛 − 𝑘] = { 4
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒

i. For 𝑛 < 0,
No common overlap interval between 𝑥[𝑘] and ℎ[𝑛 − 𝑘]

Dept. of ECE, CMRIT

 ∞

𝑦[𝑛] = ∑ 𝑥[𝑘]ℎ[𝑛 − 𝑘]
𝑘=−∞
𝒚[𝒏] = 𝟎

ii. When 𝑛 ≥ 0 and 𝑛 − 3 ≤ 0

0 ≤ 𝑛 ≤ 3,
There is a common overlap between 𝑥[𝑘] and ℎ[𝑛 − 𝑘] for 𝑘 = 0 to 𝑛.

𝑦[𝑛] = ∑ 𝑥[𝑘]ℎ[𝑛 − 𝑘]
𝑘=−∞
𝑛

= ∑ 𝑥[𝑘]ℎ[𝑛 − 𝑘]
𝑘=0
𝑛
1
=∑
4
𝑘=0

Dept. of ECE, CMRIT

 𝟏
𝒚(𝒏) = (𝒏 + 𝟏), 𝟎≤𝒏≤𝟑
𝟒

iii. When 𝑛 − 3 > 0 and 𝑛 ≤ 9

3 < 𝑛 ≤ 9,
There is a common overlap between 𝑥[𝑘] and ℎ[𝑛 − 𝑘] for 𝑘 = 𝑛 − 3 to 𝑛.

𝑦[𝑛] = ∑ 𝑥[𝑘]ℎ[𝑛 − 𝑘]
𝑘=−∞
𝑛

= ∑ 𝑥[𝑘]ℎ[𝑛 − 𝑘]
𝑘=𝑛−3
𝑛
1
= ∑
4
𝑘=𝑛−3
1
= (𝑛 − (𝑛 − 3) + 1)
4

𝒚[𝒏] = 𝟏, 𝟑<𝒏≤𝟗

iv. When 𝑛 > 9 and 𝑛 − 3 ≤ 9

9 < 𝑛 ≤ 12,
There is a common overlap between 𝑥[𝑘] and ℎ[𝑛 − 𝑘] for 𝑘 = 𝑛 − 3 to 9.
∞

𝑦[𝑛] = ∑ 𝑥[𝑘]ℎ[𝑛 − 𝑘]
𝑘=−∞
9

= ∑ 𝑥[𝑘]ℎ[𝑛 − 𝑘]
𝑘=𝑛−3

Dept. of ECE, CMRIT

 𝑛
1
= ∑
4
𝑘=𝑛−3
1
= (9 − (𝑛 − 3) + 1)
4

𝟏𝟑 − 𝒏
𝒚[𝒏] = , 𝟗 ≤ 𝒏 ≤ 𝟏𝟐
𝟗

v. For 𝑛 > 12,

No common overlap interval between 𝑥[𝑘] and ℎ[𝑛 − 𝑘]
∞

𝑦[𝑛] = ∑ 𝑥[𝑘]ℎ[𝑛 − 𝑘]
𝑘=−∞
𝒚[𝒏] = 𝟎, 𝒏 > 𝟏𝟐

Dept. of ECE, CMRIT

 0, 𝑛<0
1
(𝑛 + 1), 0≤𝑛≤3
4
𝑦[𝑛] = 1, 3<𝑛≤9
13 − 𝑛
, 9 < 𝑛 ≤ 12
9
{ 0, 𝑛 > 12
Observations:
1. Length of 𝑥[𝑛] is 𝑙𝑥 = 10 defined for 0 ≤ 𝑛 ≤ 9
2. Length of ℎ[𝑛] is 𝑙ℎ = 4 for defined 0 ≤ 𝑛 ≤ 3
3. Length of 𝑦[𝑛] is 𝑙𝑦 = 13 for defined 0 ≤ 𝑛 ≤ 12
𝒍𝒚 = 𝒍𝒙 + 𝒍𝒉 − 𝟏

Tabulation Method:
𝑥[𝑛] = {1, 1, 1, 1, 1, 1, 1, 1, 1, 1}
↑
1 1 1 1
ℎ[𝑛] = { , , , }
4 4 4 4
↑

1 1 3 3 1 1
𝑦[𝑛] = { , , , 1, 1, 1, 1, 1, 1, 1, , , }
4 2 4 4 2 4
↑

Dept. of ECE, CMRIT

Ex: Evaluate the following discrete convolution sums:
(a) 𝑦[𝑛] = 𝑢[𝑛] ∗ 𝑢[𝑛 − 3]
Sol: Let 𝑥[𝑛] = 𝑢[𝑛] and ℎ[𝑛] = 𝑢[𝑛 − 3]
𝑦[𝑛] = 𝑥[𝑛] ∗ ℎ[𝑛]
∞

𝑦[𝑛] = ∑ 𝑥[𝑘]ℎ[𝑛 − 𝑘]
𝑘=−∞

1, 𝑘≥0
𝑥[𝑘] = 𝑢[𝑘] = {
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒

1, 𝑘 − 3 ≥ 0
ℎ[𝑘] = 𝑢[𝑘 − 3] = {
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
1, 𝑘≥3
ℎ[𝑘] = {
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒

ℎ[𝑛 − 𝑘] = 𝑢[𝑛 − 𝑘 − 3]
↑
Constant
1, 𝑛−𝑘−3≥ 0
ℎ[𝑛 − 𝑘] = {
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
1, 𝑘 ≤𝑛−3
ℎ[𝑛 − 𝑘] = {
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒

Dept. of ECE, CMRIT

 i. When 𝑛 − 3 < 0

𝒚[𝒏] = 𝟎 , 𝒏 < 𝟑

ii. When 𝑛 − 3 ≥ 0

There is a common overlap between 𝑥[𝑘] and ℎ[𝑛 − 𝑘] for 𝑘 = 0 to 𝑛 − 3.

𝑛−3

𝑦[𝑛] = ∑ 𝑥[𝑘]ℎ[𝑛 − 𝑘]
𝑘=0

Dept. of ECE, CMRIT

 𝑛−3

= ∑1
𝑘=0
= (𝑛 − 3) − 0 + 1
𝒚[𝒏] = 𝒏 − 𝟐, 𝒏≥𝟑

Hence
𝟎, 𝒏<𝟑
𝒚[𝒏] = {
(𝒏 − 𝟐), 𝒏 ≥ 𝟑

1 𝑛
(b) 𝑦[𝑛] = (2) 𝑢[𝑛 − 2] ∗ 𝑢[𝑛]

1 𝑛
Sol: Let 𝑥[𝑛] = (2) 𝑢[𝑛 − 2] and ℎ[𝑛] = 𝑢[𝑛]

𝑦[𝑛] = 𝑥[𝑛] ∗ ℎ[𝑛]

∞

𝑦[𝑛] = ∑ 𝑥[𝑘]ℎ[𝑛 − 𝑘]
𝑘=−∞

1 𝑘
𝑥[𝑘] = ( ) 𝑢[𝑘 − 2]
2
1 𝑘
= {(2) , 𝑘 − 2 ≥ 0
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒

1 𝑘
𝑥[𝑘] = { 2) , 𝑘 ≥ 2
(
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒

Dept. of ECE, CMRIT

 ℎ[𝑘] = 𝑢[𝑘]
1, 𝑘≥0
={
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
ℎ[𝑛 − 𝑘] = 𝑢[𝑛 − 𝑘]
1, 𝑛−𝑘 ≥0
={
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
1, 𝑘≤𝑛
ℎ[𝑛 − 𝑘] = {
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒

i. When 𝑛 < 2

𝒚[𝒏] = 𝟎 , 𝒏<𝟐

Dept. of ECE, CMRIT

 ii. When 𝑛 ≥ 2

There is a common overlap between 𝑥[𝑘] and ℎ[𝑛 − 𝑘] for 𝑘 = 2 to 𝑛.

𝑛

𝑦[𝑛] = ∑ 𝑥[𝑘]ℎ[𝑛 − 𝑘]
𝑘=2
𝑛
1 𝑘
= ∑( )
2
𝑘=2
1 2 1 𝑛+1
(2) − (2)
=
1
1 − (2)
1 1 𝑛−2
− (
=4 2)
1
2
1 1 𝑛−1
= −( )
2 2
Hence
𝟎, 𝒏<𝟐
𝒏−𝟏
𝒚[𝒏] = {𝟏 𝟏
−( ) , 𝒏≥𝟐
𝟐 𝟐

Dept. of ECE, CMRIT

Causal Linear Time-Invariant Systems
In the case of a linear time-invariant system, causality can be translated to a condition on the
impulse response.
Let us consider a linear time-invariant system having an output at time 𝑛 = 𝑛0 given by the
convolution formula
∞

𝑦(𝑛0 ) = ∑ ℎ(𝑘)𝑥(𝑛0 − 𝑘)
𝑘=−∞

Suppose that we subdivide the sum into two sets of terms, one set involving present and past
values of the input [i.e., 𝑥(𝑛) 𝑓𝑜𝑟 𝑛 ≤ 𝑛0 ] and one set involving future values of the input
[i.e., 𝑥(𝑛), 𝑛 > 𝑛0 ]. Thus we obtain
∞ −∞

𝑦(𝑛0 ) = ∑ ℎ(𝑘)𝑥(𝑛0 − 𝑘) + ∑ ℎ(𝑘)𝑥(𝑛0 − 𝑘)

𝑘=0 𝑘=−1

= [ℎ(0)𝑥(𝑛0 ) + ℎ(1)𝑥(𝑛0 − 1) + ℎ(2)𝑥(𝑛0 − 2)+. . . . ]

+[ℎ(−1)𝑥(𝑛0 + 1) + ℎ(−2)𝑥(𝑛0 + 2)+. . . . ]

We observe 𝑥(𝑛0 ), 𝑥(𝑛0 − 1), 𝑥(𝑛0 − 2) are the present and past values of the input signal.

𝑥(𝑛0 − 1), 𝑥(𝑛0 − 2) are the future values of the input signal.

If the output at time 𝑛 = 𝑛0 is to depend only on the present and past inputs, then, clearly, the
impulse response of the system must satisfy the condition

𝒉(𝒏) = 𝟎 , 𝒏 < 𝟎
Hence an LTI system is causal if and only if its impulse response is zero for negative values
of 𝒏.

Ex: For given impulse responses, determine whether the corresponding systems is causal.
1 𝑛
i. ℎ(𝑛) = (2) 𝑢[𝑛]

1 𝑛
Sol: ℎ[𝑛] = (2) 𝑢[𝑛]

1 𝑛
ℎ(𝑛) = { 2) ,
( 𝑛≥0
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒

Dept. of ECE, CMRIT

 The given impulse response ℎ[𝑛] = 0 for 𝑛 < 0
Hence the system is causal.
𝜋
Ex: A discrete-time system has impulse response ℎ[𝑛] = cos (2 𝑛) 𝑢[𝑛 + 3]. Is the system
causal? Justify your answers.
Sol:
𝜋
cos ( 𝑛) 𝑢[𝑛 + 3], 𝑛 ≥ −3
ℎ[𝑛] = { 2
0, 𝑛 < −3

The given impulse response ℎ[𝑛] ≠ 0 for 𝑛 < 0

Hence the system is not causal.

Dept. of ECE, CMRIT

Dept. of ECE, CMRIT