Lecture 6: Thu Sep 3, 2020 · 2020. 9. 8. · 0 Lecture 6: Thu Sep 3, 2020 Reminder: •HW2 due...

15
Lecture 6: Thu Sep 3, 2020 Reminder: HW2 due midnight tonight. HW3 posted later today. Lecture convolution: more properties convolution examples

Transcript of Lecture 6: Thu Sep 3, 2020 · 2020. 9. 8. · 0 Lecture 6: Thu Sep 3, 2020 Reminder: •HW2 due...

Page 1: Lecture 6: Thu Sep 3, 2020 · 2020. 9. 8. · 0 Lecture 6: Thu Sep 3, 2020 Reminder: •HW2 due midnight tonight. •HW3 posted later today. Lecture •convolution: more properties

0

Lecture 6: Thu Sep 3, 2020

Reminder:

• HW2 due midnight tonight.

• HW3 posted later today.

Lecture

• convolution: more properties

• convolution examples

Page 2: Lecture 6: Thu Sep 3, 2020 · 2020. 9. 8. · 0 Lecture 6: Thu Sep 3, 2020 Reminder: •HW2 due midnight tonight. •HW3 posted later today. Lecture •convolution: more properties

1

Reminder: Reading Assignment

(TODAY)

Page 3: Lecture 6: Thu Sep 3, 2020 · 2020. 9. 8. · 0 Lecture 6: Thu Sep 3, 2020 Reminder: •HW2 due midnight tonight. •HW3 posted later today. Lecture •convolution: more properties

2

Reading Assignment

(TODAY)

Page 4: Lecture 6: Thu Sep 3, 2020 · 2020. 9. 8. · 0 Lecture 6: Thu Sep 3, 2020 Reminder: •HW2 due midnight tonight. •HW3 posted later today. Lecture •convolution: more properties

3

Reminder: Videos from Summerhttps://www.youtube.com/playlist?list=PLOunECWxELQRYwsuj4BL4Hu1nvj9dxRQ6

Over 60 minutes of convolution examples:

Page 5: Lecture 6: Thu Sep 3, 2020 · 2020. 9. 8. · 0 Lecture 6: Thu Sep 3, 2020 Reminder: •HW2 due midnight tonight. •HW3 posted later today. Lecture •convolution: more properties

4

Why Impulse Response is ImportantAn LTI system is completely characterized by h( t ).

Response to any input can be found by convolving it with h( t ):

x( t ) y( t ) = x( t ) * h( t ) LTIh( t )

Page 6: Lecture 6: Thu Sep 3, 2020 · 2020. 9. 8. · 0 Lecture 6: Thu Sep 3, 2020 Reminder: •HW2 due midnight tonight. •HW3 posted later today. Lecture •convolution: more properties

5

Convolution Properties

• commutative: x( t ) h( t ) = h( t ) x( t ) X• associative (x( t ) h( t )) z( t ) = x( t ) (h( t ) z( t )) X

• Convolving with an impulse:

x( t ) ( t ) = x( t ) X x( t ) ( t – t) = x( t– t) X

• Delay property: x( t ) h( t – t) = x( t – t) h( t )

• Derivative: (x( t ) h( t )) = ( x( t )) h( t ) = x( t ) * ( h( t ))

y( t ) =Z–∞

∞x( )h( t – )d

ddt----- d

dt----- d

dt-----

Page 7: Lecture 6: Thu Sep 3, 2020 · 2020. 9. 8. · 0 Lecture 6: Thu Sep 3, 2020 Reminder: •HW2 due midnight tonight. •HW3 posted later today. Lecture •convolution: more properties

6

Pop QuizFind the “step response” of a filter with an exponential impulse response:

h( t ) = e–tu( t )

UNIT STEP y( t ) = ?

u( t )

0 1

1

h( t )

t

Page 8: Lecture 6: Thu Sep 3, 2020 · 2020. 9. 8. · 0 Lecture 6: Thu Sep 3, 2020 Reminder: •HW2 due midnight tonight. •HW3 posted later today. Lecture •convolution: more properties

7

Pop QuizFind the “step response” of a filter with an exponential impulse response:

h( t ) = e–tu( t )

UNIT STEP y( t ) = (1 – e–t)u( t )

u( t )

0 1

1

h( t )

t

y( t )

0 1

1

t

Page 9: Lecture 6: Thu Sep 3, 2020 · 2020. 9. 8. · 0 Lecture 6: Thu Sep 3, 2020 Reminder: •HW2 due midnight tonight. •HW3 posted later today. Lecture •convolution: more properties

8

Sketch Convolution of x(t) with h(t):

0 2 4 7

10

h( t ) = 10e–10tu( t )

1

1

2

3

t

x( t )

Page 10: Lecture 6: Thu Sep 3, 2020 · 2020. 9. 8. · 0 Lecture 6: Thu Sep 3, 2020 Reminder: •HW2 due midnight tonight. •HW3 posted later today. Lecture •convolution: more properties

9

Sketch Convolution of x(t) with h(t):

10 2 4 7

10

h( t ) = 10e–10tu( t )

1

2

3

t

Page 11: Lecture 6: Thu Sep 3, 2020 · 2020. 9. 8. · 0 Lecture 6: Thu Sep 3, 2020 Reminder: •HW2 due midnight tonight. •HW3 posted later today. Lecture •convolution: more properties

10

Convolution Tips

• graphical approach mimics convolution integral

• graphs are plotted versus dummy variable

• convolution is commutative ⇒ pick “simpler” signal to flip!

• Getting the right limits of integration is 90% of job, integration itself is easy

• Use cconvdemo to learn the basics

y( t ) =Z–∞

∞x( )h( t – )d

Page 12: Lecture 6: Thu Sep 3, 2020 · 2020. 9. 8. · 0 Lecture 6: Thu Sep 3, 2020 Reminder: •HW2 due midnight tonight. •HW3 posted later today. Lecture •convolution: more properties

11

Example of ConvolutionConvolve these two rectangles:

1 2

31

1

1

t

t

Page 13: Lecture 6: Thu Sep 3, 2020 · 2020. 9. 8. · 0 Lecture 6: Thu Sep 3, 2020 Reminder: •HW2 due midnight tonight. •HW3 posted later today. Lecture •convolution: more properties

12

Answer is a Trapezoid

One rectangle floats past another

Once they begin to overlap, amount of overlap increases linearly with time ⇒ ramp up

While they overlap fully, the signal plateaus

When the amount of overlap decreases it decreases linearly ⇒ ramp back down

* =1 231

11

tt 2 3

1

t4 5

Page 14: Lecture 6: Thu Sep 3, 2020 · 2020. 9. 8. · 0 Lecture 6: Thu Sep 3, 2020 Reminder: •HW2 due midnight tonight. •HW3 posted later today. Lecture •convolution: more properties

13

Fun Convolution Facts• convolving two steps yields a ramp

• convolving a rect with itself yields a triangle

• convolving rects of different widths yields a trapezoid

• duration of convolution is sum of durations of each

• convolving a signal that starts at time t1 with another starting at t2 yields a signal that starts at t1 + t2

Page 15: Lecture 6: Thu Sep 3, 2020 · 2020. 9. 8. · 0 Lecture 6: Thu Sep 3, 2020 Reminder: •HW2 due midnight tonight. •HW3 posted later today. Lecture •convolution: more properties

14

Example 1:

* = ?e–tu(t)

0 t

e–tu(t)

0 t