Lecture 5: Discrete -Time Convolution

Foundations of Digital Signal Processing

Outline
• Review
• Discrete-Time Convolution
• Properties of Discrete-Time Convolution
• Combining Systems
• Properties of the Impulse Response

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 1

News
 Monday
 No class

 Homework #2
 Due Wednesday by 11:59 PM

 Submit via canvas

 Coding Assignment #1
 Due Wednesday by 11:59 PM

 Submit via canvas

◊ Submit answers as a PDF
◊ Submit code as .m files

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 2

Lecture 5: Discrete -Time Convolution
Foundations of Digital Signal Processing

Outline
• Review
• Discrete-Time Convolution
• Properties of Discrete-Time Convolution
• Combining Systems
• Properties of the Impulse Response

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 3

Review
 What did we learn in the last class?

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 4

Convolution Examples
 Solving a convolution
∞
𝑦𝑦 𝑡𝑡 = 𝑥𝑥 𝑡𝑡 ∗ ℎ 𝑡𝑡 = � 𝑥𝑥 𝜏𝜏 ℎ 𝑡𝑡 − 𝜏𝜏 𝑑𝑑𝑑𝑑
−∞

𝑥𝑥 𝑡𝑡 = 𝑒𝑒 −𝑡𝑡 u t , ℎ 𝑡𝑡 = 𝑢𝑢 𝑡𝑡

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 5

Convolution Examples
 Solving a convolution
∞
𝑦𝑦 𝑡𝑡 = 𝑥𝑥 𝑡𝑡 ∗ ℎ 𝑡𝑡 = � 𝑥𝑥 𝜏𝜏 ℎ 𝑡𝑡 − 𝜏𝜏 𝑑𝑑𝑑𝑑
−∞

𝑥𝑥 𝑡𝑡 = 𝑒𝑒 −𝑡𝑡 u t , ℎ 𝑡𝑡 = 𝑢𝑢 𝑡𝑡
∞
𝑦𝑦 𝑡𝑡 = � 𝑒𝑒 −𝜏𝜏 𝑢𝑢 𝜏𝜏 𝑢𝑢 𝑡𝑡 − 𝜏𝜏 𝑑𝑑𝑑𝑑
−∞

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 6

Convolution Examples
 Solving a convolution
∞
𝑦𝑦 𝑡𝑡 = 𝑥𝑥 𝑡𝑡 ∗ ℎ 𝑡𝑡 = � 𝑥𝑥 𝜏𝜏 ℎ 𝑡𝑡 − 𝜏𝜏 𝑑𝑑𝑑𝑑
−∞

𝑥𝑥 𝑡𝑡 = 𝑒𝑒 −𝑡𝑡 u t , ℎ 𝑡𝑡 = 𝑢𝑢 𝑡𝑡
∞
𝑦𝑦 𝑡𝑡 = � 𝑒𝑒 −𝜏𝜏 𝑢𝑢 𝜏𝜏 𝑢𝑢 𝑡𝑡 − 𝜏𝜏 𝑑𝑑𝑑𝑑 𝑢𝑢 𝑡𝑡 − 𝜏𝜏
−∞

𝑡𝑡 𝜏𝜏

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 7

Convolution Examples
 Solving a convolution
∞
𝑦𝑦 𝑡𝑡 = 𝑥𝑥 𝑡𝑡 ∗ ℎ 𝑡𝑡 = � 𝑥𝑥 𝜏𝜏 ℎ 𝑡𝑡 − 𝜏𝜏 𝑑𝑑𝑑𝑑
−∞

𝑥𝑥 𝑡𝑡 = 𝑒𝑒 −𝑡𝑡 u t , ℎ 𝑡𝑡 = 𝑢𝑢 𝑡𝑡
∞
𝑦𝑦 𝑡𝑡 = � 𝑒𝑒 −𝜏𝜏 𝑢𝑢 𝜏𝜏 𝑢𝑢 𝑡𝑡 − 𝜏𝜏 𝑑𝑑𝑑𝑑 𝑢𝑢 𝑡𝑡 − 𝜏𝜏
−∞

𝑡𝑡 𝜏𝜏

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 8

Convolution Examples
 Solving a convolution
∞
𝑦𝑦 𝑡𝑡 = 𝑥𝑥 𝑡𝑡 ∗ ℎ 𝑡𝑡 = � 𝑥𝑥 𝜏𝜏 ℎ 𝑡𝑡 − 𝜏𝜏 𝑑𝑑𝑑𝑑
−∞

𝑥𝑥 𝑡𝑡 = 𝑒𝑒 −𝑡𝑡 u t , ℎ 𝑡𝑡 = 𝑢𝑢 𝑡𝑡
∞
𝑦𝑦 𝑡𝑡 = � 𝑒𝑒 −𝜏𝜏 𝑢𝑢 𝜏𝜏 𝑢𝑢 𝑡𝑡 − 𝜏𝜏 𝑑𝑑𝑑𝑑 𝑢𝑢 𝑡𝑡 − 𝜏𝜏
−∞

𝑡𝑡 𝜏𝜏

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 9

Convolution Examples
 Solving a convolution
∞
𝑦𝑦 𝑡𝑡 = 𝑥𝑥 𝑡𝑡 ∗ ℎ 𝑡𝑡 = � 𝑥𝑥 𝜏𝜏 ℎ 𝑡𝑡 − 𝜏𝜏 𝑑𝑑𝑑𝑑
−∞

𝑥𝑥 𝑡𝑡 = 𝑒𝑒 −𝑡𝑡 u t , ℎ 𝑡𝑡 = 𝑢𝑢 𝑡𝑡
∞
𝑦𝑦 𝑡𝑡 = � 𝑒𝑒 −𝜏𝜏 𝑢𝑢 𝜏𝜏 𝑢𝑢 𝑡𝑡 − 𝜏𝜏 𝑑𝑑𝑑𝑑
−∞
𝑡𝑡
𝑦𝑦 𝑡𝑡 = � 𝑒𝑒 −𝜏𝜏 𝑑𝑑𝑑𝑑
0

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 10

Convolution Examples
 Solving a convolution
∞
𝑦𝑦 𝑡𝑡 = 𝑥𝑥 𝑡𝑡 ∗ ℎ 𝑡𝑡 = � 𝑥𝑥 𝜏𝜏 ℎ 𝑡𝑡 − 𝜏𝜏 𝑑𝑑𝑑𝑑
−∞

𝑥𝑥 𝑡𝑡 = 𝑒𝑒 −𝑡𝑡 u t , ℎ 𝑡𝑡 = 𝑢𝑢 𝑡𝑡
∞
𝑦𝑦 𝑡𝑡 = � 𝑒𝑒 −𝜏𝜏 𝑢𝑢 𝜏𝜏 𝑢𝑢 𝑡𝑡 − 𝜏𝜏 𝑑𝑑𝑑𝑑
−∞
𝑡𝑡
𝑦𝑦 𝑡𝑡 = � 𝑒𝑒 −𝜏𝜏 𝑑𝑑𝑑𝑑
0
𝑡𝑡 𝑡𝑡
𝑦𝑦 𝑡𝑡 = � 𝑒𝑒 −𝜏𝜏 𝑑𝑑𝑑𝑑 = −𝑒𝑒 −𝜏𝜏 � = −𝑒𝑒 −𝑡𝑡 + 1 = 1 − 𝑒𝑒 −𝑡𝑡
0 0

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 11

Lecture 5: Discrete -Time LTI Systems
Foundations of Digital Signal Processing

Outline
• Review
• Discrete-Time Convolution
• Properties of Discrete-Time Convolution
• Combining Systems
• Properties of the Impulse Response

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 12

Convolution
 Consider the system defined by the input-output relationship
𝑛𝑛

𝑦𝑦 𝑛𝑛 = � 𝑥𝑥 𝑚𝑚
𝑚𝑚=−∞

 Compute the output for input 𝑥𝑥 𝑛𝑛 = 𝛿𝛿 𝑛𝑛 − 1 − 3𝛿𝛿 𝑛𝑛 − 3

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 13

Convolution
 Consider the system defined by the input-output relationship
𝑛𝑛

𝑦𝑦 𝑛𝑛 = � 𝑥𝑥 𝑚𝑚
𝑚𝑚=−∞

 Compute the output for input 𝑥𝑥 𝑛𝑛 = 𝛿𝛿 𝑛𝑛 − 1 − 3𝛿𝛿 𝑛𝑛 − 3

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 14

Convolution
 Consider the system defined by the input-output relationship
𝑛𝑛

𝑦𝑦 𝑛𝑛 = � 𝑥𝑥 𝑚𝑚
𝑚𝑚=−∞

 Compute the output for input 𝑥𝑥 𝑛𝑛 = 𝛿𝛿 𝑛𝑛 − 1 − 3𝛿𝛿 𝑛𝑛 − 3

Solution:

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 15

Convolution
 Consider the system defined by the input-output relationship
𝑛𝑛

𝑦𝑦 𝑛𝑛 = � 𝑥𝑥 𝑚𝑚
𝑚𝑚=−∞

 Compute the output for input 𝑥𝑥 𝑛𝑛 = 𝑢𝑢 𝑛𝑛

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 16

Convolution
 Consider the system defined by the input-output relationship
𝑛𝑛

𝑦𝑦 𝑛𝑛 = � 𝑥𝑥 𝑚𝑚
𝑚𝑚=−∞

 Compute the output for input 𝑥𝑥 𝑛𝑛 = 𝑢𝑢 𝑛𝑛

Solution:

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 17

Convolution
 Definition of convolution
∞

𝑦𝑦 𝑛𝑛 = 𝑥𝑥 𝑛𝑛 ∗ ℎ 𝑛𝑛 = � 𝑥𝑥 𝑚𝑚 ℎ 𝑛𝑛 − 𝑚𝑚
𝑚𝑚=−∞

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 18

Convolution
 Definition of convolution
∞

𝑦𝑦 𝑛𝑛 = 𝑥𝑥 𝑛𝑛 ∗ ℎ 𝑛𝑛 = � 𝑥𝑥 𝑚𝑚 ℎ 𝑛𝑛 − 𝑚𝑚
𝑚𝑚=−∞
Shift that Domain
corresponds that we
to the final multiply
point and sum
over

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 19

Signal Operations
 Consider this
 Plotting 𝑥𝑥 3 − 𝑛𝑛
 We follow this procedure
◊ 𝑥𝑥 𝑛𝑛 → 𝑥𝑥 𝑛𝑛 + 3 → 𝑥𝑥 −𝑛𝑛 + 3 = 𝑥𝑥 3 − 𝑛𝑛
Shift Time
left 3 reverse

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 20

Signal Operations
 Consider this
 Plotting 𝑥𝑥 3 − 𝑛𝑛
 We follow this procedure
◊ 𝑥𝑥 𝑛𝑛 → 𝑥𝑥 𝑛𝑛 + 3 → 𝑥𝑥 −𝑛𝑛 + 3 = 𝑥𝑥 3 − 𝑛𝑛
Shift Time
left 3 reverse

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 21

Signal Operations
 Consider this
 Plotting 𝑥𝑥 3 − 𝑛𝑛
 We follow this procedure
◊ 𝑥𝑥 𝑛𝑛 → 𝑥𝑥 𝑛𝑛 + 3 → 𝑥𝑥 −𝑛𝑛 + 3 = 𝑥𝑥 3 − 𝑛𝑛
Shift Time
left 3 reverse

Shift
left 3

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 22

Signal Operations
 Consider this
 Plotting 𝑥𝑥 3 − 𝑛𝑛
 We follow this procedure
◊ 𝑥𝑥 𝑛𝑛 → 𝑥𝑥 𝑛𝑛 + 3 → 𝑥𝑥 −𝑛𝑛 + 3 = 𝑥𝑥 3 − 𝑛𝑛
Shift Time
left 3 reverse

Shift Time
left 3 reverse

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 23

Signal Operations
 Consider this
 Plotting 𝑥𝑥 3 − 𝑛𝑛
 We follow this procedure
◊ 𝑥𝑥 𝑛𝑛 → 𝑥𝑥 𝑛𝑛 + 3 → 𝑥𝑥 −𝑛𝑛 + 3 = 𝑥𝑥 3 − 𝑛𝑛
Shift Time
left 3 reverse

 We can go in the other direction (used to illustrate convolution)

◊ 𝑥𝑥 𝑛𝑛 → 𝑥𝑥 −𝑛𝑛 → 𝑥𝑥 −(𝑛𝑛 − 3) = 𝑥𝑥 −𝑛𝑛 + 3 = 𝑥𝑥 3 − 𝑛𝑛

Time Shift
reverse right 3

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 24

Signal Operations
 Consider this
 Plotting 𝑥𝑥 3 − 𝑛𝑛

 We can go in the other direction (used to illustrate convolution)

◊ 𝑥𝑥 𝑛𝑛 → 𝑥𝑥 −𝑛𝑛 → 𝑥𝑥 −(𝑛𝑛 − 3) = 𝑥𝑥 −𝑛𝑛 + 3 = 𝑥𝑥 3 − 𝑛𝑛

Time Shift
reverse right 3

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 25

Signal Operations
 Consider this
 Plotting 𝑥𝑥 3 − 𝑛𝑛

Time
reverse

 We can go in the other direction (used to illustrate convolution)

◊ 𝑥𝑥 𝑛𝑛 → 𝑥𝑥 −𝑛𝑛 → 𝑥𝑥 −(𝑛𝑛 − 3) = 𝑥𝑥 −𝑛𝑛 + 3 = 𝑥𝑥 3 − 𝑛𝑛

Time Shift
reverse right 3

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 26

Signal Operations
 Consider this
 Plotting 𝑥𝑥 3 − 𝑛𝑛

Time Shift
reverse right 3

 We can go in the other direction (used to illustrate convolution)

◊ 𝑥𝑥 𝑛𝑛 → 𝑥𝑥 −𝑛𝑛 → 𝑥𝑥 −(𝑛𝑛 − 3) = 𝑥𝑥 −𝑛𝑛 + 3 = 𝑥𝑥 3 − 𝑛𝑛

Time Shift
reverse right 3

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 27

Signal Operations
 Consider this
 Plotting 𝑥𝑥 3 − 𝑛𝑛

Time Shift
reverse right 3

Shift Time
left 3 reverse

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 28

Convolution Illustration

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 29

 ∞
 Convolution: 𝑥𝑥 𝑛𝑛 ∗ ℎ 𝑛𝑛 = ∑𝑚𝑚=−∞ 𝑥𝑥 𝑚𝑚 ℎ 𝑛𝑛 − 𝑚𝑚

Input
Second 1 0 3 8 7 -4 9 3 -7
Impulse response
First ∗
0 1 2

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution

 ∞
 Convolution: 𝑥𝑥 𝑛𝑛 ∗ ℎ 𝑛𝑛 = ∑𝑚𝑚=−∞ 𝑥𝑥 𝑚𝑚 ℎ 𝑛𝑛 − 𝑚𝑚
Delay of 2 units
1
0.5

Input
Second 1 0 3 8 7 -4 9 3 -7
Impulse response
First ∗
0 1 2

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution

 ∞
 Convolution: 𝑥𝑥 𝑛𝑛 ∗ ℎ 𝑛𝑛 = ∑𝑚𝑚=−∞ 𝑥𝑥 𝑚𝑚 ℎ 𝑛𝑛 − 𝑚𝑚

Input
Second 1 0 3 8 7 -4 9 3 -7
Impulse response
First ∗
0 1 2

0.5

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution

 ∞
 Convolution: 𝑥𝑥 𝑛𝑛 ∗ ℎ 𝑛𝑛 = ∑𝑚𝑚=−∞ 𝑥𝑥 𝑚𝑚 ℎ 𝑛𝑛 − 𝑚𝑚

Input
Second 1 0 3 8 7 -4 9 3 -7
Impulse response
First ∗
0 1 2

0.5 0

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution

 ∞
 Convolution: 𝑥𝑥 𝑛𝑛 ∗ ℎ 𝑛𝑛 = ∑𝑚𝑚=−∞ 𝑥𝑥 𝑚𝑚 ℎ 𝑛𝑛 − 𝑚𝑚

Input
Second 1 0 3 8 7 -4 9 3 -7
Impulse response
First ∗
0 1 2

0.5 0 1.5+1

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution

 ∞
 Convolution: 𝑥𝑥 𝑛𝑛 ∗ ℎ 𝑛𝑛 = ∑𝑚𝑚=−∞ 𝑥𝑥 𝑚𝑚 ℎ 𝑛𝑛 − 𝑚𝑚

Input
Second 1 0 3 8 7 -4 9 3 -7
Impulse response
First ∗
0 1 2

0.5 0 1.5+1 4+0

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution

 ∞
 Convolution: 𝑥𝑥 𝑛𝑛 ∗ ℎ 𝑛𝑛 = ∑𝑚𝑚=−∞ 𝑥𝑥 𝑚𝑚 ℎ 𝑛𝑛 − 𝑚𝑚

Input
Second 1 0 3 8 7 -4 9 3 -7
Impulse response
First ∗
0 1 2

0.5 0 1.5+1 4+0 3.5+3

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution

 ∞
 Convolution: 𝑥𝑥 𝑛𝑛 ∗ ℎ 𝑛𝑛 = ∑𝑚𝑚=−∞ 𝑥𝑥 𝑚𝑚 ℎ 𝑛𝑛 − 𝑚𝑚

Input
Second 1 0 3 8 7 -4 9 3 -7
Impulse response
First ∗
0 1 2

0.5 0 1.5+1 4+0 3.5+3 -2+8

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution

 ∞
 Convolution: 𝑥𝑥 𝑛𝑛 ∗ ℎ 𝑛𝑛 = ∑𝑚𝑚=−∞ 𝑥𝑥 𝑚𝑚 ℎ 𝑛𝑛 − 𝑚𝑚

Input
Second 1 0 3 8 7 -4 9 3 -7
Impulse response
First ∗
0 1 2

0.5 0 1.5+1 4+0 3.5+3 -2+8 4.5+7

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution

 ∞
 Convolution: 𝑥𝑥 𝑛𝑛 ∗ ℎ 𝑛𝑛 = ∑𝑚𝑚=−∞ 𝑥𝑥 𝑚𝑚 ℎ 𝑛𝑛 − 𝑚𝑚

Input
Second 1 0 3 8 7 -4 9 3 -7
Impulse response
First ∗
0 1 2

0.5 0 1.5+1 4+0 3.5+3 -2+8 4.5+7 1.5-4

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution

 ∞
 Convolution: 𝑥𝑥 𝑛𝑛 ∗ ℎ 𝑛𝑛 = ∑𝑚𝑚=−∞ 𝑥𝑥 𝑚𝑚 ℎ 𝑛𝑛 − 𝑚𝑚

Input
Second 1 0 3 8 7 -4 9 3 -7
Impulse response
First ∗
0 1 2

0.5 0 1.5+1 4+0 3.5+3 -2+8 4.5+7 1.5-4 -3.5+9

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution

 ∞
 Convolution: 𝑥𝑥 𝑛𝑛 ∗ ℎ 𝑛𝑛 = ∑𝑚𝑚=−∞ 𝑥𝑥 𝑚𝑚 ℎ 𝑛𝑛 − 𝑚𝑚

Input
Second 1 0 3 8 7 -4 9 3 -7
Impulse response
First ∗
0 1 2

0.5 0 1.5+1 4+0 3.5+3 -2+8 4.5+7 1.5-4 -3.5+9 3

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution

 ∞
 Convolution: 𝑥𝑥 𝑛𝑛 ∗ ℎ 𝑛𝑛 = ∑𝑚𝑚=−∞ 𝑥𝑥 𝑚𝑚 ℎ 𝑛𝑛 − 𝑚𝑚

Input
Second 1 0 3 8 7 -4 9 3 -7
Impulse response
First ∗
0 1 2

0.5 0 1.5+1 4+0 3.5+3 -2+8 4.5+7 1.5-4 -3.5+9 3 -7

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution

 ∞
 Convolution: 𝑥𝑥 𝑛𝑛 ∗ ℎ 𝑛𝑛 = ∑𝑚𝑚=−∞ 𝑥𝑥 𝑚𝑚 ℎ 𝑛𝑛 − 𝑚𝑚

Input
Second 1 0 3 8 7 -4 9 3 -7
Impulse response
First ∗
0 1 2

0.5 0 1.5+1 4+0 3.5+3 -2+8 4.5+7 1.5-4 -3.5+9 3 -7

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution

 ∞
 Convolution: 𝑥𝑥 𝑛𝑛 ∗ ℎ 𝑛𝑛 = ∑𝑚𝑚=−∞ 𝑥𝑥 𝑚𝑚 ℎ 𝑛𝑛 − 𝑚𝑚

Input
Second 1 0 3 8 7 -4 9 3 -7
Impulse response
First ∗
0 1 2

0.5 0 2.5 4 6.5 6 11.5 -2.5 5.5 3 -7

Output

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution

Example Problem #1

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 45

Convolution
 Linear and Time-Invariant (LTI) System
 Consider the system with input-output relationship:
1
𝑦𝑦 𝑛𝑛 = 𝑥𝑥 𝑛𝑛 + 𝑥𝑥 𝑛𝑛 − 1
2
 Compute response of the system to input of
𝑥𝑥 𝑛𝑛 = 2 𝛿𝛿 𝑛𝑛 − 1 + 2 𝛿𝛿 𝑛𝑛 − 2

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 46

Convolution
 Linear and Time-Invariant (LTI) System
 Consider the system with input-output relationship:
1
𝑦𝑦 𝑛𝑛 = 𝑥𝑥 𝑛𝑛 + 𝑥𝑥 𝑛𝑛 − 1
2
 Compute response of the system to input of
𝑥𝑥 𝑛𝑛 = 2 𝛿𝛿 𝑛𝑛 − 1 + 2 𝛿𝛿 𝑛𝑛 − 2
 Solution:
1
𝑦𝑦 𝑛𝑛 = 2 𝛿𝛿 𝑛𝑛 − 1 + 2 𝛿𝛿 𝑛𝑛 − 2 + 2 𝛿𝛿 𝑛𝑛 − 2 + 2 𝛿𝛿 𝑛𝑛 − 3
2
1
= 2 𝛿𝛿 𝑛𝑛 − 1 + 4 𝛿𝛿 𝑛𝑛 − 2 + 2 𝛿𝛿 𝑛𝑛 − 3
2

= 𝛿𝛿 𝑛𝑛 − 1 + 2 𝛿𝛿 𝑛𝑛 − 2 + 𝛿𝛿 𝑛𝑛 − 3

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 47

Convolution
 Linear and Time-Invariant (LTI) System
 Consider the system with input-output relationship:
1
𝑦𝑦 𝑛𝑛 = 𝑥𝑥 𝑛𝑛 + 𝑥𝑥 𝑛𝑛 − 1
2
 Compute response of the system to input of
𝑥𝑥 𝑛𝑛 = 2 𝛿𝛿 𝑛𝑛 − 1 + 2 𝛿𝛿 𝑛𝑛 − 2
 Solution:
1
𝑦𝑦 𝑛𝑛 = 2 𝛿𝛿 𝑛𝑛 − 1 + 2 𝛿𝛿 𝑛𝑛 − 2 + 2 𝛿𝛿 𝑛𝑛 − 2 + 2 𝛿𝛿 𝑛𝑛 − 3
2
1 Output
= 2 𝛿𝛿 𝑛𝑛 − 1 + 4 𝛿𝛿 𝑛𝑛 − 2 + 2 𝛿𝛿 𝑛𝑛 − 3 y 𝑛𝑛
2 2
1
= 𝛿𝛿 𝑛𝑛 − 1 + 2 𝛿𝛿 𝑛𝑛 − 2 + 𝛿𝛿 𝑛𝑛 − 3

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 48

Convolution
 Linear and Time-Invariant (LTI) System
 Consider the system with input-output relationship:
1
𝑦𝑦 𝑛𝑛 = 𝑥𝑥 𝑛𝑛 + 𝑥𝑥 𝑛𝑛 − 1
2
 Compute the impulse response ℎ 𝑛𝑛 of the system

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 49

Convolution
 Linear and Time-Invariant (LTI) System
 Consider the system with input-output relationship:
1
𝑦𝑦 𝑛𝑛 = 𝑥𝑥 𝑛𝑛 + 𝑥𝑥 𝑛𝑛 − 1
2
 Compute the impulse response ℎ 𝑛𝑛 of the system
 Solution:
𝑥𝑥 𝑛𝑛 = 𝛿𝛿 𝑛𝑛
1
ℎ 𝑛𝑛 = 𝛿𝛿 𝑛𝑛 + 𝛿𝛿 𝑛𝑛 − 1
2

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 50

Convolution
 Linear and Time-Invariant (LTI) System
 Consider the system with input-output relationship:
1
𝑦𝑦 𝑛𝑛 = 𝑥𝑥 𝑛𝑛 + 𝑥𝑥 𝑛𝑛 − 1
2
 What do you think this system does?

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 51

Convolution
 Linear and Time-Invariant (LTI) System
 Consider the system with input-output relationship:
1
𝑦𝑦 𝑛𝑛 = 𝑥𝑥 𝑛𝑛 + 𝑥𝑥 𝑛𝑛 − 1
2
 Compute the convolution of:
𝑦𝑦 𝑛𝑛 = ℎ 𝑛𝑛 ∗ 𝑥𝑥 𝑛𝑛
ℎ 𝑛𝑛 = 1/2 𝛿𝛿 𝑛𝑛 + 𝛿𝛿 𝑛𝑛 − 1
𝑥𝑥 𝑛𝑛 = 2 𝛿𝛿 𝑛𝑛 − 1 + 2 𝛿𝛿 𝑛𝑛 − 2

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 53

Convolution
 Compute the convolution of:
𝑦𝑦 𝑛𝑛 = ℎ 𝑛𝑛 ∗ 𝑥𝑥 𝑛𝑛
ℎ 𝑛𝑛 = 1/2 𝛿𝛿 𝑛𝑛 + 𝛿𝛿 𝑛𝑛 − 1
𝑥𝑥 𝑛𝑛 = 2 𝛿𝛿 𝑛𝑛 − 1 + 2 𝛿𝛿 𝑛𝑛 − 2

Input
𝑥𝑥 𝑛𝑛 Impulse response
ℎ 𝑚𝑚
2
∗
1/2

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution

Convolution
 Compute the convolution of: Step 1: Time-reverse a signal
𝑦𝑦 𝑛𝑛 = ℎ 𝑛𝑛 ∗ 𝑥𝑥 𝑛𝑛
ℎ 𝑛𝑛 = 1/2 𝛿𝛿 𝑛𝑛 + 𝛿𝛿 𝑛𝑛 − 1
𝑥𝑥 𝑛𝑛 = 2 𝛿𝛿 𝑛𝑛 − 1 + 2 𝛿𝛿 𝑛𝑛 − 2
Time-reversed and shifted signal
ℎ 0 − 𝑚𝑚

Input
𝑥𝑥 𝑛𝑛 Impulse response
ℎ 𝑚𝑚
2
∗
1/2

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution

Convolution
 Compute the convolution of: Step 1: Time-reverse a signal
𝑦𝑦 𝑛𝑛 = ℎ 𝑛𝑛 ∗ 𝑥𝑥 𝑛𝑛 Step 2: Shift that signal by n
Step 3: Multiply the signals & sum the result
ℎ 𝑛𝑛 = 1/2 𝛿𝛿 𝑛𝑛 + 𝛿𝛿 𝑛𝑛 − 1
𝑥𝑥 𝑛𝑛 = 2 𝛿𝛿 𝑛𝑛 − 1 + 2 𝛿𝛿 𝑛𝑛 − 2
Time-reversed and shifted signal
ℎ 0 − 𝑚𝑚

Input
𝑥𝑥 𝑛𝑛 Impulse response
ℎ 𝑚𝑚
2
∗
1/2

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution

Convolution
 Compute the convolution of: Step 1: Time-reverse a signal
𝑦𝑦 𝑛𝑛 = ℎ 𝑛𝑛 ∗ 𝑥𝑥 𝑛𝑛 Step 2: Shift that signal by n
Step 3: Multiply the signals & sum the result
ℎ 𝑛𝑛 = 1/2 𝛿𝛿 𝑛𝑛 + 𝛿𝛿 𝑛𝑛 − 1 Step 4: Assign the sum to y[n] for shift n

𝑥𝑥 𝑛𝑛 = 2 𝛿𝛿 𝑛𝑛 − 1 + 2 𝛿𝛿 𝑛𝑛 − 2
Time-reversed and shifted signal
Output ℎ 0 − 𝑚𝑚
y 𝑛𝑛
2 2
1

Input
𝑥𝑥 𝑛𝑛 Impulse response
ℎ 𝑚𝑚
2
∗
1/2

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution

Convolution
 Compute the convolution of: Step 1: Time-reverse a signal
𝑦𝑦 𝑛𝑛 = ℎ 𝑛𝑛 ∗ 𝑥𝑥 𝑛𝑛 Step 2: Shift that signal by n
Step 3: Multiply the signals & sum the result
ℎ 𝑛𝑛 = 1/2 𝛿𝛿 𝑛𝑛 + 𝛿𝛿 𝑛𝑛 − 1 Step 4: Assign the sum to y[n] for shift n
Step 5: Repeat for all shifts −∞ < 𝑛𝑛 < ∞
𝑥𝑥 𝑛𝑛 = 2 𝛿𝛿 𝑛𝑛 − 1 + 2 𝛿𝛿 𝑛𝑛 − 2

 Output is zero for all Time-reversed and shifted signal

Output ℎ 0 − 𝑚𝑚
y 𝑛𝑛
remaining right-shifts
2 2
1

Input
𝑥𝑥 𝑛𝑛 Impulse response
ℎ 𝑚𝑚
2
∗
1/2

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution

Convolution
 Compute the convolution of: Step 1: Time-reverse a signal
𝑦𝑦 𝑛𝑛 = ℎ 𝑛𝑛 ∗ 𝑥𝑥 𝑛𝑛 Step 2: Shift that signal by n
Step 3: Multiply the signals & sum the result
ℎ 𝑛𝑛 = 1/2 𝛿𝛿 𝑛𝑛 + 𝛿𝛿 𝑛𝑛 − 1 Step 4: Assign the sum to y[n] for shift n
Step 5: Repeat for all shifts −∞ < 𝑛𝑛 < ∞
𝑥𝑥 𝑛𝑛 = 2 𝛿𝛿 𝑛𝑛 − 1 + 2 𝛿𝛿 𝑛𝑛 − 2
Time-reversed and shifted signal
Output ℎ 1 − 𝑚𝑚
y 𝑛𝑛
2 2
1

Input
𝑥𝑥 𝑛𝑛 Impulse response
ℎ 𝑚𝑚
2
∗
1/2

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution

Convolution
 Compute the convolution of: Step 1: Time-reverse a signal
𝑦𝑦 𝑛𝑛 = ℎ 𝑛𝑛 ∗ 𝑥𝑥 𝑛𝑛 Step 2: Shift that signal by n
Step 3: Multiply the signals & sum the result
ℎ 𝑛𝑛 = 1/2 𝛿𝛿 𝑛𝑛 + 𝛿𝛿 𝑛𝑛 − 1 Step 4: Assign the sum to y[n] for shift n
Step 5: Repeat for all shifts −∞ < 𝑛𝑛 < ∞
𝑥𝑥 𝑛𝑛 = 2 𝛿𝛿 𝑛𝑛 − 1 + 2 𝛿𝛿 𝑛𝑛 − 2
Time-reversed and shifted signal
Output ℎ 2 − 𝑚𝑚
y 𝑛𝑛
2 2
1

Input
𝑥𝑥 𝑛𝑛 Impulse response
ℎ 𝑚𝑚
2
∗
1/2

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution

Convolution
 Compute the convolution of: Step 1: Time-reverse a signal
𝑦𝑦 𝑛𝑛 = ℎ 𝑛𝑛 ∗ 𝑥𝑥 𝑛𝑛 Step 2: Shift that signal by n
Step 3: Multiply the signals & sum the result
ℎ 𝑛𝑛 = 1/2 𝛿𝛿 𝑛𝑛 + 𝛿𝛿 𝑛𝑛 − 1 Step 4: Assign the sum to y[n] for shift n
Step 5: Repeat for all shifts −∞ < 𝑛𝑛 < ∞
𝑥𝑥 𝑛𝑛 = 2 𝛿𝛿 𝑛𝑛 − 1 + 2 𝛿𝛿 𝑛𝑛 − 2
Time-reversed and shifted signal
Output ℎ 2 − 𝑚𝑚
y 𝑛𝑛
2 2
1

Input
𝑥𝑥 𝑛𝑛 Impulse response
ℎ 𝑚𝑚
2
∗
1/2

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution

Convolution
 Compute the convolution of: Step 1: Time-reverse a signal
𝑦𝑦 𝑛𝑛 = ℎ 𝑛𝑛 ∗ 𝑥𝑥 𝑛𝑛 Step 2: Shift that signal by n
Step 3: Multiply the signals & sum the result
ℎ 𝑛𝑛 = 1/2 𝛿𝛿 𝑛𝑛 + 𝛿𝛿 𝑛𝑛 − 1 Step 4: Assign the sum to y[n] for shift n
Step 5: Repeat for all shifts −∞ < 𝑛𝑛 < ∞
𝑥𝑥 𝑛𝑛 = 2 𝛿𝛿 𝑛𝑛 − 1 + 2 𝛿𝛿 𝑛𝑛 − 2
Time-reversed and shifted signal
Output ℎ 3 − 𝑚𝑚
y 𝑛𝑛
2 2
1

Input
𝑥𝑥 𝑛𝑛 Impulse response
ℎ 𝑚𝑚
2
∗
1/2

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution

Convolution
 Compute the convolution of: Step 1: Time-reverse a signal
𝑦𝑦 𝑛𝑛 = ℎ 𝑛𝑛 ∗ 𝑥𝑥 𝑛𝑛 Step 2: Shift that signal by n
Step 3: Multiply the signals & sum the result
ℎ 𝑛𝑛 = 1/2 𝛿𝛿 𝑛𝑛 + 𝛿𝛿 𝑛𝑛 − 1 Step 4: Assign the sum to y[n] for shift n
Step 5: Repeat for all shifts −∞ < 𝑛𝑛 < ∞
𝑥𝑥 𝑛𝑛 = 2 𝛿𝛿 𝑛𝑛 − 1 + 2 𝛿𝛿 𝑛𝑛 − 2
Time-reversed and shifted signal Output is zero for all
Output ℎ 3 − 𝑚𝑚
y 𝑛𝑛
remaining right-shifts →
2 2
1

Input
𝑥𝑥 𝑛𝑛 Impulse response
ℎ 𝑚𝑚
2
∗
1/2

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution

Convolution
 Compute the convolution of: Step 1: Time-reverse a signal
𝑦𝑦 𝑛𝑛 = ℎ 𝑛𝑛 ∗ 𝑥𝑥 𝑛𝑛 Step 2: Shift that signal by n
Step 3: Multiply the signals & sum the result
ℎ 𝑛𝑛 = 1/2 𝛿𝛿 𝑛𝑛 + 𝛿𝛿 𝑛𝑛 − 1 Step 4: Assign the sum to y[n] for shift n
Step 5: Repeat for all shifts −∞ < 𝑛𝑛 < ∞
𝑥𝑥 𝑛𝑛 = 2 𝛿𝛿 𝑛𝑛 − 1 + 2 𝛿𝛿 𝑛𝑛 − 2

Output
y 𝑛𝑛
2
1

𝑦𝑦 𝑛𝑛 = 𝛿𝛿 𝑛𝑛 − 1 + 2 𝛿𝛿 𝑛𝑛 − 2 + 𝛿𝛿 𝑛𝑛 − 3

Same as our first result!

We can compute the system output via convolution!

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution

Example Problem #2

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 65

Convolution
 Linear and Time-Invariant (LTI) System
 Consider the system with impulse response ℎ 𝑛𝑛

 Compute response of the system to input 𝑥𝑥 𝑛𝑛 below

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 66

Convolution
 Linear and Time-Invariant (LTI) System
 Solution:

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 67

Example Problem #3: Correlation

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 68

Convolution / Correlation
 Definition of convolution
∞

𝑦𝑦 𝑛𝑛 = 𝑥𝑥 𝑛𝑛 ∗ ℎ 𝑛𝑛 = � 𝑥𝑥 𝑚𝑚 ℎ 𝑛𝑛 − 𝑚𝑚
𝑚𝑚=−∞

 Definition of correlation
∞

𝑦𝑦 𝑛𝑛 = 𝑥𝑥 −𝑛𝑛 ∗ ℎ 𝑛𝑛 = � 𝑥𝑥 𝑚𝑚 ℎ 𝑛𝑛 + 𝑚𝑚
𝑚𝑚=−∞

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 69

Convolution / Correlation
 Linear and Time-Invariant (LTI) System
 Consider ℎ 𝑛𝑛 and x 𝑛𝑛 below. Compute their correlation.
𝑦𝑦 𝑛𝑛 = 𝑥𝑥 −𝑛𝑛 ∗ ℎ 𝑛𝑛

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 70

Convolution / Correlation
 Compute the correlation:
𝑦𝑦 𝑛𝑛 = 𝑥𝑥 −𝑛𝑛 ∗ ℎ 𝑛𝑛

Input signal Impulse response

x 𝑚𝑚 ℎ 𝑚𝑚

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution

Convolution / Correlation
 Compute the correlation:
𝑦𝑦 𝑛𝑛 = 𝑥𝑥 −𝑛𝑛 ∗ ℎ 𝑛𝑛

Input signal Impulse response

x 𝑚𝑚 ℎ −𝑚𝑚

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution

Convolution / Correlation
 Compute the correlation: Step 1: Time-reverse a signal
𝑦𝑦 𝑛𝑛 = 𝑥𝑥 −𝑛𝑛 ∗ ℎ 𝑛𝑛

x 0 − 𝑚𝑚

Input signal Impulse response

x 𝑚𝑚 ℎ −𝑚𝑚

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution

Convolution / Correlation
 Compute the correlation: Step 1: Time-reverse a signal
𝑦𝑦 𝑛𝑛 = 𝑥𝑥 −𝑛𝑛 ∗ ℎ 𝑛𝑛 Step 2: Shift that signal by n
Step 3: Multiply the signals & sum the result
Step 4: Assign the sum to y[n] for shift n
Step 5: Repeat for all shifts −∞ < 𝑛𝑛 < ∞

y 𝑛𝑛 x 0 − 𝑚𝑚

Input signal Impulse response

x 𝑚𝑚 ℎ −𝑚𝑚

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution

Convolution / Correlation
 Compute the correlation: Step 1: Time-reverse a signal
𝑦𝑦 𝑛𝑛 = 𝑥𝑥 −𝑛𝑛 ∗ ℎ 𝑛𝑛 Step 2: Shift that signal by n
Step 3: Multiply the signals & sum the result
Step 4: Assign the sum to y[n] for shift n
Step 5: Repeat for all shifts −∞ < 𝑛𝑛 < ∞

y 𝑛𝑛 x 0 − 𝑚𝑚

Input signal Impulse response

x 𝑚𝑚 ℎ −𝑚𝑚

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution

Convolution / Correlation
 Compute the correlation: Step 1: Time-reverse a signal
𝑦𝑦 𝑛𝑛 = 𝑥𝑥 −𝑛𝑛 ∗ ℎ 𝑛𝑛 Step 2: Shift that signal by n
Step 3: Multiply the signals & sum the result
Step 4: Assign the sum to y[n] for shift n
Step 5: Repeat for all shifts −∞ < 𝑛𝑛 < ∞

y 𝑛𝑛 x 0 − 𝑚𝑚

Input signal Impulse response

x 𝑚𝑚 ℎ −𝑚𝑚

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution

Convolution / Correlation
 Compute the correlation: Step 1: Time-reverse a signal
𝑦𝑦 𝑛𝑛 = 𝑥𝑥 −𝑛𝑛 ∗ ℎ 𝑛𝑛 Step 2: Shift that signal by n
Step 3: Multiply the signals & sum the result
Step 4: Assign the sum to y[n] for shift n
Step 5: Repeat for all shifts −∞ < 𝑛𝑛 < ∞

y 𝑛𝑛 x 1 − 𝑚𝑚

Input signal Impulse response

x 𝑚𝑚 ℎ −𝑚𝑚

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution

Convolution / Correlation
 Compute the correlation: Step 1: Time-reverse a signal
𝑦𝑦 𝑛𝑛 = 𝑥𝑥 −𝑛𝑛 ∗ ℎ 𝑛𝑛 Step 2: Shift that signal by n
Step 3: Multiply the signals & sum the result
Step 4: Assign the sum to y[n] for shift n
Step 5: Repeat for all shifts −∞ < 𝑛𝑛 < ∞
6

y 𝑛𝑛 x 2 − 𝑚𝑚

Input signal Impulse response

x 𝑚𝑚 ℎ −𝑚𝑚

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution

Convolution / Correlation
 Compute the correlation: Step 1: Time-reverse a signal
𝑦𝑦 𝑛𝑛 = 𝑥𝑥 −𝑛𝑛 ∗ ℎ 𝑛𝑛 Step 2: Shift that signal by n
Step 3: Multiply the signals & sum the result
Step 4: Assign the sum to y[n] for shift n
Step 5: Repeat for all shifts −∞ < 𝑛𝑛 < ∞
6

y 𝑛𝑛 x 3 − 𝑚𝑚

Input signal Impulse response

x 𝑚𝑚 ℎ −𝑚𝑚

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution

Convolution / Correlation
 Compute the correlation: Step 1: Time-reverse a signal
𝑦𝑦 𝑛𝑛 = 𝑥𝑥 −𝑛𝑛 ∗ ℎ 𝑛𝑛 Step 2: Shift that signal by n
Step 3: Multiply the signals & sum the result
Step 4: Assign the sum to y[n] for shift n
Step 5: Repeat for all shifts −∞ < 𝑛𝑛 < ∞
6

y 𝑛𝑛 x 4 − 𝑚𝑚

Input signal Impulse response

x 𝑚𝑚 ℎ −𝑚𝑚

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution

Convolution / Correlation
 Compute the correlation: Step 1: Time-reverse a signal
𝑦𝑦 𝑛𝑛 = 𝑥𝑥 −𝑛𝑛 ∗ ℎ 𝑛𝑛 Step 2: Shift that signal by n
Step 3: Multiply the signals & sum the result
Step 4: Assign the sum to y[n] for shift n
Step 5: Repeat for all shifts −∞ < 𝑛𝑛 < ∞
6

y 𝑛𝑛 x 5 − 𝑚𝑚

Input signal Impulse response

x 𝑚𝑚 ℎ −𝑚𝑚

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution

Convolution / Correlation
 What does this achieve? Step 1: Time-reverse a signal
Step 2: Shift that signal by n
Step 3: Multiply the signals & sum the result
Step 4: Assign the sum to y[n] for shift n
Step 5: Repeat for all shifts −∞ < 𝑛𝑛 < ∞
6

y 𝑛𝑛 x 6 − 𝑚𝑚

Input signal Impulse response

x 𝑚𝑚 ℎ −𝑚𝑚

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution

Convolution / Correlation
 Question: What are applications of correlation?

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 83

More Examples on Course Website

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 84

Lecture 5: Discrete -Time LTI Systems
Foundations of Digital Signal Processing

Outline
• Review
• Discrete-Time Convolution
• Properties of Discrete-Time Convolution
• Combining Systems
• Properties of the Impulse Response

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 85

Convolution Properties
 Definition of convolution
∞

𝑥𝑥 𝑛𝑛 ∗ ℎ 𝑛𝑛 = � 𝑥𝑥 𝑚𝑚 ℎ 𝑛𝑛 − 𝑚𝑚
𝑚𝑚=−∞

 Property #1: Commutativity

◊ 𝑥𝑥 𝑛𝑛 ∗ ℎ 𝑛𝑛 = ℎ 𝑛𝑛 ∗ 𝑥𝑥 𝑛𝑛
 Property #2: Associativity
◊ 𝑥𝑥 𝑛𝑛 ∗ ℎ 𝑛𝑛 ∗ 𝑔𝑔 𝑛𝑛 = 𝑥𝑥 𝑛𝑛 ∗ ℎ 𝑛𝑛 ∗ 𝑔𝑔 𝑛𝑛
 Property #3: Distributivity
◊ 𝑥𝑥 𝑛𝑛 ∗ ℎ 𝑛𝑛 + 𝑔𝑔 𝑛𝑛 = 𝑥𝑥 𝑛𝑛 ∗ ℎ 𝑛𝑛 + 𝑥𝑥 𝑛𝑛 ∗ 𝑔𝑔 𝑛𝑛
 Property #4: Multiplicative identity
◊ 𝑥𝑥 𝑛𝑛 ∗ 𝛿𝛿 𝑛𝑛 = 𝑥𝑥 𝑛𝑛
Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 86
Convolution Properties
 Definition of convolution
∞

𝑥𝑥 𝑛𝑛 ∗ ℎ 𝑛𝑛 = � 𝑥𝑥 𝑚𝑚 ℎ 𝑛𝑛 − 𝑚𝑚
𝑚𝑚=−∞

 Property #5: Shifting property

◊ 𝑥𝑥 𝑛𝑛 ∗ 𝛿𝛿 𝑛𝑛 − 𝑁𝑁 = 𝑥𝑥 𝑛𝑛 − 𝑁𝑁

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 87

Lecture 5: Discrete -Time LTI Systems
Foundations of Digital Signal Processing

Outline
• Review
• Discrete-Time Convolution
• Properties of Discrete-Time Convolution
• Combining Systems
• Properties of the Impulse Response

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 88

Combining Systems
 Basic System Block Diagram

𝑥𝑥 𝑛𝑛 ℎ 𝑛𝑛 𝑦𝑦 𝑛𝑛 = 𝑥𝑥 𝑛𝑛 ∗ ℎ 𝑛𝑛

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 89

Combining Systems
 Basic System Block Diagram

𝑥𝑥 𝑛𝑛 ℎ 𝑛𝑛 𝑦𝑦 𝑛𝑛 = 𝑥𝑥 𝑛𝑛 ∗ ℎ 𝑛𝑛

 Cascading System (Systems in Series)

𝑥𝑥 𝑛𝑛 ℎ 𝑛𝑛 𝑔𝑔 𝑛𝑛 𝑦𝑦 𝑛𝑛 = 𝑥𝑥 𝑛𝑛 ∗ ℎ 𝑛𝑛 ∗ 𝑔𝑔 𝑛𝑛

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 90

Combining Systems
 Basic System Block Diagram

𝑥𝑥 𝑛𝑛 ℎ 𝑛𝑛 𝑦𝑦 𝑛𝑛 = 𝑥𝑥 𝑛𝑛 ∗ ℎ 𝑛𝑛

 Systems in Parallel

𝑥𝑥 𝑛𝑛 ℎ 𝑛𝑛 + 𝑦𝑦 𝑛𝑛 = 𝑥𝑥 𝑛𝑛 ∗ ℎ 𝑛𝑛 + 𝑔𝑔 𝑛𝑛
= 𝑥𝑥 𝑛𝑛 ∗ ℎ 𝑛𝑛 + 𝑥𝑥 𝑛𝑛 ∗ 𝑔𝑔 𝑛𝑛

𝑔𝑔 𝑛𝑛

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 91

Combining Systems
 Problem: Compute the impulse response of the system below:

𝑥𝑥 𝑛𝑛 ℎ 𝑛𝑛 + 𝑟𝑟 𝑛𝑛 𝑦𝑦 𝑛𝑛

𝑔𝑔 𝑛𝑛

 ℎ 𝑛𝑛 = 𝛿𝛿 𝑛𝑛 − 2 , g 𝑛𝑛 = 𝛿𝛿 𝑛𝑛 − 1 , r 𝑛𝑛 = 𝛿𝛿 𝑛𝑛 − 1

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 92

Combining Systems
 Problem: Compute the impulse response of the system below:

𝑥𝑥 𝑛𝑛 ℎ 𝑛𝑛 + 𝑟𝑟 𝑛𝑛 𝑦𝑦 𝑛𝑛

𝑔𝑔 𝑛𝑛

 ℎ 𝑛𝑛 = 𝛿𝛿 𝑛𝑛 − 2 , g 𝑛𝑛 = 𝛿𝛿 𝑛𝑛 − 1 , r 𝑛𝑛 = 𝛿𝛿 𝑛𝑛 − 1

 Solution
 𝑦𝑦 𝑛𝑛 = 𝛿𝛿 𝑛𝑛 ∗ 𝛿𝛿 𝑛𝑛 − 2 + 𝛿𝛿 𝑛𝑛 − 1 ∗ 𝛿𝛿 𝑛𝑛 − 1

 𝑦𝑦 𝑛𝑛 = 𝛿𝛿 𝑛𝑛 − 2 + 𝛿𝛿 𝑛𝑛 − 1

Foundations of Digital Signal Processing Lecture 5: Discrete-Time Convolution 93