6.

003: Signals and Systems

Convolution

October 4, 2011
1

Mid-term Examination #1
Wednesday, October 5, 7:30-9:30pm.

No recitations on the day of the exam. Coverage: CT and DT Systems, Z and Laplace Transforms Lectures 17 Recitations 17 Homeworks 14

Homework 4 will not collected or graded. Solutions are posted. Closed book: 1 page of notes (8 1 2 11 inches; front and back). No calculators, computers, cell phones, music players, or other aids. Designed as 1-hour exam; two hours to complete. Prior term midterm exams have been posted on the 6.003 website.
2

Multiple Representations of CT and DT Systems

Verbal descriptions: preserve the rationale. Dierence/dierential equations: mathematically compact. y [n] = x[n] + z0 y [n 1] y (t) = x(t) + s0 y (t)

Block diagrams: illustrate signal ow paths.

+ z0 R

A s0

Operator representations: analyze systems as polynomials. Y 1 Y A = = X 1 z0 R X 1 s0 A Transforms: representing di. equations with algebraic equations. z 1 H (z ) = H (s) = s s0 z z0
3

Convolution
Representing a system by a single signal.

Responses to arbitrary signals

Although we have focused on responses to simple signals ( [n], (t)) we are generally interested in responses to more complicated signals. How do we compute responses to a more complicated input signals? No problem for dierence equations / block diagrams. use step-by-step analysis.

Check Yourself

Example: Find y [3]

X R
when the input is

+ R

x[n] 1 n

1. 1

2. 2

3. 3

4. 4

5. 5

0. none of the above

6

Responses to arbitrary signals

Example.

0 R

+ 0 R

+ 0

x[n] n

y [n] n

Responses to arbitrary signals

Example.

1 R

+ 0 R

+ 0

x[n] n

y [n] n

Responses to arbitrary signals

Example.

1 R

+ 1 R

+ 0

x[n] n

y [n] n

Responses to arbitrary signals

Example.

1 R

+ 1 R

+ 1

x[n] n

y [n] n

10

Responses to arbitrary signals

Example.

0 R

+ 1 R

+ 1

x[n] n

y [n] n

11

Responses to arbitrary signals

Example.

0 R

+ 0 R

+ 1

x[n] n

y [n] n

12

Responses to arbitrary signals

Example.

0 R

+ 0 R

+ 0

x[n] n

y [n] n

13

Check Yourself

What is y [3]?

0 R

+ 0 R

+ 0

x[n] 1 n

y [n] n

1. 1

2. 2

3. 3

4. 4

5. 5

0. none of the above

14

Superposition
Break input into additive parts and sum the responses to the parts.

x[n] n = n + n + n 1 0 1 2 3 4 5 = n 1 0 1 2 3 4 5 + n + n n y [n]

15

Linearity
A system is linear if its response to a weighted sum of inputs is equal to the weighted sum of its responses to each of the inputs. Given

x1 [n]
and

system

y1 [n]

x2 [n]

system

y2 [n]

the system is linear if

x1 [n] + x2 [n]
is true for all and .

system

y1 [n] + y2 [n]

16

Superposition
Break input into additive parts and sum the responses to the parts.

x[n] n = n + n + n 1 0 1 2 3 4 5 = n 1 0 1 2 3 4 5
Superposition works if the system is linear.
17

y [n] n + n + n

Superposition
Break input into additive parts and sum the responses to the parts.

x[n] n = n + n + n 1 0 1 2 3 4 5 = n 1 0 1 2 3 4 5
Reponses to parts are easy to compute if system is time-invariant.
18

y [n] n + n + n

Time-Invariance
A system is time-invariant if delaying the input to the system simply delays the output by the same amount of time. Given

x[n]
the system is time invariant if

system

y [n]

x[n n0 ]
is true for all n0 .

system

y [ n n0 ]

19

Superposition
Break input into additive parts and sum the responses to the parts.

x[n] n = n + n + n 1 0 1 2 3 4 5 = n 1 0 1 2 3 4 5
Superposition is easy if the system is linear and time-invariant.
20

y [n] n + n + n

Structure of Superposition
If a system is linear and time-invariant (LTI) then its output is the sum of weighted and shifted unit-sample responses.

[n]

system

h[n]

[n k ]

system

h[n k ]

x[k ] [n k ]

system

x[k ]h[n k ]

x[n] =
k=

x[k ] [n k ]

system

y [n] =
k=

x[k ]h[n k ]

21

Convolution
Response of an LTI system to an arbitrary input.

x[n]

LTI

y [n]

y [n] =
k=

x[k ]h[n k ] (x h)[n]

This operation is called convolution.

22

Notation
Convolution is represented with an asterisk.
0 k=

x[k ]h[n k ] (x h)[n]

It is customary (but confusing) to abbreviate this notation: (x h)[n] = x[n] h[n]

23

Notation
Do not be fooled by the confusing notation. Confusing (but conventional) notation: 0 x[k ]h[n k ] = x[n] h[n]
k=

x[n] h[n] looks like an operation of samples; but it is not! x[1] h[1] = (x h)[1] Convolution operates on signals not samples. Unambiguous notation: 0 x[k ]h[n k ] (x h)[n]
k=

The symbols x and h represent DT signals. Convolving x with h generates a new DT signal x h.
24

Structure of Convolution
y [n] =
0 k=

x[k ]h[n k ]

x[n] n 21 0 1 2 3 4 5

h[n] n 21 0 1 2 3 4 5

25

Structure of Convolution
y [0] =
0 k=

x[k ]h[0 k ]

x[n] n 21 0 1 2 3 4 5

h[n] n 21 0 1 2 3 4 5

26

Structure of Convolution
y [0] =
0 k=

x[k ]h[0 k ]

x[k ] 21 0 1 2 3 4 5 k

h[k ] k

21 0 1 2 3 4 5

27

Structure of Convolution
y [0] =
0 k=

x[k ]h[0 k ]

x[k ] k

ip

h[k ] k

21 0 1 2 3 4 5

21 0 1 2 3 4 5 h[k ]

21 0 1 2 3 4 5

28

Structure of Convolution
y [0] =
0 k=

x[k ]h[0 k ]

x[k ] k

shift

h[k ] k

21 0 1 2 3 4 5

21 0 1 2 3 4 5 h[0 k ]

21 0 1 2 3 4 5

29

Structure of Convolution
y [0] =
0 k=

x[k ]h[0 k ]

x[k ]

multiply k

h[k ] k

21 0 1 2 3 4 5 h[0 k ]

21 0 1 2 3 4 5 h[0 k ]

21 0 1 2 3 4 5

21 0 1 2 3 4 5

30

Structure of Convolution
y [0] =
0 k=

x[k ]h[0 k ]

x[k ]

multiply k

h[k ] k

h[0 k ] k x[k ]h[0 k ] k

21 0 1 2 3 4 5 h[0 k ]

21 0 1 2 3 4 5

21 0 1 2 3 4 5

31

Structure of Convolution
y [0] =
0 k=

x[k ]h[0 k ]

x[k ] k h[0 k ] k x[k ]h[0 k ]

k=21

sum

h[k ] k h[0 k ] k

21 0 1 2 3 4 5

0 1 2 3 4 5

32

Structure of Convolution
y [0] =
0 k=

x[k ]h[0 k ]

x[k ] k h[0 k ] k x[k ]h[0 k ]

k=21

h[k ] k h[0 k ] k

21 0 1 2 3 4 5

0 1 2 3 4 5

=1

33

Structure of Convolution
y [1] =
0 k=

x[k ]h[1 k ]

x[k ] k h[1 k ] k x[k ]h[1 k ]

k=21

h[k ] k h[1 k ] k

21 0 1 2 3 4 5

0 1 2 3 4 5

=2

34

Structure of Convolution
y [2] =
0 k=

x[k ]h[2 k ]

x[k ] k h[2 k ] k x[k ]h[2 k ]

k=21

h[k ] k h[2 k ] k

21 0 1 2 3 4 5

0 1 2 3 4 5

=3

35

Structure of Convolution
y [3] =
0 k=

x[k ]h[3 k ]

x[k ] k h[3 k ] k x[k ]h[3 k ]

k=21

h[k ] k h[3 k ] k

21 0 1 2 3 4 5

0 1 2 3 4 5

=2

36

Structure of Convolution
y [4] =
0 k=

x[k ]h[4 k ]

x[k ] k h[4 k ] k x[k ]h[4 k ]

k=21

h[k ] k h[4 k ] k

21 0 1 2 3 4 5

0 1 2 3 4 5

=1

37

Structure of Convolution
y [5] =
0 k=

x[k ]h[5 k ]

x[k ] k h[5 k ] k x[k ]h[5 k ]

k=21

h[k ] k h[5 k ] k

21 0 1 2 3 4 5

0 1 2 3 4 5

=0

38

Check Yourself

Which plot shows the result of the convolution above?

1
1. 2.

1
3. 4.

5. none of the above

39

Check Yourself
1

=
0

Express mathematically: 2 n 2 n u[n] u[n] 3 3

k= n 0 k=0 n 0 k=0

2 k u[k ] 3

2 nk u[n k ] 3

2 k 3

2 nk 3
n 2 n0 1 3 k=0

2 n = 3 2 3
n

= (n + 1) = 1,

u[n]

4 4 32 80 , , , , ... 3 3 27 81

40

Check Yourself

Which plot shows the result of the convolution above? 3

1
1. 2.

1
3. 4.

5. none of the above

41

DT Convolution: Summary
Representing an LTI system by a single signal.

x[n]

h[n]

y [n]

Unit-sample response h[n] is a complete description of an LTI system. Given h[n] one can compute the response y [n] to any arbitrary input signal x[n]: y [n] = (x h)[n]
0 k=

x[k ]h[n k ]

42

CT Convolution
The same sort of reasoning applies to CT signals.

x(t)

t
x(t) = lim where 0
k 0

x(k )p(t k )

p(t)
1

As 0, k , d , and p(t) (t):

x(t)

x( ) (t )d
43

Structure of Superposition
If a system is linear and time-invariant (LTI) then its output is the integral of weighted and shifted unit-impulse responses.

(t)

system

h(t)

(t )

system

h(t )

x( ) (t )

system

x( )h(t )

x(t) =

x( ) (t )d

system

y (t) =

x( )h(t )d

44

CT Convolution
Convolution of CT signals is analogous to convolution of DT signals. DT: y [n] = (x h)[n] =
0 k=

x[k ]h[n k ]

CT: y (t) = (x h)(t) =

x( )h(t )d

45

Check Yourself
et u(t) et u(t)

Which plot shows the result of the convolution above?

1.

2.

3.

4.

5. none of the above

46

Check Yourself
Which plot shows the result of the following convolution?

et u(t)

et u(t)

t
et u(t) et u(t) = =
0 t

e u( )e(t ) u(t )d
t 0

e e(t ) d = et

d = tet u(t)

47

Check Yourself
et u(t) et u(t)

t
4

Which plot shows the result of the convolution above?

1.

2.

3.

4.

5. none of the above

48

Convolution
Convolution is an important computational tool. Example: characterizing LTI systems Determine the unit-sample response h[n]. Calculate the output for an arbitrary input using convolution: 0 y [n] = (x h)[n] = x[k ]h[n k ]

49

Applications of Convolution
Convolution is an important conceptual tool: it provides an important new way to think about the behaviors of systems. Example systems: microscopes and telescopes.

50

Microscope
Images from even the best microscopes are blurred.
CCD camera z y Light source

Image plane

Target plane

Optical axis

Image by MIT OpenCourseWare.

51

Microscope
A perfect lens transforms a spherical wave of light from the target into a spherical wave that converges to the image.

target

image

Blurring is inversely related to the diameter of the lens.

52

Microscope
A perfect lens transforms a spherical wave of light from the target into a spherical wave that converges to the image.

target

image

Blurring is inversely related to the diameter of the lens.

53

Microscope
A perfect lens transforms a spherical wave of light from the target into a spherical wave that converges to the image.

target

image

Blurring is inversely related to the diameter of the lens.

54

Microscope
Blurring can be represented by convolving the image with the optical point-spread-function (3D impulse response).

target

image

Blurring is inversely related to the diameter of the lens.

55

Microscope
Blurring can be represented by convolving the image with the optical point-spread-function (3D impulse response).

target

image

Blurring is inversely related to the diameter of the lens.

56

Microscope
Blurring can be represented by convolving the image with the optical point-spread-function (3D impulse response).

target

image

Blurring is inversely related to the diameter of the lens.

57

Microscope
Measuring the impulse response of a microscope. Image diameter 6 times target diameter: target impulse.

Courtesy of Anthony Patire. Used with permission.

58

Microscope
Images at dierent focal planes can be assembled to form a threedimensional impulse response (point-spread function).

Courtesy of Anthony Patire. Used with permission.

59

Microscope
Blurring along the optical axis is better visualized by resampling the three-dimensional impulse response.

Courtesy of Anthony Patire. Used with permission.

60

Microscope
Blurring is much greater along the optical axis than it is across the optical axis.

Courtesy of Anthony Patire. Used with permission.

61

Microscope
The point-spread function (3D impulse response) is a useful way to characterize a microscope. It provides a direct measure of blurring, which is an important gure of merit for optics.

62

Hubble Space Telescope

Hubble Space Telescope (1990-)

http://hubblesite.org
63

Hubble Space Telescope

Why build a space telescope? Telescope images are blurred by the telescope lenses AND by atmospheric turbulence.

ha (x, y ) atmospheric blurring

hd (x, y ) blur due to mirror size

ht (x, y ) = (ha hd )(x, y ) ground-based telescope

64

Hubble Space Telescope

Telescope blur can be respresented by the convolution of blur due to atmospheric turbulence and blur due to mirror size.

ha () 2 1 0 1 2 2 1

hd () d = 12cm = 0 1 2 2 1

ht ()

ha () 2 1 0 1 2 2 1

hd () d = 1m = 0 1 2

ht ()

2 1 0 1 2 [arc-seconds]

65

Hubble Space Telescope

The main optical components of the Hubble Space Telescope are two mirrors.

http://hubblesite.org
66

Hubble Space Telescope

The diameter of the primary mirror is 2.4 meters.

http://hubblesite.org
67

Hubble Space Telescope

Hubbles rst pictures of distant stars (May 20, 1990) were more blurred than expected.

expected point-spread function http://hubblesite.org

68

early Hubble image of distant star

Hubble Space Telescope

The parabolic mirror was ground 2.2 m too at!

http://hubblesite.org
69

Hubble Space Telescope

Corrective Optics Space Telescope Axial Replacement (COSTAR): eyeglasses for Hubble!

Hubble

COSTAR

70

Hubble Space Telescope

Hubble images before and after COSTAR.

before

after

http://hubblesite.org
71

Hubble Space Telescope

Hubble images before and after COSTAR.

before

after

http://hubblesite.org
72

Hubble Space Telescope

Images from ground-based telescope and Hubble.

http://hubblesite.org
73

Impulse Response: Summary

The impulse response is a complete description of a linear, timeinvariant system. One can nd the output of such a system by convolving the input signal with the impulse response. The impulse response is an especially useful description of some types of systems, e.g., optical systems, where blurring is an important gure of merit.

74

MIT OpenCourseWare http://ocw.mit.edu

6.003 Signals and Systems

Fall 2011

For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.