Chapter 3

Discrete-Time Signals and Systems in

Time-Domain
• Discrete-time Signals
✓ Basic Sequences and Sequence Operations
• Discrete-time Systems
• LTI Systems
• Properties of LTI Systems
• Linear Constant Coefficient Difference Equations
 Discrete-time Signals
• Discrete-time signals: are mathematically represented as sequence of numbers
• A sequence of number x, in which the 𝑛𝑡ℎ number in the sequence is denoted as x[n]
is formally written as: x = {x[n]}; −∞ < 𝑛 < ∞ , and where n is an integer.
• In practical settings, the above sequence is often arisen from a periodic sampling of
analog signals.
✓ In that case, the numeric value of the 𝑛𝑡ℎ number in the sequence is equal to
the value of analog signal
𝑥𝑎 (t), at time nT; i.e.
x 𝑛 = 𝑥𝑎 𝑡 |𝑡=𝑛𝑇 = 𝑥𝑎 𝑛𝑇 , −∞ < 𝑛 < ∞
✓ T: Sampling period (spacing b/n two consecutive samples)
1
✓ The sampling frequency is 𝐹𝑇 =
𝑇
 Discrete-time Signals
• Although sequences don’t always arise from sampling of analog waveforms,
it’s convenient to refer to x[n] as the “𝑛𝑡ℎ sample” of the sequence.

Figs:(a)Graphical representation of a discrete-time signal

 Discrete-time Signals

(b)segment of a continuous-time speech signal

(c)sequence of samples obtain from (b) with T=125µs

 Discrete-time Signals
• If a discrete-time signal is written as a sequence of a numbers inside braces,
the location of the sample value associated with the time index
n = 0 is indicated by an arrow under it.
{x[n]} = {. . . ,0.95.-0.2,2.17,1.1,0.2,-3.67, . . .}

• Classes of sequences
✓ Finite /infinite (extent in n)
✓Real/complex:
✓Left and Right-sided
 Classes of Sequences
• Complex sequence : {x[n]} ={𝑥𝑟𝑒 [n]} + j{𝑥𝑖𝑚 [𝑛]}
✓ Rectangular form ; x = 𝑥𝑟𝑒 + j𝑥𝑖𝑚
2 2 −1 𝑋𝑖𝑚
where magnitude |x| = 𝑥𝑟𝑒 + 𝑥𝑖𝑚 and phase θ = 𝑡𝑎𝑛 ( )
𝑋𝑟𝑒

✓ Polar form : x =|x|𝑒 𝑗θ

= |x| cos θ +j|x|sin θ
(𝑒 𝑗θ = cos θ +jsin θ)
• (a + jb)+(c + jd) = (a + c) + j(b + d): Addition
• r𝑒 𝑗θ . δ𝑒 𝑗ϕ =rδ𝑒 𝑗(θ+ϕ) :Multiplication
 Classes of Sequences
• The complex conjugate sequence of {x[n]} usually denoted as {x*[n]} is:
{x*[n]} ={𝑥𝑟𝑒 [n]} - j{𝑥𝑖𝑚 [𝑛]}
✓Flips imaginary part / negates phase:
✓Conjugate : x*=𝑥𝑟𝑒 − 𝑗𝑥𝑖𝑚 = |𝑥|𝑒 𝑗(−θ)
✓Useful in resolving to real quantities:
x + x*=𝑥𝑟𝑒 + 𝑗𝑥𝑖𝑚 + 𝑥𝑟𝑒 − 𝑗𝑥𝑖𝑚 =2𝑥𝑟𝑒
x.x*=|x|𝑒 𝑗(θ) .|x|𝑒 𝑗(−θ) =|𝑥|2
 Classes of Sequences
• Finite–length (finite-duration): defined only for finite time interval:
𝑁1 ≤ 𝑛 ≤ 𝑁2 , where -∞ < 𝑁1 and 𝑁2 < ∞ with 𝑁2 ≥𝑁1
✓The length of the above finite-length sequence is:
N = 𝑁2 -𝑁1 +1
• A length-N discrete-time sequence: consists of N samples often referred to as a N-point
sequence.
• A finite-length sequence can be considered as infinite-length sequence by zero padding
• A right-sided sequence: x[n] has zero-valued samples for n < 𝑁1 i.e.
x[n] =0 for n < 𝑁1
where 𝑁1 ≥0, a right-sided sequence is usually called causal sequence
 Classes of Sequences
• Left-sided sequence : X[n] has zero-valued samples for n > 𝑁2 , i.e., X[n] =0 for n > 𝑁2
where 𝑁2 is finite-integer which can be negative or positive.
If 𝑁2 ≤0, a left-sided sequence is usually called anti causal.
• A general two-sided sequence is defined for all values of n in the range −∞ < 𝑛 < ∞ .

Fig(a) A left-sided sequence

(b) A right-sided sequence
(c) A two-sided sequence
 Basic Sequences and Sequence operations
Basic operations on sequences :
• Addition operation:

• Multiplication operation:
 Basic Operations On Sequences
• Product (modulation) operation:

✓Multiplying an infinite-length sequence by a finite-length window sequence extract the

region
 Basic Operations on Sequences
• Time shifting operation y[n] = x[n – N] where N is an integer
• If N > 0, it is delaying operation
• Unit delay : x[n] Z−1 y[n] y[n] = x[n – 1]

If N< 0, it is an advance operation

✓ Unit advance : x[n] Z y[n] y[n] = x[n +1]

• Time-reversal or folding : x[n] x[-n]

 Basic Operations on Sequences
• Combination of basic operations:

Y[n] = α1 X[n] + α2 X[n-1] + α3 X [n-2] + α4 X[n-3]

 Basic sequences
• In discussing the theory of discrete-time signals and systems several basic
sequences are of particular importance.
I. The unit sample sequence

0, n = 0 also called discrete-time impulse

δ[n] = 1, n = 0
𝟏 𝒏=𝒌
𝜹[𝒏 − 𝒌] = {
𝟎 𝒏≠𝒌
• The unit sample sequence plays the same role for discrete-time signals and
systems that unit impulse function (Dirac-delta function) does for continuous
s-time signals and systems.
Basic Sequences
 Basic Sequences
• An arbitrary sequence can be represented as a sum of scaled, delayed discrete-impulses.
p[n} =𝑎−3 δ[n+3] + 𝑎1 δ[n-1] + 𝑎2 δ[n-2] + 𝑎1 δ[n-7]
• More generally any sequence can be expressed as
x[n] = σ∞ 𝑘=−∞ 𝑥 𝑘 δ[𝑛 − 𝑘]
• The unit step sequence:

𝟏 𝒏≥𝒌
𝒖[𝒏 − 𝒌] = {
𝟎 𝒏<𝑘
 Basic Sequences
• The unit step is related to the impulse by:
u[n] = σ𝑛𝑘=−∞ δ[𝑘];
i.e., the values of the unit step sequence at (time) index n is equal to the accumulated sum of
the values at index n and all previous values of the impulse sequence.
• An alternative representation of the unit step in terms of the impulse is obtained by
interpreting the unit step in terms of a sum of delayed impulses.
u[n] = δ[n] +δ[n-1] + δ[n-2] + . . . . Or
u[n] = σ∞ 𝑘=0 δ[𝑛 − 𝑘]
• Conversely, the impulse sequence can be expressed as the first backward difference of the
unit step sequence.
i.e. , δ[n] = u[n] – u[n-1]
 Basic Sequences
Exponential Sequence:
✓Important in representing and analyzing LTI discrete-time systems:
x[n] = Aα𝑛
✓If A and α are real numbers, then the sequence is real.
✓If 0 < α < 1 and A is positive, then the sequence values are positive
and decrease with increasing n.
✓If -1 < α <0, the sequence values alternate in sign, but again decrease
in magnitude with increasing n.
✓If |α|>1, then the sequence grows in magnitude with increasing n.
Basic Sequences
 Basic Sequences
• Sinusoidal sequences:
✓A general form: x[n] = Acos ω0 𝑛 + ϕ , for all n, A and ϕ are real
constants.
• The exponential sequence Aα𝑛 with complex A and α has real and
imaginary parts that are exponentially weighted sinusoids.
✓ α = |α|𝑒 𝑗ω0 ,and A = |A|𝑒 𝑗ϕ
✓x[n] = Aα𝑛 = |A|𝑒 𝑗ϕ . |α|𝑛 𝑒 𝑗ω0 𝑛
= |A||α|𝑛 𝑒 𝑗(ω0𝑛+ϕ)
= |A||α|𝑛 cos ω0 𝑛 + ϕ + j|A||α|𝑛 sin ω0 𝑛 + ϕ
✓The sequence oscillates with an exponentially growing envelope if
|α|> 1 or with exponentially decaying envelope if |α| < 1.
 Basic Sequences
✓ If |α|= 1,the sequence is referred to as a complex exponential sequence
and has the form:
x[n] = |A|𝑒 𝑗(ω0 𝑛+ϕ) = |A|cos ω0 𝑛 + ϕ +j |A|sin ω0 𝑛 + ϕ
✓ Varies with n (samples)
✓ω0 is frequency of the complex exponential (radians/sample)
✓Φ is phase.
• Consider a frequency (ω0 +2π) |α|= 1, |A| = real
x[n] = |A|𝑒 𝑗 ω0 𝑛+2π 𝑛

= A𝑒 𝑗ω0 𝑛 𝑒 𝑗2𝜋𝑛 =A𝑒 𝑗ω0 𝑛

Signal Manipulations / Basic Operations
These manipulations are generally compositions of a few basic signal transformations.
The most common transformations include shifting, reversal, and scaling, which are defined below.

I. Shifting: If 𝒚[𝒏] = 𝒙[𝒏 – 𝒏𝒐], x(n) is shifted to the right by no samples if no is positive (this is referred to
as a delay), and it is shifted to the left by no samples if no is negative (referred to as an advance).
II. Reversal: If 𝒚[𝒏] = 𝒙[−𝒏], x(n) is inverted about vertical axis. This transformation involves "flipping"
the signal x(n) with respect to the index n.
𝑛
III. Time Scaling: This transformation is defined by 𝑥(𝑀𝑛) 𝑜𝑟 𝑥( ) where M and N are positive integers.
𝑁
If 𝑦[𝑛] = 𝑥[𝑀𝑛] (the sequence x(Mn) is formed by taking every Mth sample of x(n)), then this operation is
known as down sampling.
𝑛
If 𝑦(𝑛) = 𝑥( ), then this operation is known as up sampling and is defined as follows
𝑁
𝒏
𝒙[ ] 𝒏 = 𝟎, ±𝑵, ±𝟐𝑵, ±𝟑𝑵, … … … … …
𝒚𝒏 ={ 𝑵
𝟎 𝒐𝒕𝒉𝒆𝒓𝒘𝒊𝒔𝒆
Example: Consider a discrete time signal x[n] as shown in figure. For the given signal
perform the following operations.
1. x[n-2] (Shifting)
2. x[-n] (Reversal)
3. x[2n] (Time scaling-down sampling)
4. x[n/2] (Time scaling- up sampling)
Ans:-
Addition, Multiplication, and Scaling
The most common types of amplitude transformations are addition, multiplication, and scaling.
Addition: The sum of two signals is formed by the point wise addition of the signal values.
𝒚[𝒏] = 𝒙𝟏 [𝒏] + 𝒙𝟐 [𝒏] −∞ <𝑛 <∞
Multiplication: The multiplication of two signals is formed by the pointwise product of the signal values.
𝒚[𝒏] = 𝒙𝟏 [𝒏]𝒙𝟐 [𝒏] −∞ <𝑛 <∞
Scaling: Amplitude scaling of a signal x(n) by a constant c is accomplished by multiplying every signal value by c:
𝒚[𝒏] = 𝒄𝒙[𝒏 ] −∞ <𝑛 <∞
Signal Decomposition

The unit sample may be used to decompose an arbitrary signal x(n) into a sum of weighted and shifted unit
samples as follows:
𝒙[𝒏] = ⋯ + 𝒙[−𝟏]𝜹[𝒏 + 𝟏] + 𝒙[𝟎]𝜹[𝒏] + 𝒙 𝟏 𝜹[𝒏 − 𝟏] + 𝒙[𝟐]𝜹[𝒏 − 𝟐] + ⋯
This decomposition may be written concisely as
∞

𝒙[𝒏] = ෍ 𝒙 𝒌 𝜹[𝒏 − 𝒌]
𝒌=−∞
Block Diagram Representation of Discrete-Time Systems

i. Constant multiplier:

ii. Adder :

iii. Signal multiplier :

iV. Unit delay :

FIR and IIR filters:
Discrete-time LTI systems can be classified into FIR or IIR systems, that is, having finite or infinite
impulse response h(n).
An FIR filter has impulse response h(n) that extends only over a finite time interval, say0 ≤ n ≤ M, and is
identically zero beyond that:{h0, h1, h2, . . . , hM, 0, 0, 0, . . . }
The I/O equation is:

(FIR filter equation)

An IIR filter, on the other hand, has an impulse response h(n) of infinite duration, defined over the infinite
interval −∞ < n < ∞.{… h-1, h0, h1, h2, . . . }
The I/O equation is:

(IIR filter equation)

 Symmetry
• Generally complex exponential sequences with frequencies(ω0 +2πr), where r is an
integer, are indistinguishable from one another.
• Complex exponential signals of the form
x[n] = A𝑒 𝑗ω0 𝑛 or a real sinusoidal signals of the form:
x[n] = Acos ω0 𝑛 + ϕ we need only to consider frequencies in an interval of
length 2π, ( such as - π <ω0 < π or 0<ω0 < 2π ).
• Conjugate symmetric:
✓ A sequence x[n] is called conjugate-symmetric sequence if x[n] = x*[-n]
✓A real conjugate-symmetric sequence is called an even sequence.
✓ A sequence x[n] is called conjugate-antisymmetric sequence if x[n] = -x*[-n] Ɐn
 Symmetry
• A real conjugate-antisymmetric sequence is called an odd sequence
• Any complex signal may always be decomposed into a sum of a conjugate
symmetric signal and a conjugate anti-symmetric signal.
✓ Even signal: A real-valued signal is even if x[n] = x[-n], for all n
✓ Odd signal: A real-valued signal is odd if
x[n] = -x[-n], for all n
• Any signal x[n] may be decomposed into sum of its even part ,𝑥𝑒 [𝑛] and its odd
part, 𝑥𝑜 [𝑛] .
x[n] = 𝑥𝑒 [𝑛] + 𝑥𝑜 [𝑛]
1
𝑥𝑒 [𝑛] = {x[n] + x[-n]}
2
1
𝑥𝑜 [𝑛] = {x[n] - x[-n]}
2
 Symmetry
• A complex sequence x[n] can also be expressed as sum of its conjugate-symmetric
part 𝑥𝑐𝑠 [𝑛] and its conjugate-antisymmetric part 𝑥𝑐𝑎 [𝑛]:
x[n] = 𝑥𝑐𝑠 [𝑛] + 𝑥𝑐𝑎 [𝑛]
1
𝑥𝑐𝑠 [𝑛] = {x[n] + x*[-n]}
2
1
𝑥𝑐𝑎 [𝑛] = {x[n] – x*[-n]}
2
Example: - Find the even and odd parts of the following signal x[n] = u[n].
Ans:-
Let 𝑥[𝑛] = 𝑢[𝑛]
𝟏 𝟏 𝒏=𝟎
The even part of u[n] is 𝒙𝒆 𝒏 = 𝒖[𝒏] + 𝒖[−𝒏] = { 𝟏
𝟐 𝒏≠𝟎
𝟐

𝟏 𝟏
which is written as 𝒙𝒆 𝒏 = + 𝜹[𝒏]
𝟐 𝟐

1/2 𝑛 > 0
𝟏 0 𝑛 = 0)
The odd part of u[n] is, 𝒙𝒐 [𝒏] = 𝒖 𝒏 − 𝒖[−𝒏] = −1/2 𝑛 < 0
𝟐

𝟏
which is written as 𝒙𝒐 [𝒏] = 𝒔𝒈𝒏(𝒏)
𝟐
For complex sequences the symmetries of interest are slightly different.
Periodic and Aperiodic Sequences
A discrete-time signal may always be classified as either being periodic or aperiodic.
A signal x(n) is said to be periodic if, for some positive real integer N,
𝒙[𝒏] = 𝒙[𝒏 + 𝑵] 𝒇𝒐𝒓 𝒂𝒍𝒍 𝒏.
This is equivalent to saying that the sequence repeats itself every N samples.

Note: If x1(n) is a sequence that is periodic with a period N1, and x2(n) is another sequence
that is periodic with a period N2,the sum
𝒙[𝒏] = 𝒙𝟏 [𝒏] + 𝒙𝟐 [𝒏]

will always be periodic and the fundamental period is

𝑁1 𝑁2
𝑁= = 𝐿𝐶𝑀 (𝑁1 , 𝑁2 )
𝑔𝑐𝑑 𝑁1 , 𝑁2
where gcd(N1, N2) means the greatest common divisor of N1 and N2.
Example:- Determine whether the given signal is periodic or aperiodic. If it is
periodic then find its fundamental period?
𝑥 𝑛 = cos(0.125𝜋𝑛)

Ans:-
Because 0.125𝜋 = 𝝅/𝟖 , and

𝝅 𝝅
𝐜𝐨𝐬 𝒏 = 𝐜𝐨𝐬 𝒏 + 𝟏𝟔
𝟖 𝟖

2𝜋
x[n] is periodic with period N = = 16
Ω
 Energy and Power Signals
• The total energy of a sequence x[n]:
𝐸𝑥 = σ∞ 𝑛=−∞ |𝑥 𝑛 | 2

✓ The energy of a signal can be finite or infinite.

✓If 0 < 𝐸𝑥 < ∞, then x[n] is called energy signal.
✓ The average power of a periodic sequence x[n]:
𝑁
1
𝑃𝑥 = lim ෍ |𝑥 𝑛 |2
𝑁→∞ 2𝑁 + 1
𝑛=−𝑁
✓ If we define the energy of signal x[n] over finite interval –N ≤ n ≤ N as:
N

EN = ෍ |x n |2
n=−N
 Energy and Power Signals

• Then the average power of the signal x[n]:

1
𝑃𝑥 = lim 𝐸𝑁
𝑁→∞ 2𝑁+1
✓ If EN is finite, p = 0
✓ If E is infinite the average power P may be finite or infinite.
✓ If P is finite ( and non zero), the signal is called a power signal.
• The average power of a periodic sequence x[n] with period N:
1 N−1
𝑃𝑥 = σn=0 |x n |2
𝑁
Example: Determine the power and energy of the unit step sequence.
 Energy and Power Signals
• The average power of the unit step sequence.
𝑁
1
𝑃𝑥 = lim ෍ 𝑢2 [𝑛]
𝑁→∞ 2𝑁 + 1
𝑛=0
1
𝑁+1 1+ 𝑁 1
= lim = lim 1 =
𝑁→∞ 2𝑁+1 𝑁→∞ 2+𝑁 2

• The unit step sequence is a power signal. Its energy is infinite.

Exercise: Show that the complex exponential sequence x[n] = A𝑒 𝑗ω0𝑛 has
average power of 𝐴2 .
 Discrete-Time Systems
• Discrete-time system is defined mathematically as a transformation or operator that
maps an input sequence with values x[n] into an output sequence with values y[n].
y[n] = T{x[n]}

x[n] T{.} y[n]

Fig: A transformation that maps an input sequence x[n] into unique output sequence y[n]
✓ Represents a rule or formula for computing the o/p sequence from the input sequence
values.
Examples:
i. The ideal delay system
y[n] = x[n-nd ], -∞ < 𝑛 <∞ nd : fixed positive integer
 Discrete-time Systems
ii. Moving Average
1
y[n] = σ𝑀 2
𝑥[𝑛 − 𝐾]
M1 + M2 +1 𝑘=−𝑀1
1
= (x[n+M1 ) + x[n+M1 − 1] + . . . +x[n] + x[n − 1]+ . . . +x[n − M2 ])
M1 + M2 +1
Input-output relationship:
T
x[n] →y[n] y[n] is the response of the system T to the excitation of x[n].
• Classes of systems are defined by placing constraints on the properties of the
transformation
• T{ . }.
 Classification Of Discrete-time System
1. Memoryless System
✓ Output y[n] at every value of n depends only on the input x[n] at the same
value of n.
Example.:
a. y[n] = (𝑥[𝑛])2 , for each value of n is memoryless system.
b. y[n] = x[n-𝑛𝑑 ], for each value of n if 𝑛𝑑 = 0, the system is memoryless.
c. 𝑦 𝑛 = 𝑥 𝑛 + 𝑥[𝑛 − 1] --------------non memoryless system(with memory)
d. The moving average system is not memoryless unless 𝑀1 = 𝑀2 = 0.
2. Linear systems
• Defined by the principle of superposition.
• A system that is both additive and homogeneous is said to be linear.
o An additive system is one for which the response to a sum of inputs is equal to the sum of
the inputs individually.
o A system is said to be homogeneous if scaling the input by a constant results in a scaling
of the output by the same amount.
If 𝑦1 [n] and 𝑦2 𝑛 are responses of a system when 𝑥1 [n] and 𝑥2 [n] are the respective inputs,
then the system is linear if and only if:
T{𝑥1 [n] + 𝑥2 [n] } = T{𝑥1 [n] } + T{𝑥2 [n] } = 𝑦1 [n] + 𝑦2 [n] and
T{ ax[n] } = aT{x[n]} = ay[n] where a is an arbitrary constant.
✓ The first property is called additive property and the second is called the homogeneity or
scaling property.
✓ These two properties can be combined into the principle of superposition.
T{𝑎𝑥1 [n] + 𝑏𝑥2 [n] } = aT{𝑥1 [n] } + bT{𝑥2 [n] } for arbitrary constant a and b.
 Continued …
• Generalizing for many inputs
If x[n] = σk ak xk n , then the o/p of the linear system y[n] = σk ak yk n
where yk n is the system response to input xk n .
Example: The accumulator system:
y[n] = σ𝑛𝑘=−∞ 𝑥[𝑘]
✓ The o/p at time n is just the sum of the present and all previous input samples.
✓ The accumulator system is a linear system.
Proof: Show that it satisfies the superposition principle for all the inputs.
Define two arbitrary inputs x1 [𝑛] and x2 [𝑛] and there corresponding outputs
y1 [𝑛] = σ𝑛𝑘=−∞ x1 [𝑘],
y2 [𝑛] = σ𝑛𝑘=−∞ x2 [𝑘]
✓When the input is x3 [𝑛] = ax1 [𝑛] + bx2 [𝑛] , the superposition principle requires the
output y3 [𝑛] = ay1 [𝑛] + by2 [𝑛] for all possible values a and b.
 Continued …

y3 [𝑛] = σ𝑛𝑘=−∞ x3 [𝑘]

= σ𝑛𝑘=−∞ (ax1 [𝑘] + bx2 [𝑘])
= aσ𝑛𝑘=−∞ x1 [𝑘] + b σ𝑛𝑘=−∞ x2 [𝑘]
= ay1 [𝑛] + by2 [𝑛]
• Consider the system defined by
w[n] = log10 (𝑥 𝑛 ) is a nonlinear system
✓Disprove considering x1 𝑛 = 1 𝑎𝑛𝑑 x2 𝑛 = 10
 3. Time-invariant systems
✓Also referred to as shift-invariant system.
✓ Is a system for which a time-shift or delay of the input sequence causes a
corresponding shift in the output sequence.
• x[n] y[n] (a system transform the input sequence with values x[n] into
the output sequence with values y[n]).
✓ Then the system is said to be time-invariant if for all 𝑛0 ,the input sequence
with values 𝑥1 𝑛 = 𝑥[𝑛 − 𝑛0 ] produces the output sequence with values
𝑦1 𝑛 = 𝑦[𝑛 − 𝑛0 ] .
Example: The accumulator a time-invariant system
𝑥1 𝑛 = 𝑥[𝑛 − 𝑛0 ]
𝑛−𝑛0
𝑦[𝑛 − 𝑛0 ] =σ𝑘=−∞ 𝑥[𝑘]
✓ Next:
Continued …
y1 [𝑛] = σ𝑛𝑘=−∞ x1 [𝑘] = σ𝑛𝑘=−∞ x1 [𝑘 − 𝑛0 ]
Let 𝑘1 = 𝑘 − 𝑛0
𝑛−𝑛0
y1 [𝑛] = σ𝑘=−∞ x1 [𝑘1 ] = 𝑦[𝑛 − 𝑛0 ] accumulator is time-invariant system.
• 𝟐. 𝑦 𝑛 = 𝑥 𝑛 − 𝑥 𝑛 + 3 −−−− −𝑡𝑖𝑚𝑒𝑖𝑛𝑣𝑎𝑟𝑖𝑎𝑛𝑡
• 𝟑. 𝑦 𝑛 = 𝑛𝑥 𝑛 −−−−−−−− −𝑛𝑜𝑛𝑡𝑖𝑚𝑒𝑖𝑛𝑣𝑎𝑟𝑖𝑎𝑛𝑡

4. Causality
• A system is causal, if, for every choice of 𝑛0 , the output sequence value at index
n = 𝑛0 depends only on the input sequence values for n≤ 𝑛0 .
✓If 𝑥1 [𝑛] = 𝑥2 [𝑛] for n≤ 𝑛0
then 𝑦1 [𝑛] = 𝑦2 [𝑛] for n≤ 𝑛0 , the system is not anticipative.
 Continued …
Example: The forward difference system y[n] = x[n+1] –x[n]
✓This system is not causal, as the current value of the output depends on a future
value of the input.
✓The backward difference system: y[n] = x[n] –x[n-1] output depends only on the
present and past values of the input.

5. Stability
• A system is stable in the bounded-input, bounded-output (BIBO) sense if every
bounded input sequence produces a bounded output sequence.
• The input x[n] is bounded if there exists a fixed positive finite value 𝐵𝑥 ,such
that
|x[n]| ≤ 𝐵𝑥 < ∞ Ɐn.
 Continued …

• Stability requires for every bounded input there exists a fixed

positive finite value 𝐵𝑦 , such that |y[n]| ≤ 𝐵𝑦 < ∞ Ɐn
Example: Accumulator x[n] = u[n] , 𝐵𝑥 = 1
y[n] = σ𝑛𝑘=−∞ 𝑢[𝑘]
= 0,n< 0
(n+1),n≥ 0
• There is no finite choice for 𝐵𝑦 , such that (n+1)≤ 𝐵𝑦 < ∞, Ɐn.
✓ The system is unstable.
 Linear Time-invariant system
• A sequence of the linear, time-invariance property is that an (LTI) discrete-time
is completely characterized by its impulse response
• Impulse response: the response of a digital filter to unit sample
sequence {δ[n]}.
• Knowing the impulse-response, we can compute the output of the
system to any arbitrary input.
✓ h[n] denote the impulse response of the LTI discrete-system; the
response to an input δ[n],
If x[n] = δ[n-1]
y[n] = h[n-1]
 LTI Systems
δ[n+2], δ[n-4], δ[n-6]

h[n+2],h[n-4],h[n-6]
• i/p: x[n] = 0.5 δ[n+2] +1.5 δ[n-1]- δ[n-2]+ δ[n-4]+0.75 δ[n-6]
• o/p: y[n] =0.5h[n+2] 1.5h[n-1]-h[n-2]+h[n-4]+0.75h[n-6]

• An arbitrary input sequence x[n] can be expressed as a weighted linear

combination of delayed and advanced unit sample sequences.
x[n] = σ∞ −∞ 𝑥[𝑘]δ[n-k]

• The response of the LTI discrete-time system to the sequence x[k]δ[n-k],

y[n] = T{σ∞ −∞ 𝑥[𝑘]δ[n-k]}
y[n] = σ∞−∞ 𝑥[𝑘]T{δ[n-k]}
 LTI Systems

y[n] = σ∞
−∞ 𝒙[𝒌] 𝐡[n-k] ….. Convolution sum

• A LTI system is completely characterized by its impulse response h[n].

i.e. Given h[n], its possible to compute the o/p y[n] due to any input x[n].
y[n] = x[n] *h[n] -> compact representation
✓ Each sample of the o/p sequence expressed in terms of all the samples of the
i/p and impulse response sequences.
 Properties of LTI systems
• Since all time-invariant systems are described by the convolution sum, the
properties of this class of systems are defined by the properties of discrete-time
convolution.
✓ The impulse response (IR) is a complete characterization of the properties of a
specific LTI system.
• For class of linear time-invariant systems.
I. The convolution operation is commutative;
x[n]*h[n] = h[n]*x[n]
✓ The order of the sequences in a convolution is un important.
II. The convolution operation is distributive over addition;
x[n]*( ℎ1 [n]+ℎ2 [n]) = x[n]*( ℎ1 [n])+x[n]*ℎ2 [n]
 Continued …
• LTI systems, in cascade corresponds to a LTI system with an IR that is the
convolution of the IRs of the two systems

x[n] h1 [n] h2 [n] y[n]

Fig: Three linear time-invariant systems with identical impulse responses.

 Continued …
• Parallel connection:
✓ The systems have the same input and their outputs are summed to produce an
overall output.
Fig: (a) parallel combination of LTI system (b) An equivalent system
 Stability Condition in terms of IR
• Linear time-invariant systems are BIBO stable if and only if the
impulse response is absolutely summable. i.e. if
s=σ∞ 𝑘=−∞ |ℎ 𝑘 | < ∞
✓ This can be shown as:
σ ∞ σ∞
|y[n]|= | 𝑘=−∞ ℎ 𝑘 𝑥 𝑛 − 𝑘 | ≤ 𝑘=−∞ |ℎ 𝑘 |𝑥[𝑛 − 𝑘]
if x[n] is bounded, so that |x[n]| ≤ 𝐵𝑥
• By substituting 𝐵𝑥 for x[n-k],
|y[n]|=𝐵𝑥 σ∞ 𝑘=−∞ |ℎ 𝑘 |, thus y[n] is bounded if
σ∞ 𝑘=−∞ |ℎ 𝑘 | < ∞
Causal Systems
• Causal systems: y[𝑛0 ] depends only on the input samples x[n], for n ≤ 𝑛0
✓ This implies the condition h[n] = 0, n< 0 for causality of LTI systems.
• Properties of linear time-invariant systems reflected in the impulse response:
• Examples:
I. Ideal Delay
h[n] = δ[n-𝑛𝑑 ], 𝑛𝑑 a positive fixed integer.

II. Moving average

1 𝑀2
h[n] = σ𝑘=−𝑀 δ[𝑛 − 𝑘]
𝑀1 +𝑀2 +1 1
1
= , 𝑀1 ≤ 𝑛 ≤ 𝑀2
𝑀1 +𝑀2 +1
0,otherwise
 Continued …
III. Accumulator:
h[n] = σnk=−∞ δ[k]
1,n≥ 0
= 0,n< 0
=u[n]
IV. Forward difference:
h[n] = δ[𝑛 + 1]- δ[𝑛]
✓ Given the impulse response of these basic systems, we can test the stability of
each by computing:
s = σ∞n=−∞ |h n |
 Continued …
• For ideal delay, moving average and forward difference, s < ∞, since the
impulse response has only a finite number of non zero samples.
✓ Such systems are called finite duration impulse response (FIR) systems.
✓ FIR systems are always stable.
• The accumulator is unstable
s = σ∞n=0 𝑢 n = ∞
✓The impulse response of the accumulator is infinite duration. These class of
systems are referred to as infinite-duration impulse response (IIR) systems.
✓ h[n] = 𝑎𝑛 u[n] with |a| <1 is IIR system that is stable.
∞ 𝑛 1
s = σn=0 |𝑎| = < ∞ (formula: sum of the terms of an infinite
1−|𝑎|
geometric series)
 Convolution
The relationship between the input to a linear time shift-invariant (LTI) system,
x[n],and the output, y[n],is given by the convolution sum:
∞

𝑦 𝑛 =𝑥 𝑛 ∗ℎ 𝑛 = ෍ 𝑥 𝑘 ℎ 𝑛−𝑘 − − − − 𝐿𝑇𝐼 𝑓𝑜𝑟𝑚

𝑘=−∞
∞

𝑦 𝑛 =ℎ 𝑛 ∗𝑥 𝑛 = ෍ ℎ 𝑘 𝑥 𝑛−𝑘 − − − − 𝐷𝑖𝑟𝑒𝑐𝑡 𝑓𝑜𝑟𝑚

𝑘=−∞
Convolution Properties:
Commutative Property:𝑥 𝑛 ∗ ℎ 𝑛 = ℎ 𝑛 ∗ 𝑥 𝑛
Associative Property: 𝑥 𝑛 ∗ ℎ1 𝑛 ∗ ℎ2 [𝑛] = 𝑥 𝑛 ∗ {ℎ1 𝑛 ∗ ℎ2 [𝑛]}
Distributive Property: 𝑥 𝑛 ∗ ℎ1 𝑛 + ℎ2 𝑛 = 𝑥 𝑛 ∗ ℎ1 𝑛 + 𝑥 𝑛 ∗ ℎ2 [𝑛]
Performing Convolutions

There are several different approaches that may be used, and the one that is the easiest will
depend upon the form and type of sequences that are to be convolved.

i. Direct Evaluation:
In performing convolutions directly, it is usually necessary to evaluate finite or infinite sums
involving terms of the form 𝛼 𝑛 or 𝑛𝛼 𝑛 .
Example: Let us perform the convolution of the two signals:
𝑥 𝑛 = 𝑎𝑛 𝑢 𝑛 𝑎𝑛𝑑 ℎ 𝑛 = 𝑢(𝑛)
With the direct evaluation of the convolution sum we find
∞ ∞ 𝑛
1 − 𝑎 𝑛+1
𝑦 𝑛 = 𝑥 𝑛 ∗ ℎ 𝑛 = ෍ 𝑥 𝑘 ℎ 𝑛 − 𝑘 = ෍ 𝑎𝑘 𝑢 𝑘 𝑢 𝑛 − 𝑘 = ෍ 𝑎𝑘 =
1−𝑎
𝑘=−∞ 𝑘=−∞ 𝑘=0
1−𝑎𝑛+1
Therefore 𝑦𝑛 = 𝑢(𝑛)
1−𝑎
•Consider a system with impulse response:
h[n] = u[n] –u[n-N]

• The input is: x[n] = 𝑎𝑛 u[n]

✓ To find the o/p at a particular index n, we must form the sums over all k of
the product x[k] and h[n-k].
✓ Clearly, the non zero portions of the sequences x[k] and h[n-k] don’t
overlap for all negative value of n.
y[n] =0, n< 0
 Continued …
•For 0≤ 𝑛 ≤ 𝑁 − 1
x[k]h[n-k] = 𝑎𝑘
y[n] = σ𝑛𝑘=0 𝑎𝑘 for 0≤ 𝑛 ≤ 𝑁 − 1
✓ From geometric series:
𝑁2 𝑘 α𝑁1 −α𝑁2 +1
σ𝑘=𝑁 α = , 𝑁2 ≥ 𝑁1
1 1−α
1−𝑎𝑛+1
y[n] = , 0≤ 𝑛 ≤ 𝑁 − 1
1−𝑎
• For n-N+1 ≤ 𝑘 ≤ 𝑛
x[k]h[n-k] = 𝑎𝑘 , n-N+1 ≤ 𝑘 ≤ 𝑛
y[n] = σ𝑛𝑘=𝑛−𝑁+1 𝑎𝑘 for N-1< 𝑛
𝑎𝑛−𝑁+1 −𝑎𝑛+1 1−𝑎 𝑁
y[n] = = 𝑎𝑛−𝑁+1 ( )
1−𝑎 1−𝑎
 Continued …
•The closed-form expression of y[n] as a function of the index n:
0, n< 0
1−𝑎𝑛+1
y[n] = , 0≤ 𝑛 ≤ 𝑁 − 1
1−𝑎
1−𝑎 𝑁
𝑎 𝑛−𝑁+1 ( ), 𝑁 − 1 < 𝑛
1−𝑎
Closed-form Expressions for Some Commonly Encountered Series
ii. Graphical Approach:

The steps involved in using the graphical approach are as follows:

✓ Plot both sequences, 𝑥[𝑘] and ℎ[𝑘], as functions of k.
✓ Choose one of the sequences, say ℎ[𝑘], and time-reverse it to form the sequence
ℎ −𝑘 .
✓ Shift the time-reversed sequence by n. [Note: If n > 0, this corresponds to a shift to
the right (delay),whereas if n <0, this corresponds to a shift to the left (advance).]
✓ Multiply the two sequences x[k] and h[n – k] and sum the product for all values of
k. The resulting value will be equal to 𝑦[𝑛]. This process is repeated for all possible
shifts, n.
A useful fact to remember in performing the convolution of two finite-length sequences
is that if x(n) is of length L1and h [n] is of length L2, y[n] = x[n] ∗ h[n] will be of
length L1 + L2 − 1.
 ✓The sequence h[n-k], - ∞ < 𝑘 < ∞,is obtained by;
I. Reflecting h[k] about the origin to obtain h[-k]
II. Shifting the origin of the reflected sequence to k =n

Note: h[n-k] = h[-(k-n)]

Example: To illustrate the graphical approach to convolution, let us evaluate y(n) = x(n)*h(n)
where x(n) and h(n) are the sequences shown in Fig. ( a ) and (b), respectively.
To perform this convolution, we follow the steps listed above:

1. Because x(k) and h(k) are both plotted as a function of k in Fig. (a) and (b), we next choose
one of the sequences to reverse in time. In this example, we time-reverse h(k), which is shown
in Fig. (c).
2. Forming the product, x(k)h(-k), and summing over k, we find that y(0) = 1.
3. Shifting h(k) to the right by one results in the sequence h(l - k ) shown in Fig. (d). Forming
the product, x(k)h(l - k), and summing over k, we find that y(1) = 3.
4. Shifting h(l - k) to the right again gives the sequence h(2 - k) shown in Fig. (e). Forming the
product, x(k)h(2 - k), and summing over k, we find that y(2) = 6.
5. Continuing in this manner, we find that y(3) = 5. y(4) = 3, and y(n) = 0 for n > 4.
6. We next take h(-k) and shift it to the left by one as shown in Fig. (f ). Because the product,
x(k)h(- 1 - k), is equal to zero for all k, we find that y(- I ) = 0. In fact. y(n) = 0 for all n < 0.
Figure (g) shows the convolution for all n.
iii. Slide Rule Method:
The steps involved in the slide rule method are as follows:
1. Write the values of 𝑥[𝑘]along the top of a piece of paper, and the values of ℎ[−𝑘]along the top
of another piece of paper
2. Line up the two sequence values 𝑥[0]and ℎ[0], multiply each pair of numbers, and add the
products to form the value of 𝑦[0].
3. Slide the paper with the time-reversed sequence h[k] to the right by one, multiply each pair of
numbers, sum the products to find the value 𝑦[𝑛] , and repeat for all shifts to the right by 𝑛 > 0.
Do the same, shifting the time-reversed sequence to the left, to find the values of y[n] for 𝑛 < 0.

Example: Find the convolution of the two finite-length sequences using the slide rule.
𝑛𝜋
𝑥 𝑛 = 0.5𝑛 𝑢 𝑛 − 𝑢 𝑛 − 5 𝑎𝑛𝑑 ℎ 𝑛 = 2 sin [𝑢 𝑛 + 2 − 𝑢 𝑛 − 4 ]
2
iV. Convolution Table:
Consider finite length input and finite impulse response, each output y(n) is the sum of all
possible products hixj with i + j = n.
Example for i=3 & j=4, the convolution table is:

Fig 2.3 Convolution table.

➢ The nth row of the table is filled by multiplying the xnsamples by the corresponding hn
sample for that row.
➢ Then, the table is “folded” along its anti-diagonal lines. In the ij- plane, the condition i+j = n
represents the nth anti-diagonal straight line.
➢ Therefore, the entries within each anti-diagonal strip are summed together to form the
corresponding output value. There are as many anti-diagonal strips as output samples yn.
Example : Calculate the convolution of the following impulse response and input signals:
h(n)= [1, 2,−1, 1],
x(n)= [1, 1, 2, 1, 2, 2, 1, 1]
Solution: The convolution table, with h arranged vertically and x horizontally, is

Folding the table, we get

y(n) = [1, 3, 3, 5, 3, 7, 4, 3, 3, 0, 1]
Note that there are Ly = Lx +Lh− 1 = 8 + 3 = 11 output samples.
V. LTI Form:
Consider finite length of the impulse response h(n)= [h0, h1, h2, h3] and the input signal
x(n) = [x0, x1, x2, x3, x4]

The input signal can be written as a linear combination of delayed impulses:

x(n) = x0[1, 0, 0, 0, 0]
+ x1[0, 1, 0, 0, 0]
+ x2[0, 0, 1, 0, 0]
+ x3[0, 0, 0, 1, 0]
+ x4[0, 0, 0, 0, 1]
It can also be written analytically for all n as a sum of delta functions:
x(n) = x0δ(n) + x1δ(n − 1) + x2δ(n − 2) + x3δ(n − 3) + x4δ(n − 4)
The effect of the filter is to replace each delayed impulse by the corresponding delayed
impulse response, that is,
y(n)= x0h(n) + x1h(n − 1) + x2h(n − 2) + x3h(n − 3) + x4h(n − 4)
We can represent the input and output signals as blocks:

The indicated linear combinations in the right-hand side give:

 y (n) = [h0x0, x0h1 + x1h0,x0h2 + x1h1 + x2h0, . . . , x4h3]
= [y0, y1, y2, . . . , y7]
For computational purposes, the LTI form can be represented pictorially in a table form,
as shown in Fig. below. The impulse response h(n) is written horizontally, and the input
x(n)vertically.

▪ The rows of the table correspond to the successive delays (right shifts) of the h(n) sequence-the mth row
corresponds to delay by m units.
▪ Each row is scaled by the corresponding input sample, that is, the mth row represents the term xmhn−min the
LTI form. After the table is filled, the table entries are summed column-wise to obtain the output samples
y(n), which is equivalent to forming the sum:
Example: Calculate the convolution of the following impulse response and input signals
using the LTI form:
h(n)= [1, 2,−1, 1], x(n)= [1, 1, 2, 1, 2, 2, 1, 1]
Solution: The corresponding LTI table is in this case:

Solution; The output samples are obtained by summing the entries in each column.

y(n) = [1, 3, 3, 5, 3, 7, 4, 3, 3, 0, 1]
Vi. Matrix Form:
The convolution can also be written in the linear matrix form:
y = Hx
where H is built out of the filter’s impulse response h. Because the output vector y has
length L+M and the input vector x length L, the filter matrix H must be rectangular with
dimensions Ly × Lx = (L + M)×L
The output samples can be arranged in the matrix form
Note that the columns of H are the successively delayed replicas of the impulse response
vector h. There are as many columns as input samples. Note also that H is a so-called
Toeplitz matrix, in the sense that it has the same entry along each diagonal. The Toeplitz
property is a direct consequence of the time invariance of the impulse response.
Example: Calculate the convolution of the following impulse response and input signals
using the matrix form:
h(n)= [1, 2,−1, 1], x(n)= [1, 1, 2, 1, 2, 2, 1, 1]
Solution: Because Ly = 11 and Lx = 8, the filter matrix will be 11 × 8 dimensional. We have,
There is also an alternative matrix form written as follows:

❖ Matrix representations of convolution are very useful in some applications, such as image
processing, and in more advanced DSP methods such as parametric spectrum estimation
and adaptive filtering.
Linear constant-coefficient Difference Equation
• Subclass of LTI systems consists of systems for which the input x[n] and y[n]
satisfy an N th order linear constant-coefficient difference equation:
The general form of a linear constant-coefficient difference equation is:
σN 𝑎
k=0 𝑘 𝑦[𝑛 − 𝐾] = σ𝑀
𝑚=0 𝑏𝑚 𝑥[𝑛 − 𝑚]
• Example: Difference equation representation of the accumulation
y[n] = σnk=−∞ 𝑥 k
The output for n-1:
y[n-1] = σn−1
k=−∞ 𝑥 k
y[n] = x[n] + σn−1
k=−∞ 𝑥 k
y[n] = x[n] + y[n-1]
y[n] – y[n-1] =x[n]
 Continued …
✓ Satisfy a linear constant-coefficient difference equation of the form
with N = 1, 𝑎0 = 1, 𝑎1 = −1,M =0 and 𝑏0 = 1
Solving Linear constant-coefficient Difference Equation
• The procedure for computing the solution of constant-coefficient difference
equation is very similar to that employed for solving the constant-coefficient
differential equation in the case of an LTI continuous-time system.
• The output response y[n] consists of two components which are computed
independently and then added to yield the total solution.
y[n] = 𝑦𝑝 𝑛 + 𝑦ℎ 𝑛
𝑦ℎ 𝑛 :homogeneous solution with x[n] =0, i.e. solution to the equation
 Solving Linear Constant coefficient Difference Equation

σNk=0 𝑎𝑘 𝑦ℎ 𝑛 − 𝑘 = 0
• 𝑦𝑝 𝑛 : Particular solution resulting from the specified into x[n], often
called forcing function
• First compute the homogeneous solution 𝑦ℎ 𝑛 , assume it is of the form:
yh 𝑛 = λ𝑛
Substituting σ𝑁 𝑘=0 𝑘 ℎ𝑎 𝑦 [n-k] = σ 𝑁
𝑎
𝑘=0 𝑘 λ (𝑛−𝑘)
=0
=λ(𝑛−𝑁) (𝑎0 λ𝑁 + 𝑎1 λ𝑁−1 + . . . +𝑎𝑁−1 λ + 𝑎𝑁 ) = 0
✓Characteristic polynomial of the discrete-time system
✓Let λ1 , λ2 , λ3 , . . . , λ𝑁 denote its N roots
 Continued …
✓If these roots are all distinct, then general form of the homogenous solution is:
yh 𝑛 = α1 λ1 𝑛 + α2 λ2 𝑛 + . . . + α𝑁 λ𝑁 𝑛
Where α1 , α2 , . . . , α𝑁 are constants determined from the specified initial
conditions of the discrete-time system.
✓ The determination of the particular solution yp 𝑛 is of the form as the
specified input x[n], if x[n] has the form λ0 𝑛
(λ0 ≠ λi , i = 1, 2, . . . , N) for all N.
✓If x[n] is constant, yp 𝑛 is also assumed constant.
✓If x[n] is sinusoidal sequence, then yp 𝑛 is also assumed to be a sinusoidal
sequence and so on.
 Continued …

Terms in x[n] Particular Solution

C 𝐶1
Cn 𝐶1 n +𝐶2
C𝑎𝑛 𝐶1 𝑎𝑛
Ccos(𝑛ω0 ) 𝐶1 cos(𝑛ω0 ) + 𝐶2 sin(𝑛ω0 )
Csin(𝑛ω0 ) 𝐶1 cos(𝑛ω0 ) + 𝐶2 sin(𝑛ω0 )
C𝑎𝑛 cos(𝑛ω0 ) 𝐶1 𝑎𝑛 cos(𝑛ω0 ) + 𝐶2
𝑎𝑛 sin(𝑛ω0 )
 Continued …
• The complete solution:
1. Find the form of the homogeneous solution yh 𝑛 from the roots of the
characteristic equation.
2. Find a particular solution yp 𝑛 by assuming that it is of the same form as
the input yet independent of all the terms in homogeneous solution.
3. Determine the coefficients in the homogeneous solution so that the complete
solution.
y[n] = yh 𝑛 + yp 𝑛 satisfies the initial condition.
The constants A1 and A2 must now be found so that the total solution satisfies the given initial
conditions, y(-1) = 1 and y(-2) = 0. Because the solution given in the above only applies for n  0,
we must derive an equivalent set of initial conditions for y(0) and y(1). Evaluating the equ. at n = 0
and n = 1. we have
✓ Although we have focused thus far on linear difference equations with constant
coefficients, not all systems and not all difference equations of interest are linear, and
not all have constant coefficients.
✓ A system that computes a running average of a signal x(n) over the interval [0, n], for
example, is defined by

This system may be represented by a difference equation that has time-varying coefficients:

✓ Although more complicated and difficult to solve, nonlinear difference equations or

difference equations with time-varying coefficients are important and arise frequently
in many applications.