Basics of Digital Signal Processing

Class Notes
Dr. Kaustubh Bhattacharyya
Signal is any physical phenomenon that varies with respect to time or any spatial
coordinate. Mathematically signal is a function of one or more independent variable.
A discrete time signal 𝒙[𝒏] is defined only for integer values of 𝒏 and is undefined for non-
integer values of 𝒏.
An infinite-duration discrete time signal x[n] can be written as a sequence of numbers
inside braces.
𝒙[𝒏] = {… … … . , 𝒙[−𝟑], 𝒙[−𝟐], 𝒙[−𝟏], 𝒙[𝟎], 𝒙[𝟏], 𝒙[𝟐], 𝒙[𝟑], 𝒙[𝟒], … … … }

Where the arrow denotes the value of the signal 𝒙[𝒏] at 𝒏 = 𝟎. The sample values to the
right of the arrow are for positive values of n, and the sample values to its left are for the
negative values of 𝒏.
A system is defined as a physical device that performs an operation on signal.
System is characterized by the type of operation that performs on the signal. Such
operations are referred to as signal processing.
Advantages of DSP
1. A digital programmable system allows flexibility in reconfiguring the digital signal processing
operation by changing the program. In analog redesign of hardware is required
2. In digital accuracy depends on word length, floating Vs fixed point arithmetic etc. in analog it
depends on components.
3. Can be stored on disk
4. It is very difficult to perform precise mathematical operations on signal in analog form but these
operations can be routinely implemented on a digital computer using software.
5. Cheaper to implement.
6. Small size
7. Several filters need several boards in analog, whereas in digital same DSP processor is used for many
filters.
Disadvantages of DSP
1. When analog signal is changing very fast, it is difficult to convert digital form.
2. Finite word length problem.
3. When the signal is weak, within a few tenths of millivolts, we can not amplify the signal after it
digitized.
4. DSP hardware is more expensive than general purpose microprocessor & microcontrollers.
5. Dedicated DSP can do better than general purpose DSP
Applications of DSP
1. Filtering
 2. Speech synthesis
3. Speech compression and expansion
4. Speech recognition
5. Signal analysis
6. Image processing
7. PCM used in telephone communication
8. High speed NODEM
9. Wave form generation
Classification of Discrete time Signal:
1. Based on number of independent variables: 1-Dimensional and Multidimensional signal
2. Finite length and Infinite length signals
A discrete-time signal may be finite-length or an infinite-length. A finite-length (also called finite -
duration) signal is defined only for a finite time interval:
𝑁1 ≤ 𝑛 ≤ 𝑁2
Where, −∞ < 𝑁1 and 𝑁2 < ∞ and 𝑁2 > 𝑁1 . The length or duration 𝑁 of the finite length signal is
𝑁 = 𝑁2 − 𝑁1 + 1
A length-𝑁 discrete time signal consists of 𝑁 samples and is often referred to as an 𝑁-point signal.
For example, a finite-duration signal can be represented as
𝑥[𝑛] = {3, −2, 5, 4, 6, 2, −6, 5, 8, 2, 7} −4≤𝑛≤6

Here 𝑁1 = −4 and 𝑁2 = 6, so 𝑁 = 𝑁2 − 𝑁1 + 1 = 11. Thus, the given signal consists of 11 samples

and is referred to as 11-point signal.
3. Right sided, causal, left sided, non-causal, both sided and anticausal signal
Right sided: An infinite length signal 𝑥[𝑛] is called a right sided signal if it is zero for 𝑛 < 𝑁1 , i.e,
𝑥[𝑛] = 0 𝑓𝑜𝑟 𝑛 < 𝑁1
Where, 𝑁1 is positive or negative integer.
Causal: An infinite length sequence x[n] is called causal sequence if it is zero for 𝑛 < 0 i.e,
𝑥[𝑛] = 0 𝑓𝑜𝑟 𝑛 < 0
Left sided: An infinite length signal x[n] is called a left sided signal if it is zero for 𝑛 > 𝑁2 , i.e,
𝑥[𝑛] = 0 𝑓𝑜𝑟 𝑛 > 𝑁2
Where, 𝑁2 is positive or negative integer.
Anticausal: An infinite length signal x[n] is said to be anticuasal signal if it is zero for 𝑛 ≥ 0, i.e,
𝑥[𝑛] = 0 𝑓𝑜𝑟 𝑛 ≥ 0
Both Sided: An infinite length signal 𝑥[𝑛] is called a both sided signal if it is non zero for 𝑛 >
𝑁 𝑎𝑛𝑑 𝑛 < 𝑁.
Non-Causal: An infinite length sequence x[n] is called non-causal sequence if it is zero for |𝑛| > 𝑁
i.e,
𝑥[𝑛] = 0 𝑓𝑜𝑟 |𝑛| > 𝑁
Where N is any integer.
4. Based on the Symmetry: Odd signal, even signal and general signal
A real-valued discrete time signal 𝑥[𝑛] is referred to as an even signal if it is identical with its
reflection about the origin, that is,
𝑥[𝑛] = 𝑥[−𝑛], 𝑓𝑜𝑟 𝑎𝑙𝑙 𝑛
An even signal has the same value at the instants 𝑛 and −𝑛 for all values of n. clearly an even signal
is symmetrical about the vertical axis.
A real valued discrete time signal 𝑥[𝑛] is said to be an odd signal if
𝑥[𝑛] = −𝑥[−𝑛], 𝑓𝑜𝑟 𝑎𝑙𝑙 𝑛
The value of an odd signal at the instant 𝑛 is negative of its value at the instants −𝑛. Therefore, an
odd signal is antisymmetric about the vertical axis. An odd signal must necessarily be 0 at 𝑛 = 0.
A general signal 𝑥[𝑛] can be expressed as sum of even and odd components.
𝑥[𝑛] = 𝑥𝑒 [𝑛] + 𝑥𝑜 [𝑛]
Where 𝑥𝑒 [𝑛] is even component and 𝑥𝑜 [𝑛] is odd component of the signal 𝑥[𝑛].
𝑥[𝑛] + 𝑥[−𝑛]
𝑥𝑒 [𝑛] =
2
𝑥[𝑛] − 𝑥[−𝑛]
𝑥𝑜 [𝑛] =
2
Properties of discrete time even and odd signals.
1. If 𝑥1 [𝑛] is an odd signal [𝑥1 [𝑛] = −𝑥1 [−𝑛]] and 𝑥2 [𝑛] is even signal [𝑥2 [𝑛] = 𝑥2 [−𝑛]], then
a. 𝑥3 [𝑛] = 𝑥1 [𝑛]𝑥2 [𝑛] is an odd signal, that is,
𝑒𝑣𝑒𝑛 𝑠𝑖𝑔𝑛𝑎𝑙 × 𝑜𝑑𝑑 𝑠𝑖𝑔𝑛𝑎𝑙 = 𝑜𝑑𝑑 𝑠𝑖𝑔𝑛𝑎𝑙
Proof: consider signal 𝑥3 [𝑛]
𝑥3 [𝑛] = 𝑥1 [𝑛]𝑥2 [𝑛]
𝑥3 [−𝑛] = 𝑥1 [−𝑛]𝑥2 [−𝑛] = −𝑥1 [𝑛]𝑥2 [𝑛] = −𝑥3 [𝑛]
b. 𝑥4 [𝑛] = 𝑥1 [𝑛]𝑥1 [𝑛] is an even signal, that is,
𝑜𝑑𝑑 𝑠𝑖𝑔𝑛𝑎𝑙 × 𝑜𝑑𝑑 𝑠𝑖𝑔𝑛𝑎𝑙 = 𝑒𝑣𝑒𝑛 𝑠𝑖𝑔𝑛𝑎𝑙
Proof: consider signal 𝑥4 [𝑛]
𝑥4 [𝑛] = 𝑥1 [𝑛]𝑥1 [𝑛]
𝑥4 [−𝑛] = 𝑥1 [−𝑛]𝑥1 [−𝑛] = (−𝑥1 [𝑛])(−𝑥1 [𝑛]) = 𝑥1 [𝑛]𝑥1 [𝑛] = 𝑥4 [𝑛]
c. 𝑥5 [𝑛] = 𝑥2 [𝑛]𝑥2 [𝑛] is an even signal, that is,
𝑒𝑣𝑒𝑛 𝑠𝑖𝑔𝑛𝑎𝑙 × 𝑒𝑣𝑒𝑛 𝑠𝑖𝑔𝑛𝑎𝑙 = 𝑒𝑣𝑒𝑛 𝑠𝑖𝑔𝑛𝑎𝑙
Proof: consider signal 𝑥5 [𝑛]
𝑥5 [𝑛] = 𝑥2 [𝑛]𝑥2 [𝑛]
𝑥5 [−𝑛] = 𝑥2 [−𝑛]𝑥2 [−𝑛] = (𝑥2 [𝑛])(𝑥2 [𝑛]) = 𝑥5 [𝑛]
2. If 𝑥[𝑛] is an even signal, then
∞ ∞

∑ 𝑥[𝑛] = 𝑥[0] + 2 ∑ 𝑥[𝑛]

𝑛=−∞ 𝑛=1
Proof:
∞ −1 ∞ ∞ ∞ ∞ ∞

∑ 𝑥[𝑛] = ∑ 𝑥[𝑛] + ∑ 𝑥[𝑛] = ∑ 𝑥[−𝑛] + ∑ 𝑥[𝑛] = ∑ 𝑥[𝑛] + 𝑥[0] + ∑ 𝑥[𝑛]

𝑛=−∞ 𝑛=−∞ 𝑛=0 𝑛=1 𝑛=0 𝑛=1 𝑛=1
∞

= 𝑥[0] + 2 ∑ 𝑥[𝑛] ; ∵ 𝑥[𝑛] 𝑖𝑠 𝑒𝑣𝑒𝑛, 𝑥[𝑛] = 𝑥[−𝑛]

𝑛=1
 3. If 𝑥[𝑛] is an odd signal, then
∞

∑ 𝑥[𝑛] = 0
𝑛=−∞
Proof:
∞ −1 ∞

∑ 𝑥[𝑛] = ∑ 𝑥[𝑛] + ∑ 𝑥[𝑛]

𝑛=−∞ 𝑛=−∞ 𝑛=0
∞ ∞ ∞ ∞

= ∑ 𝑥[−𝑛] + ∑ 𝑥[𝑛] = − ∑ 𝑥[𝑛] + 𝑥[0] + ∑ 𝑥[𝑛] = 𝑥[0] = 0

𝑛=1 𝑛=0 𝑛=1 𝑛=1
∵ 𝑥[𝑛] 𝑖𝑠 𝑜𝑑𝑑, 𝑥[𝑛] = −𝑥[−𝑛] ; 𝑥[0] = 0
4. If x[n] is a general signal with even and odd parts denoted by 𝑥𝑒 [𝑛] 𝑎𝑛𝑑 𝑥𝑜 [𝑛], then
∞ ∞ ∞

∑ 𝑥 2 [𝑛] = ∑ 𝑥𝑒2 [𝑛] + ∑ 𝑥𝑜2 [𝑛]

𝑛=−∞ 𝑛=−∞ 𝑛=−∞
Proof:
∞ ∞
2 [𝑛] 2
∑ 𝑥 = ∑ [ 𝑥𝑒 [𝑛] + 𝑥𝑜 [𝑛]]
𝑛=−∞ 𝑛=−∞
∞ ∞ ∞

= ∑ 𝑥𝑒2 [𝑛] + ∑ 𝑥𝑜2 [𝑛] + 2 ∑ 𝑥𝑒 [𝑛] 𝑥𝑜 [𝑛]

𝑛=−∞ 𝑛=−∞ 𝑛=−∞
∞ ∞ −1 ∞

= ∑ 𝑥𝑒2 [𝑛] + ∑ 𝑥𝑜2 [𝑛] + 2 [ ∑ 𝑥𝑒 [𝑛] 𝑥𝑜 [𝑛] + ∑ 𝑥𝑒 [𝑛] 𝑥𝑜 [𝑛]]

𝑛=−∞ 𝑛=−∞ 𝑛=−∞ 𝑛=0
∞ ∞ ∞ ∞

= ∑ 𝑥𝑒2 [𝑛] + ∑ 𝑥𝑜2 [𝑛] + 2 [∑ 𝑥𝑒 [−𝑛] 𝑥𝑜 [−𝑛] + ∑ 𝑥𝑒 [𝑛] 𝑥𝑜 [𝑛]]

𝑛=−∞ 𝑛=−∞ 𝑛=1 𝑛=0
∵ 𝑥𝑒 [−𝑛] = 𝑥𝑒 [𝑛], 𝑥𝑜 [−𝑛] = −𝑥𝑜 [𝑛] 𝑎𝑛𝑑 𝑥[0] = 0, we have
∞ ∞ ∞
2 [𝑛]
∑ 𝑥 = ∑ 𝑥𝑒2 [𝑛] + ∑ 𝑥𝑜2 [𝑛]
𝑛=−∞ 𝑛=−∞ 𝑛=−∞

5. Bounded, absolutely summable, and square summable signals:

A discrete time signal 𝑥[𝑛] is said to be bounded if each of its samples is of magnitude less than or
equal to a finite positive number say 𝐵𝑥 , that is,
|𝑥[𝑛]| ≤ 𝐵𝑥 < ∞
A signal 𝑥[𝑛] is said to be absolutely summable if
∞

∑ |𝑥[𝑛]| < ∞
𝑛=−∞
A signal 𝑥[𝑛] is said to be absolutely square-summable if
∞

∑ |𝑥[𝑛]|2 < ∞
𝑛=−∞
 6. Energy signal and power signal
The energy of a discrete time signal 𝑥[𝑛] is given by
∞

𝐸𝑥 = ∑ |𝑥[𝑛]|2
𝑛=−∞
The average power of a discrete time signal 𝑥[𝑛] is given by
𝑁
1
𝑃𝑥 = lim ∑ |𝑥[𝑛]|2
𝑁→∞ 2𝑁 + 1
𝑛=−𝑁
If 0 < 𝐸𝑥 < ∞, that is, if Energy 𝐸𝑥 of the signal 𝑥[𝑛] is finite and the power 𝑃𝑥 = 0, then the signal
𝑥[𝑛] is referred to as an energy signal. A necessary condition for the energy to be finite is that the
signal amplitude must → 0 as |𝑛| → ∞. All the finite length signals are energy signal.
If 0 < 𝑃𝑥 < ∞, that is, if power 𝑃𝑥 of the signal 𝑥[𝑛] is finite and the power 𝐸𝑥 = ∞, then the signal
𝑥[𝑛] is referred to as a power signal.
A signal cannot both be an energy signal and power signal. On the other hand, there are signals that
are neither energy nor power signals. All practical signals have finite energies and are therefore
energy signals. It is impossible to generate a true power signal in practice because such signal has
infinite duration and infinite energy. All finite periodic signals are power signals; however not all
power signals are periodic, for example, unit step signal.
7. Periodic and aperiodic signal
A discrete time signal x[n] is periodic if and only if
𝑥[𝑛] = 𝑥[𝑛 + 𝑘𝑁]; 𝑓𝑜𝑟 𝑎𝑙𝑙 𝑛
Where 𝑘 is an integer and 𝑁 is a positive integer and is the period of the signal measured in terms of
number of samples spacings. The smallest value of 𝑁 is called the fundamental period of the
periodic signal. If there is no value of 𝑁 that satisfy the above equation, then the signal is called
aperiodic or nonperiodic signal.
Properties of periodic signal:
1. A periodic signal is an everlasting signal
2. The sum of samples of 𝑥[𝑛] over any interval 𝑁 is the same, that is, for any real integers 𝑝 and 𝑞,
𝑝+𝑁 𝑞+𝑁

∑ 𝑥[𝑛] = ∑ 𝑥[𝑛]
𝑛=𝑝 𝑛=𝑞
This results from the fact that a periodic signal takes the same values at the intervals of 𝑁.
3. The sum of 𝑀 periodic signal is always periodic with period 𝑁 which equals the LCM of their
individual periods, that is,
𝑁 = 𝐿𝐶𝑀(𝑁1 , 𝑁2 , 𝑁3 , … … … . . , 𝑁𝑀 )
Where, 𝑁1 , 𝑁2 , 𝑁3 , … … … . . , 𝑁𝑀 are the periods of individual periodic signals.
Discrete Time Unitary Signals:
1. Unit Step Signal:
1; 𝑛 ≥ 0
𝑢[𝑛] = {
0; 𝑛 < 0
 𝑢[𝑛] = {1, 1, 1, 1 … … … … . . }

𝑦[𝑛] = 𝑥[𝑛]𝑢[𝑛]
Where 𝑥[𝑛] = 0; 𝑓𝑜𝑟 𝑎𝑙𝑙 𝑛 < 0

2. Unit Impulse:
1; 𝑛 = 0
𝛿[𝑛] = {
0; 𝑛 ≠ 0

𝛿[𝑛] = {… … . .0, 0, 0, 1, 0, 0, 0 … … … … . . }

Properties of Unit Impulse:

1. Time Scalling
𝛿[𝑘𝑛] = 𝛿[𝑛]; 𝑘 𝑖𝑠 𝑎𝑛𝑦 𝑖𝑛𝑡𝑒𝑔𝑒𝑟

1; 𝑘𝑛 = 0 − −→ 𝑛 = 0 1; 𝑛 = 0
𝛿[𝑘𝑛] = { = = 𝛿[𝑛]
0; 𝑘𝑛 ≠ 0 − −→ 𝑛 ≠ 0 0; 𝑛 ≠ 0

2. Relationship between 𝜹[𝒏] and 𝒖[𝒏]

𝛿[𝑛]

n=0
u[n]

………………

n=0 n=1 n=2 n=3 n=4 n=5

u[n-1]

n=0
𝜹[𝒏] = 𝒖[𝒏] − 𝒖[𝒏 − 𝟏]
𝒖[𝒏] = 𝜹[𝒏] + 𝜹[𝒏 − 𝟏] + 𝜹[𝒏 − 𝟐] + ⋯
∝

𝒖[𝒏] = ∑ 𝜹[𝒏 − 𝒌]
𝒌=𝟎

𝜹[𝒏] 𝜹[𝒏 − 𝟑]=1; n=3

n=0

3. Multiplication Property:
𝒙[𝒏] = 𝒙[𝒏]𝜹[𝒏 − 𝒌] = 𝒙[𝒌]𝜹[𝒏 − 𝒌]
𝟏; 𝒏 = 𝒌
𝜹[𝒏 − 𝒌] = {
𝟎; 𝒏 ≠ 𝒌
4. Sum Property
∞

∑ 𝜹[𝒏] = 𝟏
𝒏=−∞
∞ −𝟏 ∞

∑ 𝜹[𝒏] = ∑ 𝜹[𝒏] + 𝜹[𝟎] + ∑ 𝜹[𝒏] = 𝟎 + 𝟏 + 𝟎 = 𝟏

𝒏=−∞ 𝒏=−∞ 𝒏=𝟏
5. Shifting Property
 ∞

𝒙[𝒏] = ∑ 𝒙[𝒌]𝜹[𝒏 − 𝒌]
𝒌=−∞
3. Unit Ramp Signal
𝒏 ;𝒏 ≥ 𝟎
𝒓[𝒏] = {
𝟎 ;𝒏 < 𝟎
𝒓[𝒏] = {𝟎, 𝟏, 𝟐, 𝟑, 𝟒, 𝟓, … … … … }
𝑟[𝑛] = 0. 𝛿[𝑛] + 1. 𝛿[𝑛 − 1] + 2. 𝛿[𝑛 − 2] + 3. 𝛿[𝑛 − 3] + ⋯.
∞

𝒓[𝒏] = ∑ 𝒌𝜹[𝒏 − 𝒌]
𝒌=𝟎
𝑟[𝑛] = 𝑛. 𝑢[𝑛]
4. Discrete time real exponential signal
5. Discrete time complex exponential signal
6. Discrete time sinusoidal signal
7.
Discrete Time Signal Operation
1. Transformation of Dependent Variable (Amplitude)
a. Signal Addition
𝒚[𝒏] = 𝒙[𝒏] + 𝒘[𝒏]
𝒙[𝒏] = {𝟏, 𝟓, 𝟕, 𝟑, −𝟔, 𝟒, 𝟖}

𝒏𝒙 = −𝟐, −𝟏, 𝟎, 𝟏, 𝟐, 𝟑, 𝟒
𝒏𝒙𝟏 = −𝟐, 𝒏𝒙𝟐 = 𝟒
𝒘[𝒏] = {𝟖, −𝟓, 𝟐, 𝟒, −𝟑, 𝟒, 𝟕}

𝒏𝒘 = −𝟏, 𝟎, 𝟏, 𝟐, 𝟑, 𝟒, 𝟓
𝒏𝒘𝟏 = −𝟏, 𝒏𝒘𝟐 = 𝟓
𝒏𝒚𝟏 = 𝒎𝒊𝒏(𝒏𝒙𝟏 , 𝒏𝒘𝟏 ), 𝒏𝒚𝟐 = 𝒎𝒂𝒙(𝒏𝒙𝟐 , 𝒏𝒘𝟐 ),
𝒏𝒚𝟏 = −𝟐, 𝒏𝒚𝟐 = 𝟓
𝒏𝒚 = 𝒏𝒚𝟏 : 𝒏𝒚𝟐
𝒏𝒚 : − 𝟐, −𝟏, 𝟎, 𝟏, 𝟐, 𝟑 𝟒, 𝟓
𝒚[𝒏] = {𝟏, 𝟏𝟑, 𝟐, 𝟓, −𝟐, 𝟏, 𝟏𝟐, 𝟕}

b. Scalar Addition
𝒚[𝒏] = 𝒂 + 𝒙[𝒏]
𝒙[𝒏] = {𝟏, 𝟓, 𝟕, 𝟑, −𝟔, 𝟒, 𝟖}

𝒏𝒙 = −𝟐, −𝟏, 𝟎, 𝟏, 𝟐, 𝟑, 𝟒
𝒂=𝟓
𝒚[𝒏] = {𝟔, 𝟏𝟎, 𝟏𝟐, 𝟖, −𝟏, 𝟗, 𝟏𝟑}
 𝒏𝒚 = −𝟐, −𝟏, 𝟎, 𝟏, 𝟐, 𝟑, 𝟒
c. Signal Multiplication
𝒚[𝒏] = 𝒙[𝒏] ∗ 𝒘[𝒏]
𝒙[𝒏] = {𝟏, 𝟓, 𝟕, 𝟑, −𝟔, 𝟒, 𝟖}

𝒏𝒙 = −𝟐, −𝟏, 𝟎, 𝟏, 𝟐, 𝟑, 𝟒
𝒏𝒙𝟏 = −𝟐, 𝒏𝒙𝟐 = 𝟒
𝒘[𝒏] = {𝟖, −𝟓, 𝟐, 𝟒, −𝟑, 𝟒, 𝟕}

𝒏𝒘 = −𝟏, 𝟎, 𝟏, 𝟐, 𝟑, 𝟒, 𝟓
𝒏𝒘𝟏 = −𝟏, 𝒏𝒘𝟐 = 𝟓
𝒏𝒚𝟏 = 𝒎𝒊𝒏(𝒏𝒙𝟏 , 𝒏𝒘𝟏 ), 𝒏𝒚𝟐 = 𝒎𝒂𝒙(𝒏𝒙𝟐 , 𝒏𝒘𝟐 ),
𝒏𝒚𝟏 = −𝟐, 𝒏𝒚𝟐 = 𝟓
𝒏𝒚 = 𝒏𝒚𝟏 : 𝒏𝒚𝟐
𝒏𝒚 : − 𝟐, −𝟏, 𝟎, 𝟏, 𝟐, 𝟑 𝟒, 𝟓
𝒚[𝒏] = {𝟎, 𝟒𝟎, −𝟑𝟓, 𝟔, −𝟐𝟒, −𝟏𝟐, 𝟑𝟐, 𝟎}
d. Scalar Multiplication
𝒚[𝒏] = 𝒂 ∗ 𝒙[𝒏]
𝒙[𝒏] = {𝟏, 𝟓, 𝟕, 𝟑, −𝟔, 𝟒, 𝟖}

𝒏𝒙 = −𝟐, −𝟏, 𝟎, 𝟏, 𝟐, 𝟑, 𝟒
𝒂=𝟓
𝒚[𝒏] = {𝟓, 𝟐𝟓, 𝟑𝟓, 𝟏𝟓, −𝟑𝟎, 𝟐𝟎, 𝟒𝟎}

𝒏𝒚 = −𝟐, −𝟏, 𝟎, 𝟏, 𝟐, 𝟑, 𝟒
2. Transformation of Independent Variable (Time)
a. Time-Shifting
𝒚[𝒏] = 𝒙[𝒏 − 𝒌]
i. Delay (Right Shifted): if ‘k’ is +ve
Then, 𝒚[𝒏] = 𝒙[𝒏 − 𝒌]
𝒙[𝒏] = {𝟏, 𝟓, 𝟕, 𝟑, −𝟔, 𝟒, 𝟖}

𝒏𝒙 = −𝟐, −𝟏, 𝟎, 𝟏, 𝟐, 𝟑, 𝟒
𝒌=𝟑
𝒏𝒙𝟏 = −𝟐, 𝒏𝒙𝟐 = 𝟒
𝒚[𝒏] = 𝒙[𝒏 − 𝟑]

𝒚[𝒏] = {𝟎, 𝟎, 𝟎, 𝟏, 𝟓, 𝟕, 𝟑, −𝟔, 𝟒, 𝟖}

𝒏𝒚𝟏 = 𝒏𝒙𝟏 = −𝟐; 𝒏𝒚𝟐 = 𝒏𝒙𝟐 + 𝒌 = 𝟕

𝒏𝒚 = 𝒏𝒚𝟏 : 𝒏𝒚𝟐
𝒏𝒚 = −𝟐, −𝟏, 𝟎, 𝟏, 𝟐, 𝟑, 𝟒, 𝟓, 𝟔, 𝟕
 ii. Advance (Left Shifted): if ‘k’ is -ve
Then, 𝒚[𝒏] = 𝒙[𝒏 + 𝒌]
𝒙[𝒏] = {𝟏, 𝟓, 𝟕, 𝟑, −𝟔, 𝟒, 𝟖}

𝒏𝒙 = −𝟐, −𝟏, 𝟎, 𝟏, 𝟐, 𝟑, 𝟒
𝒏𝒙𝟏 = −𝟐, 𝒏𝒙𝟐 = 𝟒
𝒌 = −𝟑
𝒚[𝒏] = 𝒙[𝒏 + 𝟑]
𝒚[𝒏] = {𝟏, 𝟓, 𝟕, 𝟑, −𝟔, 𝟒, 𝟖, 𝟎, 𝟎, 𝟎 }

𝒏𝒚𝟏 = 𝒏𝒙𝟏 + 𝒌 = −𝟓; 𝒏𝒚𝟐 = 𝒏𝒙𝟐 = 𝟒

𝒏𝒚 = 𝒏𝒚𝟏 : 𝒏𝒚𝟐
𝒏𝒚 = −𝟓, −𝟒, −𝟑, −𝟐, −𝟏, 𝟎, 𝟏, 𝟐, 𝟑, 𝟒
b. Time Scaling
i. Down sampling/decimation/Compression
𝒚[𝒏] = 𝒙[𝒌𝒏]
𝒙[𝒏] = {−𝟒, 𝟏, 𝟓, 𝟕, 𝟑, −𝟔, 𝟒, 𝟖, 𝟗}

𝒏𝒙 = −𝟑, −𝟐, −𝟏, 𝟎, 𝟏, 𝟐, 𝟑, 𝟒, 5

𝒏𝒙𝟏 = −𝟑, 𝒏𝒙𝟐 = 𝟓
𝒌=𝟐
𝒚[𝒏] = 𝒙[𝟐𝒏]
Remove k-1 samples in between
𝒚[𝒏] = {𝟏, 𝟕, −𝟔, 𝟖}

𝒏𝒙𝟏 𝒏𝒙𝟐
𝒏𝒚𝟏 = +𝒗𝒆 𝒓𝒐𝒖𝒏𝒅 ( ) = −𝟏; 𝒏𝒚𝟐 = −𝒗𝒆 𝒓𝒐𝒖𝒏𝒅 ( )=𝟐
𝒌 𝒌
𝒏𝒚 = 𝒏𝒚𝟏 : 𝒏𝒚𝟐
𝒏𝒚 = −𝟏, 𝟎, 𝟏, 𝟐
ii. p sampling/Interpolation/Expansion
𝒏
𝒚[𝒏] = 𝒙 [ ]
𝒌
𝒙[𝒏] = {−𝟒, 𝟏, 𝟓, 𝟕, 𝟑, −𝟔, 𝟒, 𝟖, 𝟗}

𝒏𝒙 = −𝟑, −𝟐, −𝟏, 𝟎, 𝟏, 𝟐, 𝟑, 𝟒, 5

𝒏𝒙𝟏 = −𝟑, 𝒏𝒙𝟐 = 𝟓
𝟏
𝒌 = 𝟐; = 𝟎. 𝟓
𝒌
𝒏
𝒚[𝒏] = 𝒙[ ]
𝟐
Input k-1 samples in between
𝒚[𝒏] = {−𝟒, 𝟎, 𝟏, 𝟎, 𝟓, 𝟎, 𝟕, 𝟎, 𝟑, 𝟎, −𝟔, 𝟎, 𝟒, 𝟎, 𝟖, 𝟎, 𝟗 }
 𝒏𝒚𝟏 = 𝒌 ∗ 𝒏𝒙𝟏 = −𝟔; 𝒏𝒚𝟐 = 𝒌 ∗ 𝒏𝒙𝟐 = 𝟏𝟎
𝒏𝒚 = 𝒏𝒚𝟏 : 𝒏𝒚𝟐
𝒏𝒚 = −𝟔, −𝟓, −𝟒, −𝟑, −𝟐, −𝟏, 𝟎, 𝟏, 𝟐, 𝟑, 𝟒, 𝟓, 𝟔, 𝟕, 𝟖, 𝟗, 𝟏𝟎
c. Time Reversal/Folding
𝒙[𝒏] = {−𝟒, 𝟏, 𝟓, 𝟕, 𝟑, −𝟔, 𝟒, 𝟖, 𝟗}

𝒏𝒙 = −𝟑, −𝟐, −𝟏, 𝟎, 𝟏, 𝟐, 𝟑, 𝟒, 5

𝒏𝒙𝟏 = −𝟑, 𝒏𝒙𝟐 = 𝒏𝒙𝟏 + 𝑵𝒖𝒎𝒃𝒆𝒓 𝒐𝒇 𝑬𝒍𝒆𝒎𝒆𝒏𝒕 − 𝟏 = 𝟓
𝒚[𝒏] = 𝒙[−𝒏] = {𝟗, 𝟖, 𝟒, −𝟔, 𝟑, 𝟕, 𝟓, 𝟏, −𝟒}

𝒏𝒚𝟏 = −𝒏𝒙𝟐 = −𝟓; 𝒏𝒚𝟐 = −𝒏𝒙𝟏 = 𝟑

𝒏𝒚 = 𝒏𝒚𝟏 : 𝒏𝒚𝟐
𝒏𝒚 = −𝟓, −𝟒, −𝟑, −𝟐, −𝟏, 𝟎, 𝟏, 𝟐, 𝟑

Combined Operation:
𝒏
𝒚[𝒏] = 𝒙 [− − 𝒌]
𝒎

Time Shifting-----→ Time Scaling ----→ Time Reversal

⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗
𝒏 𝒏 𝒏
𝒙[𝒏] ⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗
𝑻𝒊𝒎𝒆 𝑺𝒉𝒊𝒇𝒕𝒊𝒏𝒈 𝒃𝒚 𝒌 𝒙[𝒏 − 𝒌] 𝑻𝒊𝒎𝒆 𝑺𝒄𝒂𝒍𝒊𝒏𝒈 𝒃𝒚 𝒎 (𝒓𝒆𝒑𝒍𝒂𝒄𝒆 𝒏 𝒃𝒚 ) 𝒙 [ − 𝒌]] 𝑻𝒊𝒎𝒆 ⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗⃗
𝑹𝒆𝒗𝒆𝒓𝒔𝒂𝒍 𝒙 [− − 𝒌]
𝒎 𝒎 𝒎

Time reversal (Folding) and Time delaying (or advancing) on a signal are not commutative

TD=Time delay, FD=Folding

𝑻𝑫𝒌 [𝒙[𝒏]] = 𝒙[𝒏 − 𝒌], 𝒌 > 𝟎

𝑭𝑫[𝒙[𝒏]] = 𝒙[−𝒏]

𝑻𝑫𝒌 [𝑭𝑫[𝒙[𝒏]]] = 𝑻𝑫𝒌 {𝒙[−𝒏]} = 𝒙[−(𝒏 − 𝒌)] = 𝒙[−𝒏 + 𝒌]

Whereas
𝑭𝑫[𝑻𝑫𝒌 [𝒙[𝒏]]] = 𝑭𝑫{𝒙[𝒏 − 𝒌]} = 𝒙[−𝒏 − 𝒌] = 𝒙[−𝒏 − 𝒌]

𝒙[𝒏] → 𝒅𝒆𝒍𝒂𝒚 𝒃𝒚 𝒌 → 𝒙[𝒏 − 𝒌] → 𝒇𝒐𝒍𝒅𝒊𝒏𝒈 → 𝒙[−𝒏 − 𝒌]

𝒙[𝒏] → 𝑭𝒐𝒍𝒅𝒊𝒏𝒈 → 𝒙[−𝒏] → 𝒅𝒆𝒍𝒂𝒚𝒊𝒏𝒈 → 𝒙[−𝒏 + 𝒌]

𝒙[−𝒏 − 𝒌]𝒄𝒂𝒏 𝒃𝒆 𝒐𝒃𝒕𝒂𝒊𝒏𝒆𝒅 𝒃𝒚

1. 𝒙[𝒏] → 𝒅𝒆𝒍𝒂𝒚 𝒃𝒚 𝒌 → 𝒙[𝒏 − 𝒌] → 𝒇𝒐𝒍𝒅𝒊𝒏𝒈 → 𝒙[−𝒏 − 𝒌]

2. 𝒙[𝒏] → 𝒇𝒐𝒍𝒅𝒊𝒏𝒈 → 𝒙[−𝒏] → 𝑨𝒅𝒗𝒂𝒏𝒄𝒊𝒏𝒈 → 𝒙[−𝒏 − 𝒌]
𝒙[−𝒏 + 𝒌]𝒄𝒂𝒏 𝒃𝒆 𝒐𝒃𝒕𝒂𝒊𝒏𝒆𝒅 𝒃𝒚
1. 𝒙[𝒏] → 𝑭𝒐𝒍𝒅𝒊𝒏𝒈 → 𝒙[−𝒏] → 𝒅𝒆𝒍𝒂𝒚𝒊𝒏𝒈 → 𝒙[−𝒏 + 𝒌]
2. 𝒙[𝒏] → 𝑨𝒅𝒗𝒂𝒏𝒄𝒊𝒏𝒈 → 𝒙[𝒏 + 𝒌] → 𝒇𝒐𝒍𝒅𝒊𝒏𝒈 → 𝒙[−𝒏 + 𝒌]
 System
Physical System is broadly characterized by its ability to accept an input such as voltage, current,
force, pressure etc. to produce an output in response to the input.

System
Input x[n] Output y[n]
𝑻[. ]
𝒙[𝒏] → 𝒚[𝒏]
𝒚[𝒏] = 𝑻[𝒙[𝒏]]
𝑻[. ]
𝐒𝐈𝐒𝐎: Single Input Single Output
SIMO: Single Input Multiple Output
MISO: Multiple Input Single Output
MIMO: Multiple Input Multiple Output
Properties of the System:
1. Linearity: Linear System or Non-Linear System
𝒂𝒙𝟏 [𝒏] → 𝒂𝒚𝟏 [𝒏]
𝒃𝒙𝟐 [𝒏] → 𝒃𝒚𝟐 [𝒏]
𝒚[𝒏] = 𝒚𝟏 [𝒏] + 𝒚𝟐 [𝒏]
𝒂𝒙𝟏 [𝒏] + 𝒃𝒙𝟐 [𝒏] → 𝒚′ [𝒏] = 𝒂𝒚𝟏 [𝒏] + 𝒃𝒚𝟐 [𝒏]
𝑰𝒇, 𝒚[𝒏] = 𝒚′ [𝒏]; 𝒕𝒉𝒆𝒏 𝒕𝒉𝒆 𝒔𝒚𝒔𝒕𝒆𝒎 𝒊𝒔 𝑳𝒊𝒏𝒆𝒂𝒓

Example : 𝒚[𝒏] = 𝒏𝒙[𝒏]

𝒙𝟏 [𝒏] → 𝒚𝟏 [𝒏] = 𝒏𝒙𝟏 [𝒏]

𝒙𝟐 [𝒏] → 𝒚𝟐 [𝒏] = 𝒏𝒙𝟐 [𝒏]

𝒚′ [𝒏] = 𝒂𝒚𝟏 [𝒏] + 𝒃𝒚𝟐 [𝒏]

𝒙[𝒏] = 𝒂𝒙𝟏 [𝒏] + 𝒃𝒙𝟐 [𝒏] → 𝒚[𝒏] = 𝒏(𝒂𝒙𝟏 [𝒏] + 𝒃𝒙𝟐 [𝒏]) = 𝒂𝒏𝒙𝟏 [𝒏] + 𝒃𝒏𝒙𝟐 [𝒏] = 𝒂𝒚𝟏 [𝒏] + 𝒃𝒚𝟐 [𝒏]
= 𝒚′ [𝒏]

𝒚[𝒏] = 𝒚′ [𝒏]; Hence 𝒚[𝒏] = 𝒏𝒙[𝒏] is a Linear System

2. Causality: Causal System or Non-Causal System

𝒚[𝒏] = 𝒏𝒙[𝒏]
𝒚[𝒏]|𝒏=𝒏𝟎 = 𝒚[𝒏𝟎 ] = 𝒏𝟎 𝒙[𝒏𝟎 ]; 𝒘𝒉𝒆𝒓𝒆 𝒏𝟎 𝒑𝒐𝒔𝒊𝒕𝒊𝒗𝒆 𝒕𝒊𝒎𝒆
𝒚[𝒏]|𝒏=−𝒏𝟎 = 𝒚[−𝒏𝟎 ] = −𝒏𝟎 𝒙[−𝒏𝟎 ]; 𝒘𝒉𝒆𝒓𝒆 𝒏𝟎 𝒏𝒆𝒈𝒂𝒕𝒊𝒗𝒆 𝒕𝒊𝒎𝒆
𝒚[𝒏] = 𝒙[−𝒏]
3. Time Varying Property: Time Variant or Time Invariant System
 𝒙[𝒏] → 𝒚[𝒏]

𝒚[𝒏] = 𝑻[𝒙[𝒏]]

𝒚[𝒏, 𝒌] = 𝑻[𝒙[𝒏 − 𝒌]] − −→ 𝑹𝒆𝒔𝒑𝒐𝒏𝒔𝒆 𝒅𝒖𝒆 𝒅𝒆𝒍𝒂𝒚𝒆𝒅 𝒊𝒏𝒑𝒖𝒕

𝒙[𝒏] → 𝒚[𝒏]−→ 𝒚[𝒏 − 𝒌]

𝒊𝒇, 𝒚[𝒏, 𝒌] = 𝒚[𝒏 − 𝒌] ==> 𝑻𝒊𝒎𝒆 𝒊𝒏𝒗𝒂𝒓𝒊𝒂𝒏𝒕
𝒊𝒇, 𝒚[𝒏, 𝒌] ≠ 𝒚[𝒏 − 𝒌] ==> 𝑻𝒊𝒎𝒆 𝒗𝒂𝒓𝒊𝒂𝒏𝒕

Example : 𝒚[𝒏] = 𝒏𝒙[𝒏]

𝒙[𝒏] → 𝒚[𝒏] = 𝒏𝒙[𝒏]

𝒚[𝒏, 𝒌] = 𝒏𝒙[𝒏 − 𝒌]

𝒚[𝒏 − 𝒌] = (𝒏 − 𝒌)𝒙[𝒏 − 𝒌]

𝑺𝒊𝒏𝒄𝒆, 𝒚[𝒏, 𝒌] ≠ 𝒚[𝒏 − 𝒌] ==> 𝑻𝒊𝒎𝒆 𝒗𝒂𝒓𝒊𝒂𝒏𝒕

4. Stability: Stable System or Unstable System

BIBO Stable: Bounded IN Bounded OUT
For Input |𝒙[𝒏]| ≤ 𝑩𝒙 < ∞; 𝒇𝒐𝒓 𝒂𝒍𝒍 𝒏 i.e., x[n] is bounded, if the response |𝒚[𝒏]| ≤ 𝑩𝒚 <
∞; 𝒇𝒐𝒓 𝒂𝒍𝒍 𝒏, then the system is said to be Stable.
Example: 𝒚[𝒏] = 𝒏𝒙[𝒏]
Assume that, |𝒙[𝒏]| ≤ 𝑩𝒙 < ∞; 𝒇𝒐𝒓 𝒂𝒍𝒍 𝒏

𝒚[𝒏] = 𝒏𝒙[𝒏]
|𝒚[𝒏]| = |𝒏𝒙[𝒏]| = |𝒏||𝒙[𝒏]| = |𝒏|𝑩𝒙
𝒊𝒇 𝒏 → ∞; 𝒕𝒉𝒆𝒏 |𝒚[𝒏]| → ∞ => 𝒘𝒉𝒊𝒄𝒉 𝒍𝒆𝒂𝒅𝒔 𝒕𝒐 𝒖𝒏𝒃𝒐𝒖𝒏𝒅𝒆𝒅
The condition that the input signal is bounded is not sufficient to guarantee a bounded output.
Hence the system is Unstable
5. Dynamicity Property: System can be Static or Dynamic: Also called System
without memory or System with Memory
𝒚[𝒏] = 𝟓𝒙[𝒏] + 𝒙[𝒏 − 𝟏]
Example: 𝒚[𝒏] = 𝒏𝒙[𝒏]
6. Passivity: Active System or Passive System
If the energy of the output does not exceed the energy of the input, then the system is Passive
System. A system is said to be Passive, if for finite energy input signal x[n], the output signal
y[n]has at most the same energy of x[n].
∞ ∞

∑ |𝒚[𝒏]|𝟐 ≤ ∑ |𝒙[𝒏]|𝟐
𝒏=−∞ 𝒏=−∞
If the energy of the output exceeds the energy of the input, then the system is Active System.
∞ ∞

∑ |𝒚[𝒏]|𝟐 > ∑ |𝒙[𝒏]|𝟐

𝒏=−∞ 𝒏=−∞
 7. Inevitability: Invertible and Inverse System
System 1: w[n] System 2: y[n]=x[n]
x[n]
Invertible Inverse

Linear and Time Invariant System

LTI System
𝒙[𝒏] = 𝜹[𝒏] y[n]

h[n]
𝒙[𝒏] → 𝒚[𝒏]
𝒚[𝒏] = 𝑻[𝒙[𝒏]]
𝜹[𝒏] → 𝒉[𝒏] => 𝑼𝒏𝒊𝒕 𝑰𝒎𝒑𝒖𝒍𝒔𝒆 𝑹𝒆𝒔𝒑𝒐𝒏𝒔𝒆
𝒚[𝒏] = 𝒉[𝒏] = 𝑻[𝜹[𝒏]]
𝒚[𝒏] = 𝑻[𝒙[𝒏]]
∞

𝒚[𝒏] = 𝑻 [ ∑ 𝒙[𝒌]𝜹[𝒏 − 𝒌]]

𝒌=−∞
𝑺𝒊𝒏𝒄𝒆, 𝑳𝒊𝒏𝒆𝒂𝒓
∞

𝒚[𝒏] = ∑ 𝒙[𝒌]𝑻[𝜹[𝒏 − 𝒌]]

𝒌=−∞
𝑺𝒊𝒏𝒄𝒆 𝑻𝒊𝒎𝒆 𝑰𝒏𝒗𝒂𝒓𝒊𝒂𝒏𝒕
∞

𝒚[𝒏] = ∑ 𝒙[𝒌]𝒉(𝒏, 𝒌)
𝒌=−∞
∞

𝒚[𝒏] = ∑ 𝒙[𝒌]𝒉[𝒏 − 𝒌]
𝒌=−∞
𝑪𝒐𝒏𝒗𝒐𝒍𝒖𝒕𝒊𝒐𝒏 𝑺𝒖𝒎
𝒚[𝒏] = 𝒙[𝒏] ∗ 𝒉[𝒏]
1. 𝒙[𝒏] and 𝒉[𝒏]
2. 𝒙[𝒌] and 𝒉[𝒌]
3. 𝒉[−𝒌]--------Folding
4. 𝒉[𝒏 − 𝒌]-------Shifting
5. 𝒙[𝒌]. 𝒉[𝒏 − 𝒌]-----Multiplication
6. ∑∞
𝒌=−∞ 𝒙[𝒌]𝒉[𝒏 − 𝒌] − − − − − 𝑺𝒖𝒎𝒎𝒂𝒕𝒊𝒐𝒏

Folding (Once) →(Shifting→Multiplication→Summation) repeating k times

Property Of Convolution/Property LTI System

1. Commutative:
Convolution is Commutative:
𝒙[𝒏] ∗ 𝒉[𝒏] = 𝒉[𝒏] ∗ 𝒙[𝒏]
2. Associative:
𝒙[𝒏] ∗ (𝒉𝟏 [𝒏] ∗ 𝒉𝟐 [𝒏]) = (𝒙[𝒏] ∗ 𝒉𝟏 [𝒏]) ∗ 𝒉𝟐 [𝒏]

𝒘[𝒏] = 𝒙[𝒏] ∗ 𝒉𝟏 [𝒏]

LTI System LTI System 𝒚[𝒏] = 𝒘[𝒏] ∗ 𝒉𝟐 [𝒏] = (𝒙[𝒏] ∗ 𝒉𝟏[𝒏]) ∗ 𝒉𝟐 [𝒏]

h1[n] h2[n]

𝒙[𝒏] LTI System 𝒚[𝒏] = 𝒙[𝒏] ∗ (𝒉𝟏 [𝒏] ∗ 𝒉𝟐 [𝒏])

h1[n]* h2[n]

3. Distributive:
𝒙[𝒏] ∗ (𝒉𝟏 [𝒏] + 𝒉𝟐 [𝒏]) = 𝒙[𝒏] ∗ 𝒉𝟏 [𝒏] + 𝒙[𝒏] ∗ 𝒉𝟐 [𝒏]

LTI System
h1[n]
𝒙[𝒏] + 𝒚[𝒏]

LTI System
h2[n]

LTI System
𝒙[𝒏] 𝒚[𝒏]
h1[n]+h2[n]
4. Impulse Convolution
𝒙[𝒏] ∗ 𝜹[𝒏] = 𝒙[𝒏]
5. Shift Property
𝑰𝒇, 𝒙[𝒏] ∗ 𝒉[𝒏] = 𝒚[𝒏]
𝑻𝒉𝒆𝒏, 𝒙[𝒏] ∗ 𝒉[𝒏 − 𝒌] = 𝒙[𝒏 − 𝒌] ∗ 𝒉[𝒏] = 𝒚[𝒏 − 𝒌]
6. Width Property
𝒚[𝒏] = 𝒙[𝒏] ∗ 𝒉[𝒏]
𝒏𝒙𝟏 = 𝑺𝒕𝒂𝒓𝒕𝒊𝒏𝒈 𝒕𝒊𝒎𝒆 𝒊𝒏𝒅𝒆𝒙 𝒐𝒇 𝒙[𝒏]
𝒏𝒙𝟐 = 𝑬𝒏𝒅𝒊𝒏𝒈 𝒕𝒊𝒎𝒆 𝒊𝒏𝒅𝒆𝒙 𝒐𝒇 𝒙[𝒏]
𝒏𝒉𝟏 = 𝑺𝒕𝒂𝒓𝒕𝒊𝒏𝒈 𝒕𝒊𝒎𝒆 𝒊𝒏𝒅𝒆𝒙 𝒐𝒇 𝒉[𝒏]
𝒏𝒉𝟐 = 𝑬𝒏𝒅𝒊𝒏𝒈 𝒕𝒊𝒎𝒆 𝒊𝒏𝒅𝒆𝒙 𝒐𝒇 𝒉[𝒏]
𝒏𝒚𝟏 = 𝑺𝒕𝒂𝒓𝒕𝒊𝒏𝒈 𝒕𝒊𝒎𝒆 𝒊𝒏𝒅𝒆𝒙 𝒐𝒇 𝒚[𝒏]
𝒏𝒚𝟏 = 𝒏𝒙𝟏 + 𝒏𝒉𝟏
𝑵𝒙 = 𝑵𝒖𝒎𝒃𝒆𝒓 𝒐𝒇 𝑺𝒂𝒎𝒑𝒍𝒆𝒔 𝒊𝒏 𝒙[𝒏]
𝑵𝒉 = 𝑵𝒖𝒎𝒃𝒆𝒓 𝒐𝒇 𝑺𝒂𝒎𝒑𝒍𝒆𝒔 𝒊𝒏 𝒉[𝒏]
𝑵𝒚 = 𝑵𝒖𝒎𝒃𝒆𝒓 𝒐𝒇 𝑺𝒂𝒎𝒑𝒍𝒆𝒔 𝒊𝒏 𝒚[𝒏]
𝑵𝒚 = 𝑵𝒙 + 𝑵𝒉 − 𝟏
7. Sum Property
𝒚[𝒏] = 𝒙[𝒏] ∗ 𝒉[𝒏]
∞ ∞ ∞

𝑺𝒚 = ∑ 𝒚[𝒏] , 𝑺𝒙 = ∑ 𝒙[𝒏] , 𝑺𝒉 = ∑ 𝒉[𝒏]

𝒏=−∞ 𝒏=−∞ 𝒏=−∞
𝑺 𝒚 = 𝑺𝒙 𝑺𝒉
∞ ∞ ∞

∑ 𝒚[𝒏] = ∑ 𝒙[𝒏] ∑ 𝒉[𝒏]

𝒏=−∞ 𝒏=−∞ 𝒏=−∞
8. Causality for LTI System
∞

𝒚[𝒏] = ∑ 𝒙[𝒌]𝒉[𝒏 − 𝒌]
𝒌=−∞
∞

𝒚[𝒏] = ∑ 𝒉[𝒌]𝒙[𝒏 − 𝒌]
𝒌=
−∞
−𝟏 ∞

= ∑ 𝒉[𝒌]𝒙[𝒏 − 𝒌] + ∑ 𝒉[𝒌]𝒙[𝒏 − 𝒌]
𝒌=−∞ 𝒌=𝟎
𝒚[𝒏] = [… 𝒉[−𝟑]. 𝒙[𝒏 + 𝟑] + 𝒉[−𝟐]𝒙[𝒏 + 𝟐] + 𝒉[−𝟏]. 𝒙[𝒏 + 𝟏]]
+ [𝒉[𝟎]𝒙[𝒏] + 𝒉[𝟏]𝒙[𝒏 − 𝟏] + ⋯ ]
𝑻𝒉𝒆 𝒄𝒐𝒏𝒅𝒊𝒕𝒊𝒐𝒏 𝒇𝒐𝒓 𝑪𝒂𝒖𝒔𝒂𝒍𝒊𝒕𝒚 𝒐𝒇 𝒂𝒏 𝑳𝑻𝑰 𝒔𝒚𝒔𝒕𝒆𝒎:
𝒉[𝒏] = 𝟎; 𝒇𝒐𝒓 𝑨𝒍𝒍 𝒏 < 𝟎
9. Stability for LTI System:
∞

𝒚[𝒏] = ∑ 𝒉[𝒌]𝒙[𝒏 − 𝒌]
𝒌=−∞
 ∞

|𝒚[𝒏]| = | ∑ 𝒉[𝒌]𝒙[𝒏 − 𝒌]|

𝒌=−∞
∞

|𝒚[𝒏]| ≤ ∑ |𝒉[𝒌]||𝒙[𝒏 − 𝒌]|

𝒌=−∞
If we assume that the x[n] is bounded, i.e, |𝒙[𝒏]| ≤ 𝑩𝒙 < ∞, then |𝒙[𝒏 − 𝒌]| ≤ 𝑩𝒙 < ∞,
∞

|𝒚[𝒏]| ≤ 𝑩𝒙 ∑ |𝒉[𝒌]|
⏟
𝒌=−∞
𝑩𝒚
|𝒚[𝒏]| ≤ 𝑩𝒙 𝑩𝒚
From this equation, we can conclude that if the impulse response is absolutely summable, that is ,
if
∞

∑ |𝒉[𝒌]| = 𝑩𝒚 < ∞
𝒌=−∞
Then y[n] is bounded and hence the system is stable.
10. Invertibility for LTI System
Invertible Inverse
LTI System 1 w[n] LTI System 2
y[n]=x[n]
x[n] 𝒉𝟏 [𝒏] 𝒉𝟐 [𝒏]

𝒚[𝒏] = 𝒙[𝒏] ∗ (𝒉𝟏 [𝒏] ∗ 𝒉𝟐 [𝒏]) = 𝒙[𝒏]

𝒉𝟏 [𝒏] ∗ 𝒉𝟐 [𝒏] = 𝜹[𝒏]

11.LTI System with memory and without memory

∞

𝒚[𝒏] = ∑ 𝒉[𝒌]𝒙[𝒏 − 𝒌]
𝒌=−∞
−𝟏 ∞

= ∑ 𝒉[𝒌]𝒙[𝒏 − 𝒌] + ∑ 𝒉[𝒌]𝒙[𝒏 − 𝒌]
𝒌=−∞ 𝒌=𝟎
𝒚[𝒏] = [… 𝒉[−𝟑]. 𝒙[𝒏 + 𝟑] + 𝒉[−𝟐]𝒙[𝒏 + 𝟐] + 𝒉[−𝟏]. 𝒙[𝒏 + 𝟏]]
+ [𝒉[𝟎]𝒙[𝒏] + 𝒉[𝟏]𝒙[𝒏 − 𝟏] + ⋯ ]
𝒉[𝒏] = 𝑲𝜹[𝒏], 𝑲 = 𝒉[𝟎]

Unit Step Response of LTI System s[n]---→Home Work

𝒔[𝒏] = 𝒖[𝒏] ∗ 𝒉[𝒏]

𝒉[𝒏] = 𝒔[𝒏] − 𝒔[𝒏 − 𝟏]
 Discrete Convolution (Convolution Sum) Evaluation
1. Conventional Method (Using the Formula: Analytical Method)
∞

𝒚[𝒏] = ∑ 𝒉[𝒌]𝒙[𝒏 − 𝒌]
𝒌=−∞
𝒙[𝒏]𝒂𝒏𝒅 𝒉[𝒏]−→
𝑰𝒏𝒇𝒊𝒏𝒊𝒕𝒆 𝒍𝒆𝒏𝒈𝒕𝒉 𝒂𝒏𝒅 𝒇𝒊𝒏𝒊𝒕𝒆 𝒍𝒆𝒏𝒈𝒕𝒉 𝒃𝒖𝒕 𝒂𝒓𝒆 𝒅𝒆𝒔𝒄𝒓𝒊𝒃𝒆𝒅 𝒃𝒚
𝒔𝒊𝒎𝒑𝒍𝒆 𝒆𝒏𝒐𝒖𝒈𝒉 𝒂𝒏𝒂𝒍𝒚𝒕𝒊𝒄𝒂𝒍 𝒆𝒙𝒑𝒓𝒆𝒔𝒔𝒊𝒐𝒏
𝒙[𝒏]𝒊𝒔 𝑪𝒂𝒖𝒔𝒂𝒍 = 𝒙[𝒏]𝒖[𝒏]
𝒖[𝒌]𝒂𝒏𝒅 𝒖[𝒏 − 𝒌]

Ex. 1 Find the Convolution of the following two signals

𝟐
𝒙[𝒏] = 𝒆−𝒏 𝒇𝒐𝒓 𝒂𝒍𝒍 𝒏

𝒉[𝒏] = 𝟑𝒏𝟐 𝒇𝒐𝒓 𝒂𝒍𝒍 𝒏

Sol.
By definition
∞

𝒚[𝒏] = ∑ 𝒙[𝒌]𝒉[𝒏 − 𝒌]
𝒌=−∞

∞
𝟐
𝒚[𝒏] = ∑ 𝒆−𝒌 𝟑(𝒏 − 𝒌)𝟐
𝒌=−∞

∞
𝟐
𝒚[𝒏] = ∑ 𝒆−𝒌 𝟑(𝒏𝟐 − 𝟐𝒏𝒌 + 𝒌𝟐 )
𝒌=−∞

∞ ∞ ∞
𝟐 −𝒌𝟐 −𝒌𝟐 𝟐
𝒚[𝒏] = 𝟑𝒏 ∑ 𝒆 − 𝟔𝒏 ∑ 𝒆 . 𝒌 + 𝟑 ∑ 𝒆−𝒌 . 𝒌𝟐
𝒌=−∞ 𝒌=−∞ 𝒌=−∞

∞ ∞
𝟐 −𝒌𝟐 𝟐
𝒚[𝒏] = 𝟑𝒏 ∑ 𝒆 − 𝟎 + 𝟑 ∑ 𝒆−𝒌 . 𝒌𝟐 = 𝟓. 𝟑𝟏𝟖𝒏𝟐 + 𝟐. 𝟔𝟓𝟒
𝒌=−∞ 𝒌=−∞

𝒚[𝒏] = 𝟓. 𝟑𝟏𝟖𝒏𝟐 + 𝟐. 𝟔𝟓𝟒; 𝒇𝒐𝒓 𝒂𝒍𝒍 𝒏

Ex.2. Compute the convolution 𝒚[𝒏] = 𝒙[𝒏] ∗ 𝒉[𝒏] of the following signals

𝒙[𝒏] = 𝒖[𝒏]𝒂𝒏𝒅 𝒉[𝒏] = 𝒖[𝒏]

Sol.
∞

𝒚[𝒏] = ∑ 𝒙[𝒌]𝒉[𝒏 − 𝒌]
𝒌=−∞
∞

𝒚[𝒏] = ∑ 𝒖[𝒌]𝒖[𝒏 − 𝒌]
𝒌=−∞
= 𝟏; 𝒇𝒐𝒓𝒂𝒍𝒍 𝒌 ≥ 𝟎
𝒖[𝒌] = {
𝟎; 𝑶𝒕𝒉𝒆𝒓𝒘𝒊𝒔𝒆
𝟏; 𝒇𝒐𝒓𝒂𝒍𝒍 𝒏 − 𝒌 ≥ 𝟎; 𝒏 ≥ 𝒌
𝒖[𝒏 − 𝒌] = {
𝟎; 𝑶𝒕𝒉𝒆𝒓𝒘𝒊𝒔𝒆; 𝒏 − 𝒌 < 𝟎; 𝒏 < 𝒌
𝒏

𝒚[𝒏] = ∑ 𝟏 = (𝒏 + 𝟏)𝒖[𝒏] = 𝒓[𝒏 + 𝟏]

𝒌=𝟎

Ex.3. Find the Convolution of the two sequences

𝟎; 𝒏 < −𝟓
𝒙[𝒏] = { 𝟏 𝒏
( ) ; 𝒏 ≥ −𝟓
𝟐

𝟏 𝒌
𝒙[𝒌] = ( ) ; 𝒌 = −𝟓, −𝟒, −𝟑, … … ∞
𝟐
𝟎; 𝒏 < 𝟑
𝒏
𝒉[𝒏] = { 𝟏
( ) ;𝒏 ≥ 𝟑
𝟑

𝟏 𝒏−𝒌
𝒉[𝒏 − 𝒌] = ( ) ; 𝒏 − 𝒌 ≥ 𝟑 => 𝒌 ≤ 𝒏 − 𝟑
𝟑
Do it……
2. Graphical Method
∞

𝒚[𝒏] = ∑ 𝒙[𝒌]𝒉[𝒏 − 𝒌]
𝒌=−∞
1. Plotting: Plot x[k] and h[k] as a function of k
2. Fold the h[k] to make h[-k]
3. Shift h[-k] to make it h[n-k]
 4. Multiplication and addition.

Ex. 4 Find the convolution of the following two signals

𝒙[𝒏] = {𝟏, 𝟓, 𝟑, 𝟒, 𝟐}

𝒉[𝒏] = {−𝟏, 𝟕, 𝟐, 𝟑, 𝟓}
Sol.

𝒚[𝒏] = 𝒙[𝒏] ∗ 𝒉[𝒏]

nx1=Starting point of 𝐱[𝐧] = −𝟏

nx2=Ending point of 𝐱[𝐧] = 𝟑

nh1=Starting point of 𝐡[𝐧] = −𝟐

nh2=Ending point of 𝐡[𝐧] = 𝟐

ny1=Starting point of 𝐲[𝐧] = 𝐧𝐱𝟏 + 𝐧𝐡𝟏 = −𝟏 + (−𝟐) = −𝟑

ny2=Ending point of 𝐲[𝐧] = 𝐧𝐱𝟐 + 𝐧𝐡𝟐 = 𝟑 + 𝟐 = 𝟓

Number of samples in 𝐲[𝐧], 𝐍𝐲 = 𝟓 + 𝟓 − 𝟏 = 𝟗

 3. Tabular Method (Finite Length Sequence)
Ex. 4 Find the convolution of the following two signals

𝒙[𝒏] = {𝟏, 𝟓, 𝟑, 𝟒, 𝟐}

𝒉[𝒏] = {−𝟏, 𝟕, 𝟐, 𝟑, 𝟓}
Sol.

𝒚[𝒏] = 𝒙[𝒏] ∗ 𝒉[𝒏]

nx1=Starting point of x[n]=-1
nx2=Ending point of x[n]=3
nh1=Starting point of h[n]=-2
nh2=Ending point of h[n]=2
ny1=Starting point of y[n]=nx1+nh1=-1+(-2)=-3
ny2=Ending point of y[n]=nx2+nh2=3+2=5
Nx=Number of Samples in x[n]=5
Nh=Number of Samples in h[n]=5
Ny=Number of Samples in y[n]=Nx+Nh-1=5+5-1=9
h[n] -1 7 2 3 5
x[n]
1 -1 7 2 3 5
5 -5 35 10 15 25
3 -3 21 6 9 15
4 -4 28 8 12 20
2 -2 14 4 6 10
 y[n]={-1, 2, 34, 30, 52, 56, 31, 26, 10}

4. Multiplication Method

h[n]: -1 7 2 3 5
x[n]: 1 5 3 4 2

-2 14 4 6 10
-4 28 8 12 20
-3 21 6 9 15
-5 35 10 15 25
-1 7 2 3 5
-1 2 34 30 52 56 31 26 10

y[n]={-1, 2, 34, 30, 52, 56, 31, 26, 10}

Matrix Method