I Introduction

Recommended
Problems
P1.1
Evaluate each of the following expressions for the complex number z =2e'/.
(a) Re{z}
(b) Im{z)
(c) IzI
(d) <z
(e) z*(* denotes complex conjugation)
(f) z + z*

P1.2
Let z be an arbitrary complex number.
(a) Show that

z + z*
Re{z}

(b) Show that

jIm{z} - 2

P1.3
Using Euler's formula, ei" - cos 0 + j sin 0, derive the following relations:

(a) cos 0= 2
e"- ei"
(b) sin 0 = 2j

P1.4
(a) Let z = rei". Express in polar form (i.e., determine the magnitude and angle for)
the following functions of z:
(i) z*
(ii) z2
(iii) jz
(iv) zz*

(v z
z
(vi) 1
z

P1-1
Signals and Systems
P1-2

(b) Plot in the complex plane the vectors corresponding to your answers to Problem
P1.4a(i)-(vi) for r = 3, 0 =r/6.

P1.5

Show that

(1 - el") = 2 sin (a)e i-(a ir)/2

P1.6

For x(t) indicated in Figure P1.6, sketch the following:

(a) x(-t)
(b) x(t + 2)
(c) x(2t + 2)
(d) x(1 - 3t)

x(t)

3
0F
Figure P1.6

P1.7
Evaluate the following definite integrals:

(a) fe-2tdt

(b) e dt
MIT OpenCourseWare
[Link]

Resource: Signals and Systems

Professor Alan V. Oppenheim

The following may not correspond to a particular course on MIT OpenCourseWare, but has been
provided by the author as an individual learning resource.

For information about citing these materials or our Terms of Use, visit: [Link]
2 Signals and Systems: Part I
Recommended
Problems
P2.1

Let x(t) = cos(wx(t + rx) + Ox).

(a) Determine the frequency in hertz and the period of x(t) for each of the follow­
ing three cases:

(i) r/3 0 21r

(ii) 3r/4 1/2 7r/4
(iii) 3/4 1/2 1/4

(b) With x(t) = cos(wx(t + rx) + Ox) and y(t) = sin(w,(t + -r,)+ 0,), determine for
which of the following combinations x(t) and y(t) are identically equal for all t.
WX 0
T x WY TY Oy

(i) r/3 0 2r ir/3 1 - r/3

(ii) 3-r/4 1/2 7r/4 11r/4 1 37r/8
(iii) 3/4 1/2 1/4 3/4 1 3/8

P2.2
Let x[n] = cos(Qx(n + Px) + Ox).
(a) Determine the period of x[n] for each of the following three cases:
Ox PX Ox

(i) r/3 0 27r

(ii) 3-r/4 2 r/4
(iii) 3/4 1 1/4

(b) With x[n] = cos(Q,(n + PX) + Ox) and y[n] = cos(g,(n + Py) + 6,), determine
for which of the following combinations x[n] and y[n] are identically equal for
all n.
QX PX Ox
X, Q, PO

(i) r/3 0 27 8r/3 0 0

(ii) 37r/4 2 w/ 4 3r/4 1 -ir
(iii) 3/4 1 1/4 3/4 0 1

P2.3

(a) A discrete-time signal x[n] is shown in Figure P2.3.

P2-1
Signals and Systems
P2-2

x [n]

-3 -2 -1 0 1 2 3 4 5 6

Figure P2.3

Sketch and carefully label each of the following signals:

(i) x[n - 2]
(ii) x[4 - n]
(iii) x[2n]

(b) What difficulty arises when we try to define a signal as.x[n/2]?

P2.4
For each of the following signals, determine whether it is even, odd, or neither.

(a) (b)

1­ P2.4-1
Figure

(c) (d)

x(t) x [n ]

Figure P2.4-3 x~nn

-2 F3 P2.44n
t
0 1 2 3 4

Figure P2.4-4
 Signals and Systems: Part I / Problems
P2-3

(e) (f)

x [n] 2 x[n]

-3 -2 -1 02
1 2 3 4 n

-3 -2 -1 0 1 2 3
Figure P2.4-5 Figure P2.4-6

P2.5
Consider the signal y[n] in Figure P2.5.

y[n] 2'.

e n
-3 -2-1 0 1 2 3

Figure P2.5

(a) Find the signal x[n] such that Ev{x[n]} = y[n] for n> 0, and Od(x[n]} = y[n]
for n < 0.
(b) Suppose that Ev{w[n]} = y[n] for all n. Also assume that w[n] = 0 for n < 0.
Find w[n].

P2.6
(a) Sketch x[n] = a"for a typical a in the range -1 < a < 0.
(b) Assume that a = -e-' and define y(t) as y(t) = eO'. Find a complex number #
such that y(t), when evaluated at t equal to an integer n, is described by
(-e- )".
(c) For y(t) found in part (b), find an expression for Re{y(t)} and Im{y(t)}. Plot
Re{y(t)} and Im{y(t)} for t equal to an integer.

P2.7
4 e(-i+2
,). Sketch and label the following:
Let x(t) = \/2(1 + j)ej"1
(a) Re{x(t)}
(b) Im{x(t)}
(c) x(t + 2) + x*(t + 2)
Signals and Systems
P2-4

P2.8
Evaluate the following sums:
5

(a) T 2(3n
n=0

(b) b
n=2

2n
(c)Z -~3
n=o

Hint: Convert each sum to the form

N-1

C ( a' = SN or Cn
n=o n =0

and use the formulas

SN = C aN 11-a
C for lal < 1
1a

P2.9
(a) Let x(t) and y(t) be periodic signals with fundamental periods Ti and T2, respec­
tively. Under what conditions is the sum x(t) + y(t) periodic, and what is the
fundamental period of this signal if it is periodic?
(b) Let x[n] and y[n] be periodic signals with fundamental periods Ni and N 2 ,
respectively. Under what conditions is the sum x[n] + y[n] periodic, and what
is the fundamental period of this signal if it is periodic?
(c) Consider the signals

x(t) = cos-t + 2 sin 3 '

33
y(t) = sin irt
Show that z(t) = x(t)y(t) is periodic, and write z(t) as a linear combination of
harmonically related complex exponentials. That is, find a number T and com­
plex numbers Ck such that
z(t) = jce(21'/T'
k

P2.10

In this problem we explore several of the properties of even and odd signals.
(a) Show that if x[n] is an odd signal, then
+o0
( x[n] = 0
n=-00

(b) Show that if xi[n] is an odd signal and x2[n] is an even signal, then x,[n]x 2[n] is
an odd signal.
 Signals and Systems: Part I / Problems
P2-5

(c) Let x[n] be an arbitrary signal with even and odd parts denoted by
xe[n] = Ev{x[n]}, x[n] = Od{x[n]}
Show that

~x[n] = [n] + ~3X'[n]

(d) Although parts (a)-(c) have been stated in terms of discrete-time signals, the
analogous properties are also valid in continuous time. To demonstrate this,
show that

Jx 2 (t )dt = 2x(t )dt + J 2~(t) dt,

where xe(t) and x,(t) are, respectively, the even and odd parts of x(t).

P2.11
Let x(t) be the continuous-time complex exponential signal x(t) = ei0O' with fun­
damental frequency wo and fundamental period To = 2 7r/wo. Consider the discrete-
time signal obtained by taking equally spaced samples of x(t). That is, x[n] =
x(nT) = eswonr

(a) Show that x[n] is periodic if and only if T/TO is a rational number, that is, if and
only if some multiple of the sampling interval exactly equals a multiple of the
period x(t).
(b) Suppose that x[n] is periodic, that is, that
T p
- , (P2.11-1)
To q
where p and q are integers. What are the fundamental period and fundamental
frequency of x[n]? Express the fundamental frequency as a fraction of woT.
(c) Again assuming that T/TO satisfies eq. (P2.11-1), determine precisely how many
periods of x(t) are needed to obtain the samples that form a single period of
x[n].
MIT OpenCourseWare
[Link]

Resource: Signals and Systems

Professor Alan V. Oppenheim

The following may not correspond to a particular course on MIT OpenCourseWare, but has been
provided by the author as an individual learning resource.

For information about citing these materials or our Terms of Use, visit: [Link]
3 Signals and Systems: Part II
Recommended
Problems
P3.1

Sketch each of the following signals.

(a) x[n] = b[n] + 3[n - 3]
(b) x[n] = u[n] - u[n - 5]
(c) x[n] = 6[n] + 1n + (i)2 [n - 2] + (i)ag[n - 3]
(d) x(t) = u(t + 3) - u(t - 3)
(e) x(t) = 6(t + 2)
(f) x(t) = e-'u(t)

P3.2
Below are two columns of signals expressed analytically. For each signal in column
A, find the signal or signals in column B that are identical.
A B
(1) 6[n + 1] (a) 6
t[k]
k=-0
(2) (I)"U [n]

(3) 6(t) (b) du(t)

(4) u(t)
(5) u[n] (C) 6[k]
(6) b[n + 1]u[n] k=O

(d) L (I)k[n - k]
k=O

(e) J t6(r) dr

(f) u[n]

(g) E (I)kb[n - k]
k= -oo

(h) b[n + 1]
(i) $

P3.3

(a) Express the following as sums of weighted delayed impulses, i.e., in the form

x[n] = ( ako[n - k]
k= -o

P3-1
Signals and Systems
P3-2

x [n]

2 4
2 4?
-3 -2 -1 0 1 3 5 6 7 8

-2

Figure P3.3-1

(b) Express the following sequence as a sum of step functions, i.e., in the form

s[n] = ( aku[n - k]
k= -00

s[n] 3

-3 -2 2 3
10 1 0 n
-6-5-4 -1 0 1- 4 5 6
-1 ­

Figure P3.3-2

P3.4
For x(t) indicated in Figure P3.4, sketch the following:
(a) x(1 - t)[u(t + 1) - u(t - 2)]
(b) x(1 - t)[u(t + 1) - u(2 - 3t)]

x(t)

0 12

Figure P3.4
 Signals and Systems: Part II / Problems
P3-3

P3.5
Consider the three systems H, G, and F defined in Figure P3.5-1.

x [n] H y[n]=x2 [nJ

x[n] G . y[n]=x[n]-x[n-1]

x [n F y[n]=2x[n]+x[n-1]

Figure P3.5-1

The systems in Figures P3.5-2 to P3.5-7 are formed by parallel and cascade combi­
nation of H, G, and F. By expressing the output y[n] in terms of the input x[n],
determine which of the systems are equivalent.
(a) (b)

H G

x + y x + y

Figure P3.5-2 Figure P3.5-3

(c) (d)

x-- G H Y X- H -- + G -- WY

Figure P3.5-4 Figure P3.5-5

(e) (f)

x G F Y x -- F G Y

Figure P3.5-6 Figure P3.5-7

P3.6
Table P3.6 contains the input-output relations for several continuous-time and dis­
crete-time systems, where x(t) or x[n] is the input. Indicate whether the property
along the top row applies to each system by answering yes or no in the appropriate
boxes. Do not mark the shaded boxes.
Signals and Systems
P3-4

Properties
y(t), y[n] Memoryless Linear Time-Invariant Causal Invertible Stable

Table P3.6

P3.7

Consider the following systems

H: y(t) = x(r) dr (an integrator),

G: y(t) = x(2t),
where the input is x(t) and the output is y(t).
(a) What is H-', the inverse of H? What is G- ?
(b) Consider the system in Figure P3.7. Find the inverse F-1 and draw it in block
diagram form in terms of H-1 and G-1.

x(t) --- H G w(t)

x(t) F --- o w(t)

Figure P3.7
 Signals and Systems: Part II / Problems
P3-5

Optional
Problems
P3.8
In this problem we illustrate one of the most important consequences of the prop­
erties of linearity and time invariance. Specifically, once we know the response of a
linear system or of a linear, time-invariant (LTI) system to a single input or the
responses to several inputs, we can directly compute the responses to many other
input signals.
(a) Consider an LTI system whose response to the signal x 1(t) in Figure P3.8-1(a)
is the signal y 1(t) illustrated in Figure P8-1(b). Determine and sketch carefully
the response of the system to the input x 2(t) depicted in Figure P3.8-1(c).

x1 (t) Y1 (t)

1 1 t
0 1 2 0 1 2

(a) (b)

x 2 (t) x3 (t)

1 1 i 7
1 t4-101 2
-1

(c) (d)

x(t

0 1 2

(e)

Figure P3.8-1

(b) Determine and sketch the response of the system considered in part (a) to the
input x 3 (t) shown in Figure P3.8-1(d).
Signals and Systems
P3-6

(c) Suppose that a second LTI system has the following output y(t) when the input
is the unit step x(t) = u(t):
y(t) = e-'u(t) + u(-1 - t)
Determine and sketch the response of this system to the input x(t) shown in
Figure P3.8-1(e).
(d) Suppose that a particular discrete-time linear (but possibly not time-invariant)
system has the responses yi[n], y2[n], and yAn] to the input signals xi[n], x2[n],
and x[n], respectively, as illustrated in Figure P3.8-2(a). If the input to this
system is x[n] as illustrated in Figure P3.8-2(b), what is the output y[n]?

x, [n] y, [n]

0
1 n - .

0 1 2 3
a 0 n
X2 [n] Y 2 [n] 2

01 0 1 3
x 3 [n] y 3 [n]

0p 0pop
0 o n -..-
. 1 2 4
4 ­ .
n
-1 -1 1 3

(a)

x[n] 2

n
-1 0

-2

(b)

Figure P3.8-2

(e) If an LTI system has the response y1[n] to the input xi[n] as in Figure P3.8-2(a),
what would its responses be to x 2[n] and x 3[n]?
(f) A particular linear system has the property that the response to t' is cos kt.
What is the response of this system to the inputs
x 1 (t) = ,r + 6t 2 - 47t 5 + \/et6
1+ t0
x 2(t) = 1 +t 2
 Signals and Systems: Part II / Problems
P3-7

P3.9
(a) Consider a system with input x(t) and with output y(t) given by
+0o
y(t) = ( x(t)b(t - nT)

(i) Is this system linear?

(ii) Is this system time-invariant?
For each part, if your answer is yes, show why this is so. If your answer is no,
produce a counterexample.
(b) Suppose that the input to this system is x(t) = cos 27rt. Sketch and label care­
fully the output y(t) for each of the following values of T: T = 1, 1, 1, 1, -. All
of your sketches should have the same horizontal and vertical scales.
(c) Repeat part (b) for x(t) = e' cos 27rt.

P3.10
(a) Is the following statement true or false?
The series interconnection of two linear, time-invariant systems is itself a lin­
ear, time-invariant system.
Justify your answer.
(b) Is the following statement true or false?
The series connection of two nonlinear systems is itself nonlinear.
Justify your answer.
(c) Consider three systems with the following input-output relations:

x[n/2], n even
System 1: y[n] = 1o, n odd

System 2: y[n] = x[n] + 1x[n - 1] + ix[n - 2]

System 3: y[n] = x[2n]
Suppose that these systems are connected in series as depicted in Figure P3.10.
Find the input-output relation for the overall interconnected system. Is this sys­
tem linear? Is it time-invariant?

x[n] System 1 - System 2 System 3 y[n]

Figure P3.10

(d) Consider a second series interconnection of the form of Figure P3.10 where the
three systems are specified by the following equations, with a, b, and c real
numbers:
System 1: y[n] = x[-n]
System 2: y[n] = ax[n - 1] + bx[n] + cx[n + 1]
System 3: y[n] = x[-n]
Signals and Systems
P3-8

Find the input-output relation for the overall interconnected system. Under
what conditions on the numbers a, b, and c does the overall system have each
of the following properties?
(i) The overall system is linear and time-invariant.
(ii) The input-output relation of the overall system is identical to that of sys­
tem 2.
(iii) The overall system is causal.

P3.11
Determine whether each of the following systems is linear and/or time-invariant.
In each case, x[n] denotes the input and y[n] denotes the output. Assume that
x[] > 0.
(a) y[n] = x[n] + x[n - 1]
(b) y[n] = x[n] + x[n - 1] + x[O]

P3.12
(a) Show that causality for a continuous-time linear system implies the following
statement:
For any time to and any input x(t) such that x(t) = 0 for t < to, the correspond­
ing output y(t) must also be zero for t < to.
The analogous statement can be made for discrete-time linear systems.
(b) Find a nonlinear system that satisfies this condition but is not causal.
(c) Find a nonlinear system that is causal but does not satisfy this condition.
(d) Show that invertibility for a discrete-time linear system is equivalent to the fol­
lowing statement:
The only input that produces the output y[n] = 0 for all n is x[n] = 0 for all n.
The analogous statement is also true for continuous-time linear systems.
(e) Find a nonlinear system that satisfies the condition of part (d) but is not
invertible.
MIT OpenCourseWare
[Link]

Resource: Signals and Systems

Professor Alan V. Oppenheim

The following may not correspond to a particular course on MIT OpenCourseWare, but has been
provided by the author as an individual learning resource.

For information about citing these materials or our Terms of Use, visit: [Link]
4 Convolution
Recommended
Problems
P4.1

This problem is a simple example of the use of superposition. Suppose that a dis­
crete-time linear system has outputs y[n] for the given inputs x[n] as shown in Fig­
ure P4.1-1.

Input x[n] Outputy [n]

III xn] yi[n]

0 0 n 0 n
-1 0 1 2 -2 -1 0 1 2

x2[n] 1 2[n]

0 0 n 0 0 n
-1 0 1 2 3 -1 0 1 2 3

21

1 X3 [n] y3[n]

.-. n -n

-1 0 1 2 3 -1 0 1 2
Figure P4.1-1

Determine the response y4 [n] when the input is as shown in Figure P4.1-2.

I x4[n]

-1
012 0 2 3

-1

Figure P4.1-2

(a) Express x 4 [n] as a linear combination of x 1[n], x 2[n], and x 3 [n].

(b) Using the fact that the system is linear, determine y 4[n], the response to x 4[n].
(c) From the input-output pairs in Figure P4.1-1, determine whether the system is
time-invariant.

P4-1
Signals and Systems
P4-2

P4.2
Determine the discrete-time convolution of x[n] and h[n] for the following two
cases.

(a)

h[n] 2
x[n]

0 -1.11 0 1 11
1 2 3 4
11J
Fiur P42-
-1 0 1 2 3
00
4

Figure P4.2-1

(b)

3
h[nI
2

x [n ]
X17

0 1 2 3 4 -1 0 1 2 3 4 5
Figure P4.2-2

P4.3
Determine the continuous-time convolution of x(t) and h(t) for the following three
cases:
(a)
x(t) h (t)

t t
0 4 0 4
Figure P4.3-1
 Convolution / Problems
P4-3

(b)
x(t) h (t)

1) u (t - 1) u(t+1
1-- -(t-

t t
0 1 -1 0
Figure P4.3-2

(c)
x(t) h (t)
F6(t-2)

-l 0 1 3 0 2

Figure P4.3-3

P4.4
Consider a discrete-time, linear, shift-invariant system that has unit sample re­
sponse h[n] and input x[n].
(a) Sketch the response of this system if x[n] = b[n - no], for some no > 0, and
h[n] = (i)"u[n].
(b) Evaluate and sketch the output of the system if h[n] = (I)"u[n] and x[n] =
u[n].
(c) Consider reversing the role of the input and system response in part (b). That
is,
h[n] = u[n],
x[n] = (I)"u[n]
Evaluate the system output y[n] and sketch.

P4.5
(a) Using convolution, determine and sketch the responses of a linear, time-invar­
iant system with impulse response h(t) = e-t 2 u(t) to each of the two inputs
x 1(t), x 2 (t) shown in Figures P4.5-1 and P4.5-2. Use yi(t) to denote the response
to x 1 (t) and use y 2 (t) to denote the response to x 2 (t).
Signals and Systems
P4-4

(i)
X1 (t) = u(t)

t
0
Figure P4.5-1

(ii)
x 2 (t)

t
0 3

Figure P4.5-2

(b) x 2 (t) can be expressed in terms of x,(t) as

x 2 (t) = 2[x(t) - xi(t - 3)]

By taking advantage of the linearity and time-invariance properties, determine

how y 2(t) can be expressed in terms of yi(t). Verify your expression by evalu­
ating it with yl(t) obtained in part (a) and comparing it with y 2(t) obtained in
part (a).

Optional
Problems
P4.6
Graphically determine the continuous-time convolution of h(t) and x(t) for the
cases shown in Figures P4.6-1 and P4.6-2.
 Convolution / Problems
P4-5

(a)
h(t)

x(t) 2

t1
t
0 1 0

Figure P4.6-1

(b)

h (t)

x(t)

t t
0 1 2 0 1 2

Figure P4.6-2

P4.7
Compute the convolution y[n] = x[n] * h[n] when

x[n] =au[n], O < a< 1,

h[n] =#"u[n], 0 < #< 1

Assume that a and # are not equal.

P4.8
Suppose that h(t) is as shown in Figure P4.8 and x(t) is an impulse train, i.e.,

x(t) = ( oft-kT)
k= -o0
Signals and Systems
P4-6

(a) Sketch x(t).

(b) Assuming T = 2, determine and sketch y(t) = x(t) * h(t).

P4.9
Determine if each of the following statements is true in general. Provide proofs for
those that you think are true and counterexamples for those that you think are
false.
(a) x[n] *{h[ng[n]} = {x[n] *h[n]}g[n]
(b) If y(t) = x(t) * h(t), then y(2t) = 2x(2t) * h(2t).
(c) If x(t) and h(t) are odd signals, then y(t) = x(t) * h(t) is an even signal.
(d) If y(t) = x(t) * h(t), then Ev{y(t)} = x(t) * Ev{h(t)} + Ev{x(t)} * h(t).

P4.10
Let 1 1(t) and 2 2 (t) be two periodic signals with a common period To. It is not too
difficult to check that the convolution of 1 1(t) and t 2 (t) does not converge. However,
it is sometimes useful to consider a form of convolution for such signals that is
referred to as periodicconvolution. Specifically, we define the periodic convolution
of t 1 (t) and X 2(t) as
TO

g(t) = T 1 (r)- 2 (t - r) dr = t 1(t)* t 2 (t) (P4.10-1)

Note that we are integrating over exactly one period.

(a) Show that q(t) is periodic with period To.
(b) Consider the signal
a + T0

Pa(t) = 1(rF)t2(t - r) dr,

fa

where a is an arbitrary real number. Show that

9(t) = Ya(t)

Hint: Write a = kTo - b, where 0 b < To.

(c) Compute the periodic convolution of the signals depicted in Figure P4.10-1,
where To = 1.
 Convolution / Problems
P4-7

et

-1 0 1 2 3

R2 (t)

t
-1 - 22 1 1 3
22 2 5 3

Figure P4.10-1

(d) Consider the signals x1[n] and x 2[n] depicted in Figure P4.10-2. These signals
are periodic with period 6. Compute and sketch their periodic convolution using
No = 6.

x, [n]

...
IT I '1 I II
I-6
0II 1 1 6
T II
12
..

1? 11
X2 [n]
2

-6 0 6 12

Figure P4.10-2
(e) Since these signals are periodic with period 6, they are also periodic with period
12. Compute the periodic convolution of xi[n] and x2[n] using No = 12.

P4.11

One important use of the concept of inverse systems is to remove distortions of some
type. A good example is the problem of removing echoes from acoustic signals. For
example, if an auditorium has a perceptible echo, then an initial acoustic impulse is
Signals and Systems
P4-8

followed by attenuated versions of the sound at regularly spaced intervals. Conse­

quently, a common model for this phenomenon is a linear, time-invariant system
with an impulse response consisting of a train of impulses:

h(t) = [ hkb(t-kT) (P4.11-1)

k=O

Here the echoes occur T s apart, and hk represents the gain factor on the kth echo
resulting from an initial acoustic impulse.
(a) Suppose that x(t) represents the original acoustic signal (the music produced
by an orchestra, for example) and that y(t) = x(t) * h(t) is the actual signal
that is heard if no processing is done to remove the echoes. To remove the dis­
tortion introduced by the echoes, assume that a microphone is used to sense
y(t) and that the resulting signal is transduced into an electrical signal. We will
also use y(t) to denote this signal, as it represents the electrical equivalent of
the acoustic signal, and we can go from one to the other via acoustic-electrical
conversion systems.
The important point to note is that the system with impulse response given
in eq. (P4.11-1) is invertible. Therefore, we can find an LTI system with impulse
response g(t) such that
y(t) *g(t) = x(t)
and thus, by processing the electrical signal y(t) in this fashion and then con­
verting back to an acoustic signal, we can remove the troublesome echoes.
The required impulse response g(t) is also an impulse train:

g(t) = ( gkAot-kT)
k=O

Determine the algebraic equations that the successive gk must satisfy and solve
for gi, g 2 , and g3 in terms of the hk. [Hint: You may find part (a) of Problem 3.16
of the text (page 136) useful.]
(b) Suppose that ho = 1, hi = i, and hi = 0 for all i > 2. What is g(t) in this case?
(c) A good model for the generation of echoes is illustrated in Figure P4.11. Each
successive echo represents a fedback version of y(t), delayed by T s and scaled
by a. Typically 0 < a < 1 because successive echoes are attenuated.

x(t) ± y(t)

Delay
T

Figure P4.11

(i) What is the impulse response of this system? (Assume initial rest, i.e.,
y(t) = 0 for t < 0 if x(t) = 0 for t < 0.)
(ii) Show that the system is stable if 0 < a < 1 and unstable if a > 1.
(iii) What is g(t) in this case? Construct a realization of this inverse system
using adders, coefficient multipliers, and T-s delay elements.
 Convolution / Problems
P4-9

Although we have phrased this discussion in terms of continuous-time systems

because of the application we are considering, the same general ideas hold in
discrete time. That is, the LTI system with impulse response

h[n] = ( hkS[n-kN]
k=O

is invertible and has as its inverse an LTI system with impulse response

g[n] = (g [nkN]
­
k=O

It is not difficult to check that the gi satisfy the same algebraic equations as in
part (a).
(d) Consider the discrete-time LTI system with impulse response

h[n] = ( S[n-kN]
k=-m

This system is not invertible. Find two inputs that produce the same output.

P4.12
Our development of the convolution sum representation for discrete-time LTI sys­
tems was based on using the unit sample function as a building block for the rep­
resentation of arbitrary input signals. This representation, together with knowledge
of the response to 5[n] and the property of superposition, allowed us to represent
the system response to an arbitrary input in terms of a convolution. In this problem
we consider the use of other signals as building blocks for the construction of arbi­
trary input signals.
Consider the following set of signals:
$[n] = (i)"u[n],
#[n ] = [n - k], k = 0, 1, ±2 3, ....
(a) Show that an arbitrary signal can be represented in the form
+ 00
x[n] = ( ak4[n - k]
k=

by determining an explicit expression for the coefficient ak in terms of the values

of the signal x[n]. [Hint: What is the representation for 6[n]?]
(b) Let r[n] be the response of an LTI system to the input x[n] = #[n]. Find an
expression for the response y[n] to an arbitrary input x[n] in terms of r[n] and
x[n].
(c) Show that y[n] can be written as
y[n] = 0[n] * x[n] * r[n]

by finding the signal 0[n].

(d) Use the result of part (c) to express the impulse response of the system in terms
of r[n]. Also, show that
0[n] *#[n] = b[n]
MIT OpenCourseWare
[Link]

Resource: Signals and Systems

Professor Alan V. Oppenheim

The following may not correspond to a particular course on MIT OpenCourseWare, but has been
provided by the author as an individual learning resource.

For information about citing these materials or our Terms of Use, visit: [Link]
5 Properties of Linear,
Time-Invariant Systems
Recommended
Problems
P5.1

Consider an integrator that has the input-output relation

y(t) = x(r) dr

Determine the input-output relation for the inverse system.

P5.2
The first-order difference equation y[n] - ay[n - 1] = x[n], 0 < a < 1, describes a
particular discrete-time system initially at rest.
(a) Verify that the impulse response h[n] for this system is h[n] = a'u[n].
(b) Is the system
(i) memoryless?
(ii) causal?
(iii) stable?
Clearly state your reasoning.
(c) Is this system stable if |a I > 1?

P5.3
The first-order differential equation

dy(t) + 2y(t) =
x(t)
dt

describes a particular continuous-time system initially at rest.

(a) Verify that the impulse response of this system is h(t) = e -
2
u(t).
(b) Is this system
(i) memoryless?
(ii) causal?
(iii) stable?
Clearly state your reasoning.

P5-1
Signals and Systems
P5-2

P5.4
Consider the linear, time-invariant system in Figure P5.4, which is composed of a
cascade of two LTI systems. u(t) is a unit step signal and s(t) is the step response
of system L.

s(t)
u(t) - L d/dt y(t)

Figure P5.4

Using the fact that the overall response of LTI systems in cascade is indepen­
dent of the order in which they are cascaded, show that the impulse response of
system L is the derivative of its step response, i.e.,

h(t) = ds(t)
dit

P5.5
Consider the cascade of two systems shown in Figure P5.5. System B is the inverse
of system A.

x(t) LTI y(t) System w(t)

Input System
A B Output

Figure P5.5

(a) Suppose the input is b(t). What is the output w(t)?

(b) Suppose the input is some more general signal x(t). What is the output w(t) in
terms of x(t)?

Optional
Problems
P5.6

(a) Consider again the cascade of two systems presented in Problem P5.5. Suppose
an input x 1 (t) produces yi(t) as system A output and an input x 2(t) produces
y 2 (t) as system A output. What is w(t) if the input is such that y(t), the output
of system A, is ay 1 (t) + by 2 (t) with a, b constants?
(b) Suppose an input x 1(t) produces yi(t) as system A output. What is w(t) if x(t)
is such that y(t) = y 1(t - r)?
(c) Is system B an LTI system? Justify your answer.
 Properties of Linear, Time-Invariant Systems / Problems
P5-3

P5.7
Consider the three discrete-time signals shown in Figure P5.7.

x[n] 2 y[n]

e--- n n
0 1 -1 0

-1

w[n]

01
e e n

Figure P5.7

(a) Verify the distributive law of convolution:

(x + w) * y = (x * y) + (w * y)
(b) You may have noticed a similarity between the convolution operation and mul­
tiplication, but they are not equivalent. Verify that
(x * y) - w # x * (y - w)

P5.8
Let y(t) = x(t) * h(t). Show the following.
dy(t) dh(t) dx(t)
(a) t= x(t) * =t d *h(t)
dt dt dt
(b) y(t) = (J'_ x(r) dr) * h'(t)
(c) y(t) = f'_[x'(r) * h(r)] dr

(d) y(t) = x'(t) * J'_,h(r) dr

P5.9

Determine if each of the following statements concerning LTI systems is true or

false. Justify your answers.
(a) If h(t) is the impulse response of an LTI system and h(t) is periodic and non­
zero, the system is unstable.
Signals and Systems
P5-4

(b) The inverse of a causal LTI system is always causal.

(c) If Ih[n] 5 K for each n, where K is a given number, then the LTI system with
h[n] as its impulse response is stable.
(d) If a discrete-time LTI system has an impulse response h[n] of finite duration,
the system is stable.
(e) If an LTI system is causal, it is stable.
(f) The cascade of a noncausal LTI system with a causal one is necessarily
noncausal.
(g) A continuous-time LTI system is stable if and only if its step response s(t) is
absolutely integrable, i.e.,

|s(t)I dt < oo
(h) A discrete-time LTI system is causal if and only if its step response s[n] is zero
for n < 0.

P5.10
In Section 3.7 of the text we characterized the unit doublet through the equation

x(t) * u 1 (t) = x(t - r)ui(r) dr = x'(t) (P5.10-1)

for any signal x(t). From this equation we derived the fact that
f+
g(r)ui(r) d-r = -g'(0) (P5.10-2)

(a) Show that eq. (P5.10-2) is an equivalent characterization of ui(t) by showing

that eq. (P5.10-2) implies eq. (P5.10-1). [Hint: Fix t and define the signal g(r)
x(t - r).]
Thus we have seen that characterizing the unit impulse or unit doublet by how
it behaves under convolution is equivalent to characterizing how it behaves under
integration when multiplied by an arbitrary signal g(t). In fact, as indicated in Sec­
tion 3.7 of the text, the equivalence of these operational definitions holds for all
signals and in particular for all singularity functions.
(b) Let f(t) be a given signal. Show that
f(t)ui(t) = f(0)ui(t) - f'(0)3(t)
by showing that both have the same operational definitions.
(c) Determine the value of

J X(r)u 2 (r) dr

(d) Find an expression for f(t)u2 (t) analogous to that considered in part (b) for
f(t)ui(t).
 Properties of Linear, Time-Invariant Systems / Problems
P5-5

P5.11
Consider the cascade of two systems H and G as shown in Figure P5.11.

-1 H G

Figure P5.11

(a) If H and G are both LTI causal systems, prove that the overall system is causal.
(b) If H and G are both stable systems, show that the overall system is stable.

P5.12
Find the combined impulse response of the LTI system in Figure P5.12. Recall that
x(t)* h(t) * h-1(t) = x(t).

x(t) h2]--N h2 + h1 h-2 y(t)

..
.r.. h

Figure P5.12

P5.13
Find the necessary and sufficient condition on the impulse response h[n] such that
for any input x[n],
max {Ix[n]I} > max {Iy[n]|},
where y[n] = x[n] * h[n].
MIT OpenCourseWare
[Link]

Resource: Signals and Systems

Professor Alan V. Oppenheim

The following may not correspond to a particular course on MIT OpenCourseWare, but has been
provided by the author as an individual learning resource.

For information about citing these materials or our Terms of Use, visit: [Link]
6 Systems Represented by
Differential and Difference
Equations
Recommended
Problems
P6.1
Suppose that y 1(t) and y 2(t) both satisfy the homogeneous linear constant-coeffi­
cient differential equation (LCCDE)
dy(t) + ay(t)
= 0
dt

Show that y 3 (t) = ayi(t) + 3y2 (t), where a and # are any two constants, is also a
solution to the homogeneous LCCDE.

P6.2
In this problem, we consider the homogeneous LCCDE
d 2yt + 3 dy(t) + 2y(t) = 0 (P6.2-1)
dt 2 dt

(a) Assume that a solution to eq. (P6.2-1) is of the form y(t) = es'. Find the qua­
dratic equation that s must satisfy, and solve for the possible values of s.
(b) Find an expression for the family of signals y(t) that will satisfy eq. (P6.2-1).

P6.3
Consider the LCCDE

dy(t) + 1 y(t) = x(t), x(t) = e- t u(t) (P6.3-1)

dt 2

(a) Determine the family of signals y(t) that satisfies the associated homogeneous
equation.
(b) Assume that for t > 0, one solution of eq. (P6.3-1), with x(t) as specified, is of
the form
y 1(t) = Ae-, t > 0

Determine the value of A.

(c) By substituting into eq. (P6.3-1), show that
y 1(t) = [2e -t/2 - 2e-']u(t)

is one solution for all t.

P6-i
Signals and Systems
P6-2

P6.4

Consider the block diagram relating the two signals x[n] and y[n] given in Figure
P6.4.

x[n] + 1 y[n]

1
2

Figure P6.4

Assume that the system described in Figure P6.4 is causal and is initially at rest.
(a) Determine the difference equation relating y[n] and x[n].
(b) Without doing any calculations, determine the value of y[ -5] when x[n] = u[n].
(c) Assume that a solution to the difference equation in part (a) is given by
y[n] = Kanu[n]
when x[n] = b[n]. Find the appropriate value of K and a, and verify that y[n]
satisfies the difference equation.
(d) Verify your answer to part (c) by directly calculating y[O], y[l], and y[2].

P6.5
Figure P6.5 presents the direct form II realization of a difference equation. Assume
that the resulting system is linear and time-invariant.

x[n] O + r0n] y[n]

+1 -2
3

Figure P6.5

(a) Find the direct form I realization of the difference equation.

(b) Find the difference equation described by the direct form I realization.
(c) Consider the intermediate signal r[n] in Figure P6.5.
(i) Find the relation between r[n] and y[n].
(ii) Find the relation between r[n] and x[n].
(iii) Using your answers to parts (i) and (ii), verify that the relation between
y[n] and x[n] in the direct form II realization is the same as your answer
to part (b).
 Systems Represented by Differential and Difference Equations / Problems
P6-3

P6.6
Consider the following differential equation governing an LTI system.
dytt) dx(t)
dt + ay(t) = b di + cx(t) (P6.6-1)
dt dt
(a) Draw the direct form I realization of eq. (P6.6-1).
(b) Draw the direct form II realization of eq. (P6.6-1).

Optional
Problems
P6.7
Consider the block diagram in Figure P6.7. The system is causal and is initially at
rest.

r [n]
x [n] + y [n]

-4
Figure P6.7

(a) Find the difference equation relating x[n] and y[n].

(b) For x[n] = [n], find r[n] for all n.
(c) Find the system impulse response.

P6.8
Consider the system shown in Figure P6.8. Find the differential equation relating
x(t) and y(t).

x(t) + a r(t) + y t

a b
Figure P6.8
Signals and Systems
P6-4

P6.9
Consider the following difference equation:
y[n] - ly[n - 1] = x[n] (P6.9-1)
with
x[n] = K(cos gon)u[n] (P6.9-2)
Assume that the solution y[n] consists of the sum of a particular solution y,[n] to
eq. (P6.9-1) for n 0 and a homogeneous solution yjn] satisfying the equation
Yh[flI - 12Yhn - 1] =0.

(a) If we assume that Yh[n] = Az", what value must be chosen for zo?
(b) If we assume that for n 0,
y,[n] = B cos(Qon + 0),
what are the values of B and 0? [Hint: It is convenient to view x[n] =
Re{Kej"onu[n]} and y[n] = Re{Ye"onu[n]}, where Y is a complex number to be
determined.]

P6.10

Show that if r(t) satisfies the homogeneous differential equation

m d=r(t)
dt 0
and if s(t) is the response of an arbitrary LTI system H to the input r(t), then s(t)
satisfies the same homogeneous differential equation.

P6.11
(a) Consider the homogeneous differential equation
N dky)
k~=0 (P6.11-1)
k=ak dtk
Show that if so is a solution of the equation
N

p(s) = E akss
k=O
= 0, (P6.11-2)

then Aeso' is a solution of eq. (P6.11-1), where A is an arbitrary complex

constant.
(b) The polynomial p(s) in eq. (P6.11-2) can be factored in terms of its roots
S1, ... ,S,.:
p(s) = aN(S - SI)1P(S - S2)2 . . . (S - Sr)ar,

where the si are the distinct solutions of eq. (P6.11-2) and the a are their mul­
[Link] that
U1 + o2 + + Ur = N
In general, if a, > 1, then not only is Ae''' a solution of eq. (P6.11-1) but so
is Atiesi' as long as j is an integer greater than or equal to zero and less than or
 Systems Represented by Differential and Difference Equations / Problems
P6-5

equal to oa - 1. To illustrate this, show that if ao = 2, then Atesi is a solution

of eq. (P6.11-1). [Hint: Show that if s is an arbitrary complex number, then
N ak dtk = Ap(s)te' t + A
estI

Thus, the most general solution of eq. (P6.11-1) is

p ci-1

( Aesi ,
i=1 j=0

where the Ai, are arbitrary complex constants.

(c) Solve the following homogeneous differential equation with the specified aux­
iliary conditions.
d 2 y(t)
2
dt2
+ 2 dy(t)
dt
+ y(t) = 0, y(0) = 1, y'() = 1
MIT OpenCourseWare
[Link]

Resource: Signals and Systems

Professor Alan V. Oppenheim

The following may not correspond to a particular course on MIT OpenCourseWare, but has been
provided by the author as an individual learning resource.

For information about citing these materials or our Terms of Use, visit: [Link]
7 Continuous-Time Fourier Series
Recommended
Problems
P7.1

(a) Suppose that the signal ei" is applied as the excitation to a linear, time-invar­
iant system that has an impulse response h(t). By using the convolution inte­
gral, show that the resulting output is H(w)ew, where H(w) = f". h(r)e -jw dr.
(b) Assume that the system is characterized by a first-order differential equation
dy(t) + ay(t) = x(t).
dt

If x(t) = ejw" for all t, then y(t) = H(w)eo" for all t. By substituting into the
differential equation, determine H(w).

P7.2

(a) Suppose that z", where z is a complex number, is the input to an LTI system that
has an impulse response h[n]. Show that the resulting output is given by H(z)z",
where

H(z) = E h[n]z-".

(b) If the system is characterized by a first-order difference equation,

y[n] + ay[n - 1] = x[n],
determine H(z).

P7.3
Find the Fourier series coefficients for each of the following signals:

(a) x(t) = sin (lort +

(b) x(t) = 1 + cos (21rt)

(c) x(t) = [1 + cos (27rt)] [sin (1ort +

Hint: You may want to first multiply the terms and then use appropriate trigo­
nometric identities.

P7.4
By evaluating the Fourier series analysis equation, determine the Fourier series for
the following signals.

P7-1
Signals and Systems
P7-2

(a)

x(t)

-4 12
\ t
-3 -2 -1 0 3 4 5

-1 - -

Figure P7.4-1

(b)

x( t)

-3 -1 1 3
-2 0 2

Figure P7.4-2

P7.5
Without explicitly evaluating the Fourier series coefficients, determine which of the
periodic waveforms in Figures P7.5-1 to P7.5-3 have Fourier series coefficients with
the following properties:
(i) Has only odd harmonics
(ii) Has only purely real coefficients
(iii) Has only purely imaginary coefficients
(a)

cos t

| | t
Fu 27r

Figure P7.5-1
 Continuous-Time Fourier Series / Problems
P7-3

(b)

t
3 S

Figure P7.5-2

(c)

-5 1 2
-6 -4 -3 -l 1 2 3 4 5 6 7

Figure P7.5-3

Optional
Problems
P7.6

Suppose x(t) is periodic with period T and is specified in the interval 0 < t < T/4
as shown in Figure P7.6.

x(t)

{t
T T
8 4
Figure P7.6

Sketch x(t) in the interval 0 < t < T if

(a) the Fourier series has only odd harmonics and x(t) is an even function;
(b) the Fourier series has only odd harmonics and x(t) is an odd function.

P7.7
Let x(t) be a periodic signal, with fundamental period To and Fourier series coeffi­
cients a,. Consider the following signals. The Fourier series coefficients for each can
Signals and Systems
P7-4

be expressed in terms of the ak as in Table 4.2 (page 224) of the text. Show that the
expression in Table 4.2 is correct for each signal.
(a) x(t - to)
(b) x(-t)
(c) x*(t)
(d) x(at), a > 0 (Determine the period of the signal.)

P7.8
As we have seen in this lecture, the concept of an eigenfunction is an extremely
important tool in the study of LTI systems. The same can also be said of linear but
time-varying systems. Consider such a system with input x(t) and output y(t). We
say that a signal #(t) is an eigenfunction of the system if
#(t) - Xt)
That is, if x(t) = #(t), then y(t) = k(t), where the complex constant Xis called the
eigenvalue associatedwith #(t).
(a) Suppose we can represent the input x(t) to the system as a linear combination
of eigenfunctions $j(t), each of which has a corresponding eigenvalue Xk.
± *

xMt = >7 CAMx(

k= -o

Express the output y(t) of the system in terms of {Ck}, {O(t)}, and {Xk}.
(b) Show that the functions #(O = tk are eigenfunctions of the system character­
ized by the differential equation
d x(t) dx(t)
dt' dt
For each $k(t), determine the corresponding eigenvalue Xk.

P7.9

In the text and in Problem P4.10 in this manual, we defined the periodic convolution
of two periodic signals 2 1 (t) and 2 2(t) that have the same period To. Specifically, the
periodic convolution of these signals is defined as

9(t) = 2 1 (t) @ 2 2 (t) = 21(r)t 2 (t - r) dr (P7.9-1)

As shown in Problem P4.10, any interval of length To can be used in the integral in
eq. (P7.9-1), and 9(t) is also periodic with period To.
(a) If 2 1 (t), t 2(t), and 9(t) have Fourier series representations
+o+o0 +o0
= aike2r/o~t, 2(t) = ( bkeik( 2,rot, tt) = (3 [Link]
k=-o0 k=-o0 k=-o

show that Ck = Toakbk.

(b) Consider the periodic signal t(t) depicted in Figure P7.9-1. This signal is the
result of the periodic convolution of another periodic signal, 2(t), with itself.
 Continuous-Time Fourier Series / Problems
P7-5

Find z(t) and then use part (a) to determine the Fourier series representation
for 1(t).

x(t)

7 -5 -3 -2 0 2 3 5 7 8

Figure P7.9-1

(c) Suppose now that x 1 (t) and x 2 (t) are the finite-duration signals illustrated in
Figure P7.9-2(a) and (b). Consider forming the periodic signals 2 1(t) and 2 2 (t),
which consist of periodically repeated versions of x 1(t) and x 2 (t) as illustrated
for 2 1 (t) in Figure P7.9-2(c). Let y(t) be the usual, aperiodic convolution of x 1(t)
and x 2(t),
y(t) = x 1 (t) * X2(),
and let P(t) be the periodic convolution of t 1(t) and t 2 (t),
X0t = 21l(t) @ 2t2(t)

Show that if To is large enough, we can recover y(t) completely from one period
of P(t), that is,

9(t), |t| 5 To/2,

YM- 0, |t | > To/2

x1 (t)

11

I I I t
-T T
(a)
x2 (t)

et e

-T T
(b)

x1 (t)

"t
To -T T To

(c)

Figure P7.9-2
Signals and Systems
P7-6

P7.10
The purpose of this problem is to show that the representation of an arbitrary peri­
odic signal by a Fourier series, or more generally by a linear combination of any set
of orthogonal functions, is computationally efficient and in fact is very useful for
obtaining good approximations of signals. (See Problem 4.7 [page 254] of the text
for the definitions of orthogonal and orthonormal functions.)
Specifically, let {#o(t)}, i = 0, + 1, i 2, . . . , be a set of orthonormal functions on
the interval a t b, and let x(t) be a given signal. Consider the following approx­
imation of x(t) over the interval a t - b:
+N
N~t
= ajo(t), (P7. 10-i1)
i=-N

where the a, are constants (in general, complex). To measure the deviation between
x(t) and the series approximation tN(t), we consider the error eN(t) defined as
eN(t) = x(t) - tNAht) (P7.10-2)
A reasonable and widely used criterion for measuring the quality of the approxi­
mation is the energy in the error signal over the interval of interest, that is, the
integral of the squared-error magnitude over the interval a - t 5 b:

E= leN(t) 2 dt (P7.10-3)

(a) Show that E is minimized by choosing

b
ai = f x(t)$7(t) dt (P7.10-4)

Hint: Use eqs. (P7.10-1) to (P7.10-3) to express E in terms of aj, 4i(t), and x(t).
Then express ai in rectangular coordinates as ai = bi + jci, and show that the
equations
9E OE
- = 0 and -=0, i = 0, +1, +2, . .. , iN,
abi aci
are satisfied by the ai as given in eq. (P7.10-4).
(b) Determine how the result of part (a) changes if the {#o(t)} are orthogonal but not
orthonormal, with

Ai = |fa
I )|12 dt

(c) Let #n(t) = e "Oo'and choose any interval of length To = 27r/wo. Show that the ai
that minimize E are as given in eq. (4.45) of the text (page 180).
MIT OpenCourseWare
[Link]

Resource: Signals and Systems

Professor Alan V. Oppenheim

The following may not correspond to a particular course on MIT OpenCourseWare, but has been
provided by the author as an individual learning resource.

For information about citing these materials or our Terms of Use, visit: [Link]
8 Continuous-Time
Fourier Transform
Recommended
Problems
P8.1

Consider the signal x(t), which consists of a single rectangular pulse of unit height,
is symmetric about the origin, and has a total width T1.
(a) Sketch x(t).
(b) Sketch t(t), which is a periodic repetition of x(t) with period To = 3T 1 /2.
(c) Compute X(w), the Fourier transform of x(t). Sketch IX(w) I for Iw s 6w/Ti.
(d) Compute a,, the Fourier series coefficients of t(t). Sketch ak for k = 0 +1,
+2, 3.
(e) Using your answers to (c) and (d), verify that, for this example,
1
ak = 1-XM (w) /

(f) Write a statement that indicates how the Fourier series for a periodic function
can be obtained if the Fourier transform of one period of this periodic function
is given.

P8.2
Find the Fourier transform of each of the following signals and sketch the magni­
tude and phase as a function of frequency, including both positive and negative
frequencies.
(a) b(t - 5)
(b) e -a'u(t), a real, positive
2
(c) e (-I+j )tu(t)

P8.3
In this problem we explore the definition of the Fourier transform of a periodic
signal.
(a) Show that if x 3(t) = axi(t) + bx 2 (t), then X3(w) = aXi(w) + bX 2 (w).
(b) Verify that

e-= - 2,7b(w - wo)ej-t dw

From this observation, argue that the Fourier transform of eiwO' is 27rb(w - wo).
(c) Recall the synthesis equation for the Fourier series:

k= -00

P8-1
Signals and Systems
P8-2

By taking the Fourier transform of both sides and using the results to parts (a)
and (b), show that
2irk
k~)= 2ra T
(d) Sketch X(w) for your answer to Problem P8.1(d) for I oI 4ir/To.

P8.4

(a) Consider the often-used alternative definition of the Fourier transform, which
we will call X,(f). The forward transform is written as

2
Xa(f) = f x(t)e -j f t dt,

where f is the frequency variable in hertz. Derive the inverse transform formula
for this definition. Sketch X(f) for the signal discussed in Problem P8.1.
(b) A second, alternative definition is
1 .
Xb(v) = 1 x(t)ej" dt

Find the inverse transform relation.

P8.5
Consider the periodic signal t(t) in Figure P8.5-1, which is composed solely of
impulses.

X(t)

2 'e ' t
-7 -6 -S -4 -3 -2 -1 0 1 2 3 4 5 6 7
Figure P8.5-1

(a) What is the fundamental period To?

(b) Find the Fourier series of I(t).
(c) Find the Fourier transform of the signals in Figures P8.5-2 and P8.5-3.
(i) I I
 Continuous-Time Fourier Transform / Problems
P8-3

(ii)
x 2 (t)

1
1
2t

0 1 2 3 4 5

Figure P8.5-3

(d) 2(t) can be expressed as either xi(t) periodically repeated or x 2(t) periodically
repeated, i.e.,

)= xi(t - kTi), or (P8.5-1)

k=--00

X)= E x 2(t - kT2 ) (P8.5-2)

k= -oo

Determine T, and T2 and demonstrate graphically that eqs. (P8.5-1) and

(P8.5-2) are valid.
(e) Verify that the Fourier series of X(t) is composed of scaled samples of either
XI(w) or X 2(w).

P8.6
Find the signal corresponding to the following Fourier transforms.
1

-­
(a) X.(w) = .

(b)

2
Xb(W)

-7 7
Figure P8.6-1
1
(c) Xe(W) = 9 + W 2
See Example 4.8 in the text (page 191).
(d) Xd(w) = X.(x)Xb(w), where X(w) and Xb(w) are given in parts (a) and (b), respec­
tively. Try to simplify as much as possible.
Signals and Systems
P8-4

(e)
IXe(Co)l
4 Xe o
1 -­

-1 +1

Figure P8.6-2

Optional
Problems
P8.7

In earlier lectures, the response of an LTI system to an input x(t) was shown to be
y(t) = x(t) * h(t), where h(t) is the system impulse response.

(a) Using the fact that

y(t) = x(t) * h(t) = x(r)h(t - r) dr,

show that

Y(w) = f f x(r-)h(t - r)e j dr dt

(b) By appropriate change of variables, show that

Y(w) = X(w)H(w),
where X(w) is the Fourier transform of x(t), and H(w) is the Fourier transform
of h(t).

P8.8

Consider the impulse train

p(t) = ( ot - kT)
k=-0

shown in Figure P8.8-1.

 Continuous-Time Fourier Transform / Problems
P8-5

p (t)

t
-2T -T 0 T 2T

Figure P8.8-1

(a) Find the Fourier series of p(t).

(b) Find the Fourier transform of p(t).
(c) Consider the signal x(t) shown in Figure P8.8-2, where Ti < T.

x( t)

-T _T T1 T
2 2

Figure P8.8-2

Show that the periodic signal t(t), formed by periodically repeating x(t),
satisfies
2(t) = x(t)* p(t)

(d) Using the result to Problem P8.7 and parts (b) and (c) of this problem, find the
Fourier transform of t(t) in terms of the Fourier transform of x(t).
MIT OpenCourseWare
[Link]

Resource: Signals and Systems

Professor Alan V. Oppenheim

The following may not correspond to a particular course on MIT OpenCourseWare, but has been
provided by the author as an individual learning resource.

For information about citing these materials or our Terms of Use, visit: [Link]
9 Fourier Transform Properties
Recommended
Problems
P9.1
Determine the Fourier transform of x(t) = e-t/u(t) and sketch
(a) IX(w)I
(b) <tX(w)

(c) Re{X(w)}
(d) Im{X(w)}

P9.2

Figure P9.2 shows real and imaginary parts of the Fourier transform of a signal
x(t).

Re Ix (C) ]In I(X(O))

-W W -W W
Figure P9.2

(a) Sketch the magnitude and phase of the Fourier transform X(W).
(b) In general, if a signal x(t) is real, then X(-w) = X*(w). Determine whether x(t)
is real for the Fourier transform sketched in Figure P9.2.

P9.3

Determine which of the Fourier transforms in Figures P9.3-1 and P9.3-2 correspond
to real-valued time functions.
(a)
IX(W)| 4 X(co)

CF CO

Figure P9.3-1

P9-1
Signals and Systems
P9-2

(b)
Rel X()t = XR (W) Im IX(W)= X, (W)

2
lr

-W W
I
Figure P9.3-2

P9.4
(a) By considering the Fourier analysis equation or synthesis equation, show the
validity in general of each of the following statements:
(i) If x(t) is real-valued, then X(o) = X*(-c).
(ii) If x(t) = x*(- t), then X(w) is real-valued.
(b) Using the statements in part (a), show the validity of each of the following
statements:
(i) If x(t) is real and even, then X(o) is real and even.
(ii) If x(t) is real and odd, then X(w) is imaginary and odd.

P9.5
(a) In the lecture, we derived the transform of x(t) = e -'u(t).Using the linearity
and scaling properties, derive the Fourier transform of e -"t = x(t) + x(- t).
(b) Using part (a) and the duality property, determine the Fourier transform of
1/(1 + t2
(c) If
1
r(t) = 1 + (3t) 2
find R(w).
(d) x(t) is sketched in Figure P9.5. If y(t) = x(t/2), sketch y(t), Y(w), and X(w).

x(t)

-T T

Figure P9.5
 Fourier Transform Properties / Problems
P9-3

P9.6

Show the validity of the following statements:

(a) x(O) = - X(w) dw

27r -o

(b) X(O) = x(t) dt

P9.7

The output of a causal LTI system is related to the input x(t) by the differential
equation
ly(t) + 2y(t) = x(t)
dt
(a) Determine the frequency response H(w) = Y(w)/X(w) and sketch the phase and
magnitude of H(w).
(b) If x(t) = e- tu(t), determine Y(w), the Fourier transform of the output.
(c) Find y(t) for the input given in part (b).

P9.8

By first expressing the triangular signal x(t) in Figure P9.8 as the convolution of a
rectangular pulse with itself, determine the Fourier transform of x(t).

x (t)

t
-2 2
Figure P9.8

Optional
Problems
P9.9
Using Figure P9.9-1, determine y(t) and sketch Y(w) if X(w) is given by Figure
P9.9-2. Assume w, > wo.
Signals and Systems
P9-4

x(t) I N y(t)

cos (wet)

Figure P9.9-1

X(W)

-W0 W0
Figure P9.9-2

P9.10
Compute the Fourier transform of each of the following signals:
(a) [e~a cos wotlu(t), a > 0
(b) e 31 sin 2t

(C) sin irt sin 2irt

rt rt

P9.11
Consider the following linear constant-coefficient differential equation (LCCDE):
dy(t) + 2y(t) = A cos
wet
dt
Find the value of wo such that y(t) will have a maximum amplitude of A/3. Assume
that the resulting system is linear and time-invariant.

P9.12
Suppose an LTI system is described by the following LCCDE:
d'ytt) 2dytt) 4dx(t)
+ + 3y(t) = - x(t)
dt2 dt dt
 Fourier Transform Properties / Problems
P9-5

(a) Show that the left-hand side of the equation has a Fourier transform that can
be expressed as

A(w)Y(w), where Y(w) = J{y(t)}

Find A(w).
(b) Similarly, show that the right-hand side of the equation has a Fourier transform
that can be expressed as
B(w)X(w), where X(w) = {x(t)}

(c) Show that Y(w) can be expressed as Y(w) = H(w)X(w) and find H(w).

P9.13
From Figure P9.13, find y(t) where

x(t) = sin(wot) and h(t) = sin(2wot)

t t

x(t) h(t) yt)

Figure P9.13

P9.14
(a) Determine the energy in the signal x(t) for which the Fourier transform X(w) is
given by Figure P9.14.

(b) Find the inverse Fourier transform of X(w) of part (a).

Signals and Systems
P9-6

P9.15

Suppose that the system F takes the Fourier transform of the input, as shown in
Figure P9.15-1.

x(t) i F - y(t)=2nX(--w)

Figure P9.15-1

What is w(t) calculated as in Figure P9.15-2?

xt)M F O F F 0 w(t)

Figure P9.15-2

P9.16

Use properties of the Fourier transform to show by induction that the Fourier trans­
form of
tn-i
x(t) = - 1-U(t), a > 0
(n - 1)!

1
X(w) =
(a + jo)"
MIT OpenCourseWare
[Link]

Resource: Signals and Systems

Professor Alan V. Oppenheim

The following may not correspond to a particular course on MIT OpenCourseWare, but has been
provided by the author as an individual learning resource.

For information about citing these materials or our Terms of Use, visit: [Link]
10 Discrete-Time Fourier Series
Recommended
Problems
P10.1
Consider a discrete-time system with impulse response
h[n] = (I)"U[n]

Determine the response to each of the following inputs:

(a) x[n] = (-1)" = e-i'n for all n
(b) x[n] = ein/4 ) for all n

(c) x[n] = cos + 8 for all n

P10.2

Consider the following two periodic sequences:

!,[n] = 1 + sin 2rn (P10.2-1)

210 J
22[n] = 1 + sin n +I (P10.2-2)

(a) Determine the period of xi[n] and of x2[n].

(b) Determine the sequence of Fourier series coefficients alk for xi[n] and a 2k for
X2[fl].
(c) In each case, the sequence of Fourier series coefficients is periodic. Determine
the period of the sequence ai, and the sequence a2k.

P10.3
Determine the Fourier series coefficients for the three periodic sequences shown in
Figures P10.3-1 to P10.3-3. Since these three sequences all have the same nonzero
values over one period, we suggest that you first determine an expression for the
envelope of the Fourier series coefficients and then sample this envelope at the
appropriate spacings in each case.
(a)

P10-1
Signals and Systems
P10-2

(b)

1 2 [n]

-1 0 1 12

Figure P10.3-2

(c)

-1
III*****1
0 1
1 60
00-*p *see n

Figure P10.3-3

P10.4
(a) Determine and sketch the discrete-time Fourier transform of the sequence in
Figure P10.4-1.

x [n]

0 1 2
T 3
2
4
n

Figure P10.4-I

(b) Using your result in part (a), determine the discrete-time Fourier series of the
two periodic sequences in Figure P10.4-2.

n
-2[ I

2 0 n

0 1 2 3 4 5

Figure P10.4-2
 Discrete-Time Fourier Series / Problems
P10-3

P10.5
Consider the signal x[n] depicted in Figure P10.5. This signal is periodic with period
N = 4. The signal x[n] can be expressed in terms of a discrete-time Fourier series:
3

x[n] = akejk(2r/4)n (P10.5-1)

k=O

As mentioned in the text, one way to determine the Fourier series coefficients is to
treat eq. (P10.5-1) as a set of four linear equations [eq. (P10.5-1) for n = 0, 1, 2, 3]
in the four unknowns (aO, ai, a 2, and a3 ).

x[n]

n
-8 1-4 0 4 8 16
12
Figure P10.5

(a) Explicitly write out the four equations and solve them directly using any stan­
dard technique for solving four equations in four unknowns. (Be sure to first
reduce the complex exponentials to the simplest form.)
(b) Check your answer by calculating the coefficients ak directly, using the Fourier
series analysis equation

ak = Mne-k( 2 /4)n
4n=O

P10.6
Figure P10.6 shows a real periodic signal x[n]. Using the properties of the Fourier
series and without explicitly evaluating the Fourier series coefficients, determine
whether the following are true for the Fourier series coefficients ak.
(a) a, = a +10 for all k
(b) a = a _ for all k
(c) akejk(2,/) is real for all k
(d) ao = 0

x [n]

u 1 2 3 4
6 7 8 9 0
15
- 5 - 4 - 3 -2 -

Figure P10.6
Signals and Systems
P10-4

Optional
Problems
P10.7
In parts (a)-(d) we specify the Fourier series coefficients of a signal that is periodic
with period 8. Determine the signal x[n] in each case.

(a) ak = cos (k 4)+ sin (3k 4)

sin kr 0 k - 6
(b) ak = ( 3)
10, k =7

(c) ak as in Figure P10.7(a)

(d) ak as in Figure P10.7(b)

II~I 111111111 III~I II 111-111

.111 ~I
n

-8 0 8 16

Figure P10.7

P10.8
(a) Consider a linear, time-invariant system with impulse response
h[n] = (1)I"I

Find the Fourier series representation of the output 9[n] for each of the follow­
ing inputs.
f/37rn\
(i) t[n] = sin 34)

(ii) 1[n] = [n - 4k]

k= -0

(iii) 1[n] is periodic with period 6, and

1[n]= 1,
0,
n
n
=

=
0, ±1
±2, 3, ±4

(iv) 1[n] = j" + (- 1)"

 Discrete-Time Fourier Series / Problems
P10-5

(b) Repeat part (a) for

1, O ns2
h[n]= < 1, -2 ! n - -1
0 otherwise

P10.9

Let I[n] be a periodic sequence with period N and Fourier series representation
2[n] = 1 akejk(2/N)n (P10.9-1)
k=(N)

The Fourier series coefficients for each of the following signals can be expressed in
terms of the coefficients ak in eq. (P10.9-1). Derive these expressions.
(a) I[n - no]
(b) 1[n] - 2[n - 1]

(c) 2[n] - I N] (assume that N is even)

(d) I[n] + x N] (assume that N is even; note that this signal is periodic
with period N/2)

(e) 2*-n]

P10.10
Consider two specific periodic sequences t[n] and 9[n]. X[n] has period N and 9[n]
has period M. The sequence '[n] is defined as W[n] = I[n] + 9[n].
(a) Show that 7b[n] is periodic with period MN.
(b) Since I[n] has period N, its discrete Fourier series coefficients ak also have
period N. Similarly, since 9[n] has period M, its discrete Fourier series coeffi­
cients bk also have period M. The discrete Fourier series coefficients of qb[n], Ck,
have period MN. Determine Ck in terms of a and bk.

P10.11
Determine the Fourier series coefficients for each of the following periodic discrete-
time signals. Plot the magnitude and phase of each set of coefficients a.

(a) x[n] = sin 7r'n 1

(b) x[n] = cos (2;n) 7

(c) x[n] = cos (4b n

3)
MIT OpenCourseWare
[Link]

Resource: Signals and Systems

Professor Alan V. Oppenheim

The following may not correspond to a particular course on MIT OpenCourseWare, but has been
provided by the author as an individual learning resource.

For information about citing these materials or our Terms of Use, visit: [Link]
11 Discrete-Time Fourier Transform
Recommended
Problems
P11.1

Compute the discrete-time Fourier transform of the following signals.

(a) x[n] = (A)"u[n]
(b) x[n] = (a" sin Gon)u[nj, lal < 1
(c) x[n] as shown in Figure P11.1

x [/n|

I'll

0 1 2 3

Figure P11.1

(d) x[n] = ( )"U[n + 2]

P11.2
(a) Consider the linear constant coefficient difference equation
y[n] - ly[n - 1] =x[n],

which describes a linear, time-invariant system initially at rest. What is the sys­
tem function that describes Y(Q) in terms of X(Q)?
(b) Using Fourier transforms, evaluate y[n] if x[n] is
(i) b[n]
(ii) 6[n - no]
(iii) (j)"n]

P11.3
(a) Consider a system with impulse response

h[n] = (1)cos u[n]

Determine the system transfer function H(Q).

(b) Suppose that x[n] = cos(7rn/2). Determine the system output y[n] using the
transfer function H(Q) found in part (a).

P11.4
A particular LTI system is described by the difference equation
y[n] + {y[n - 1] - ly[n - 2] = x[n] - x[n - 1]

P11-1
Signals and Systems
P11-2

(a) Find the impulse response of the system.

(b) Evaluate the magnitude and phase of the system frequency response at 9 = 0,
9 = r/4, 9 = -r/4, and 0 = 9r/4.

P11.5
x[n] is a finite-duration signal of length N so that x[n] = 0, n < 0 and n > N - 1.
The discrete-time Fourier transform of x[n] is denoted by X(Q). We generate the
periodic signal 9[n] by periodically replicating x[n], i.e.,

9[n] = ( x[n + rN]

r= -cc

(a) Write the expression in terms of x[n] for the Fourier series coefficients ak of
R[n].
(b) Write an expression relating the Fourier series coefficients of 9[n] to the Fourier
transform of x[n].

P11.6
(a) Four different transforms have been introduced thus far:
I. Continuous-time Fourier series
II. Discrete-time Fourier series
III. Continuous-time Fourier transform
IV. Discrete-time Fourier transform
In the following table, fill in the blanks with I, II, III, or IV depending on which
transform(s) can be used to represent the signal described on the left. Finite
duration means that the signal is guaranteed to be nonzero over only a finite
interval.

Signal Description Transform

Continuous time Infinite duration Periodic

Continuous time Infinite duration Aperiodic

Continuous time Finite duration Aperiodic

Discrete time Infinite duration Periodic

Discrete time Infinite duration Aperiodic

Discrete time Finite duration Aperiodic

(b) Which of the transforms in the preceding table possess the duality property
summarized in Sections 4.6.6 and 5.9.1 of the text?
(c) Which of the transforms are always periodic?
 Discrete-Time Fourier Transform / Problems
P11-3

Optional
Problems
P11.7

For continuous-time signals, we saw that

if x(t) - X(w), then x(at) ­

|al a
Is there a similar property for discrete-time signals? If so, what is it? If not, why
not?

P11.8
If x[n] and X(Q) denote a sequence and its Fourier transform, determine in terms of
x[n] the sequence corresponding to
(a) X(Q - 00)
(b) Re{X(Q)}
(c) Im{X(Q)}
(d) |X(12)|2
Hint: Write your answer in terms of a convolution.

P11.9
Suppose we have an LTI system characterized by an impulse response
. irn
sin
h[n] = s
,rn
(a) Sketch the magnitude of the system transfer function.
(b) Evaluate y[n] = x[n] * h[n] when

x[n] = (-1)n cos n

4

P11.10
A particular discrete-time system has input x[n] and output y[n]. The Fourier trans­
forms of these signals are related by the following equation:

Y(Q) = 2X(Q) + e dXQ)

-"XCg)

d12
(a) Is the system linear? Clearly justify your answer.
(b) Is the system time-invariant? Clearly justify your answer.
(c) What is y[n] if x[n] = b[n]?
Signals and Systems
P11-4

P11.11
Consider a discrete-time sequence I[n] that is periodic with period N. We know that
.t[n]can be written as
t[n] )=7 ake jk(2x/N)n
k=(N)

(a) Show that by multiplying both sides of the equation by e -jl(2w/N)n and summing
over one period, the discrete-time Fourier series coefficients ak are obtained as

ak = 1 E tlnje -jk(21r/N)n
n=(N)

(b) The synthesis equation for an aperiodic discrete-time signal can be written as

x[n] = 1 X(Q)e I dQ

(i) Show that by multiplying both sides by e -j1" and summing over n =
-oo to n = 0o,

x[nle"3"= X(12) e"-" " di

(ii) Show that

3
n==-o
ejj-)n = 27r
n=­
( - 91 + 21rn)

Hint: Consider E* exo-Qan as the Fourier series representation of

some continuous-time periodic function.
(iii) By combining the results of parts (i) and (ii), establish that

( x[ne -1"" = X(Q)

P11.12
The Fourier transform of a discrete-time periodic signal is based on the fact that
such a series can be written as
l[n] = ake jk(2w/N)n
k=(N)

(a) Establish that the Fourier transform of ejk( 2 1/N)n is

00 2xoO + 2x
127r(- N +27n

(b) Establish that the Fourier transform of

T, ake jk(27/N)n
k =(N)

is

n=-o
S27 k=(N)
ak - 29-k + 2.7rn
N
 Discrete-Time Fourier Transform / Problems
P11-5

(c) Establish that

00 2xk 27rk
2,r (: ako Q N + 2,rn =2,r ako 0 -N '
n=-o0 k=(N) k=-0 N

which shows that

( 21rk
2[n] - 2 T,r o
ak - N J
(d) Use the result in part (c) to verify that the Fourier series coefficients
1
ak= - X(Q)
N 1=(2rk)/N

where X(Q) is the Fourier transform of x[n], which consists of a single period
of t[n].
MIT OpenCourseWare
[Link]

Resource: Signals and Systems

Professor Alan V. Oppenheim

The following may not correspond to a particular course on MIT OpenCourseWare, but has been
provided by the author as an individual learning resource.

For information about citing these materials or our Terms of Use, visit: [Link]
12 Filtering
Recommended
Problems
P12.1

Consider a lowpass filter with real frequency response H(w) as shown in Figure
P12.1.

H(o)

3 3

Figure P12.1

(a) Which of the following properties does the filter impulse response have?
(i) Real-valued
(ii) Complex-valued
(iii) Even
(iv) Odd
(v) Causal
(vi) Noncausal
(b) Consider the filter input

x(t) = 6 (t - 9n)

Sketch and label the Fourier transform of the filter output y(t).
(c) Determine the filter output for the input considered in part (b).

P12.2
Consider the system shown in Figure P12.2-1, where the frequency response H(w)
has magnitude and phase shown in Figure P12.2-2.

X1 (t)

+ H(co) y(t)

X2 (t)

Figure P12.2-1

P12-1
Signals and Systems
P12-2

IH(w)|

--4r 41T

4H(w)

Figure P12.2-2

xi(t) = sin [wit + (7r/4)] and

x 2(t) = 2 cos [w2 t - (r/3)],
where wi = 7r and C2 = 27r, write an expression for y(t).

P12.3
Consider the first-order RC circuit shown in Figure P12.3-1.

+ v,(t) ­

vs(t) R vr(t)

Figure P12.3-1

(a) Determine Hi(w), the transfer function from v, to vc, as shown in Figure P12.3-2.
Sketch the magnitude and phase of H 1 (w).

Vs(t) -- - i e)P2.3(t)

Figure P12.3-2
 Filtering / Problems
P12-3

(b) Evaluate H 2 (W), the transfer function from vs to v,, as shown in Figure P12.3-3.

s(t) i H2( P r(t)

Figure P12.3-3

Sketch the magnitude and phase of H 2 (w).

(c) What are the cutoff frequencies for H1 (w), H 2(w)? (For this problem, the cutoff
frequency is defined as the frequency at which the magnitude of the frequency
responses is 1/V2 times its maximum value.)
(d) Suppose we now consider the system shown in Figure P12.3-4.

Sketch V(w)/V,(w). What is the corresponding cutoff frequency?

P12.4

Figure P12.4 shows the frequency response H(Q) of a discrete-time differentiator.

Determine the output signal y[n] as a function of 90 if the input x[n] is
x[n] = cos [Qon + 0]

|H( 2)|

4H(92)

71
- fI
IT
2 K
Figure P12.4
Signals and Systems
P12-4

P12.5

Consider the following two LTI systems

System 1: y 1[n] = x[n] + x[n - 1]

2
System 2: y2[n] = x[nJ - x[n ­
2
(a) Without calculating the respective system functions determine the following.
(i) Is system 1 a lowpass filter, highpass filter, or bandpass filter?
(ii) Is system 2 a lowpass filter, highpass filter, or bandpass filter?
Clearly give your reasoning.
(b) Calculate the frequency responses H(Q) and H 2(Q) for systems 1 and 2 and plot
their magnitudes for the range of 9 between - 2r and 2 1r.

P12.6

In the lecture we discussed the use of a moving average as a lowpass filter. Here we
study this idea a little more closely.
(a) Let
1 N
y 1 [n] = 2N+ 1 x[n-k]
Find the impulse response h[n].
(b) In eq. (6.15) in the text (page 416), we find that Hi(Q) is given by

sn02N + 1
1 sn 2
2N + 1 n .
sin ­

Consider a new filter y 2[n] = x[n] - y,[n]. Find H 2 (Q).

(c) Plot IHI(O)| and |H2(Q)| on a linear scale, carefully noting where HI(Q) or
H 2(Q) equals 1 or 0.
(d) What type of filter is H 2 ()?

P12.7

We want to design a nonideal continuous-time lowpass filter with the specifications

indicated in Figure P12.7.
Furthermore, we require that H(0) = 1. We are restricted to designing a filter
with a transfer function of the form

_Ka + jw
 Filtering / Problems
P12-5

1.0

2 6

Figure P12.7

(a) Find K such that H(0) = 1.

(b) Find the range of values of a such that the resulting filters will meet the speci-
fications in the figure.

Optional
Problems
P12.8
A causal LTI filter has the frequency response H(w) = -2jw. For each of the follow­
ing input signals, determine the filtered output signal y(t).
(a) x(t) = e' t
(b) x(t) = (sin wt)u(t)
1
(c) XMo) = .wjw + 6)
1
(d) X(w) = .w + 2

P12.9
It was stated in Section 6.4 of the text that for a discrete-time filter to be causal and
have exactly linear phase, its impulse response must be of finite length and conse­
quently the difference equation must be nonrecursive. To focus on the insight behind
this statement, we consider a particular case for which the slope of the phase is an
integer. Thus, the frequency response is assumed to be of the form
H(Q) = H,(Q)e -j, -r < <ir, (P12.9-)
where H,(Q) is real and even. Let h[n] denote the impulse response of the filter with
frequency response H(Q) and hn] denote the impulse response of the filter with
frequency response H,(Q).
Signals and Systems
P12-6

(a) By using the appropriate properties in Table 5.1 of the text (page 335), show
the following.
(i) h,[fn = h,[-n] (i.e., h,[n] is symmetric about n = 0)
(ii) h[n] = h,[n - M]
(b) Using the result in part (a), show that with H(Q) of the form in eq. (P12.9-1),
h[n] is symmetric about n = M, that is,

h[M + n] = h[M - n] (P12.9-2)

(c) According to the result in part (b), the linear phase characteristic in eq.
(P12.9-1) imposes a symmetry in the impulse response. Show that if h[n] is
causal and has the symmetry in eq. (P12.9-2), then
h[n] = 0, n<0, n>2M,
i.e., it must be of finite length.

P12.10
In Figure P12.10-1, we show a discrete-time system consisting of a parallel combi­
nation of N LTI filters with impulse response h,[n], k = 0, 1, . . . , N - 1. For any k,
hk[n] is related to hon] by the expression
hk[n] = ej(27rnk/N)h 0[nI

N -I
y[n] = E Yk [n
- k =0

Figure P12.10-1

(a) If ho[n] is an ideal discrete-time lowpass filter with frequency response HO(Q) as
shown in Figure P12.10-2, sketch the Fourier transforms of hi[n] and hN- 1[n] for
9 in the range -7r < - +7r.
(b) Determine the value of the cutoff frequency 9, in Figure P12.10-2 in terms of N
(0 <Q, sr) such that the system of Figure P12.10-2 is an identity system; that
is, y[n] = x[n] for all n and any input x[n].
 Filtering / Problems
P12-7

1T I1

Figure P12.10-2

(c) Suppose that h[n] is no longer restricted to be an ideal lowpass filter. If h[n]
denotes the impulse response of the entire system in Figure P12.10-1 with input
x[n] and output y[n], then h[n] can be expressed in the form h[n] = r[n]ho[n].
Determine and sketch r[n].
(d) From your result in part (c), determine a necessary and sufficient condition on
ho[n] to ensure that the overall system will be an identity system (i.e., such that
for any input x[n], the output y[n] will be identical to x[n]). Your answer should
not contain any sums.

P12.11
Consider the system in Figure P12.11-1.

+
x(t) + y(t)

G(w)

Figure P12.11-1

Let G(w) be real and have the form shown in Figure P12.11-2.

G(o)

- -1

-C02 ~1 W1 W2

Figure P12.11-2

Plot the resulting frequency response magnitude and phase for the following cases.
(a) a > 1
(b) 1 > a > 0
(c) a < 0
MIT OpenCourseWare
[Link]

Resource: Signals and Systems

Professor Alan V. Oppenheim

The following may not correspond to a particular course on MIT OpenCourseWare, but has been
provided by the author as an individual learning resource.

For information about citing these materials or our Terms of Use, visit: [Link]
13 Continuous-Time Modulation
Recommended
Problems
P13.1

c(t)

x(t) x ;y(t)

Figure P13.1-1

In the amplitude modulation system in Figure P13.1-1, the input x(t) has the Fourier
transform shown in Figure P13.1-2.

IX(W)i 4X(o)

IT
2

Figure P13.1-2

For each choice of carrier c(t) in the following list, draw the magnitude and phase
of Y(w), the Fourier transform of y(t).

(a) c(t) = eisect

(b) c(t) = es'cturm
(c) c(t) = cos 3ct
(d) c(t) = sin 3wct
(e) c(t) = cos 3wct + sin 3wt

P13.2
Consider the system in Figure P13.2-1.

P13-1
Signals and Systems
P13-2

x(t) X r X X y
Y
~C WC CC WC
s(t) m(t) d(t)

Figure P13.2-1

The Fourier transform of s(t) is given by Figure P13.2-2.

S(W)

-2wC 0 2WC
Figure P13.2-2

The Fourier transform of x(t) is given by Figure P13.2-3.

X(W)

Figure P13.2-3

For which of the following choices for m(t) and d(t) is y(t) nonzero?
m(t) d(t)

(a) 1 1
(b) cos oct cos Wct
(c) sin wct sin wt
(d) cos 2wet cos 2wet
(e) cos 2wct cos oct

P13.3
In Section 7.1 of the text, we discussed the effect of a loss in synchronization in
phase between the carrier signals in the modulator and demodulator for sinusoidal
amplitude modulation. Specifically, we showed that the output of the demodulator
is attenuated by the cosine of the phase difference; in particular, when the modu­
lator and demodulator have a phase difference of ir/2, the demodulator output is
zero. As we demonstrate in this problem, it is also important to havefrequency syn­
chronization between the modulator and demodulator.
 Continuous-Time Modulation / Problems
P13-3

Consider the amplitude modulation and demodulation systems in Figure

P13.3-1, with 0e = 0 and with a change in frequency of the demodulator carrier such
that
w(t) = y(t) cos wAdt

where
y(t) = x(t) cos Wct

cos (oct + 0C)

Lowpass filter

cos (co t + 0e )

Figure P13.3-1

Let us denote the frequency difference between the modulator and demodulator as
Aw [i.e., (Wd - Wc) = Aw]. Also, assume that x(t) is bandlimited with X(W) = 0 for
Iwl > wm and assume that the cutoff frequency W of the lowpass filter in the demod­
ulator satisfies the inequality
(wM + |IA|) < W < (2we + Awl -Mo)
(a) Show that the output of the demodulator lowpass filter is proportional to
x(t)cos(AWt).
(b) If the spectrum of x(t) is that shown in Figure P13.3-2, sketch the spectrum of
the output of the demodulator.

X(o)
1

- M P M

Figure P13.3-2
Signals and Systems
P13-4

You may find it useful to use the trigonometric identity

cos A cos B = i[cos(A - B) + cos(A + B)]

P13.4
As discussed in Section 7.1.1 of the text, asynchronous modulation-demodulation
requires the injection of the carrier signal so that the modulated signal is of the form
y(t) = [A + x(t)]cos(wet + Or), (P13.4-1)
where [A + x(t)] > 0 for all t. The presence of the carrier means that more trans­
mitter power is required, representing an inefficiency.
(a) Let x(t) be given by x(t) = cos omt with wm < we and [A + x(t)] > 0. For a
periodic signal y(t) with period T, the time average power P, is defined as P, =
(1/T) fTy 2 (t)dt. Determine and sketch Py for y(t) in eq. (P13.4-1). Express your
answer as a function of the modulation index m, defined as the maximum abso­
lute value of x(t) divided by A.
(b) The efficiency of transmission of an amplitude-modulated signal is defined to be
the ratio of the power in the sidebands of the signal to the total power in the
signal. With x(t) = cos wMt and with wM < wc and [A + x(t)] > 0, determine and
sketch the efficiency E of the modulated signal as a function of the modulation
index m.

Optional
Problems
P13.5

Consider the modulated signal z(t) = A(t)cos(wc + 0,), where we, is known but 0c is
unknown. We would like to recover A(t) from z(t).
(a) Show that z(t) = x(t)cos wet + y(t)sin wet and express x(t) and y(t) in terms
of A(t) and 0e,.
(b) Show how to recover x(t) from z(t) by modulation followed by filtering.
(c) Show how to recover y(t) from z(t) by modulation followed by filtering.
(d) Express A(t) in terms of x(t) and y(t) with no reference to 0c and show in a
block diagram how to recover A(t) from z(t). The following trigonometric iden­
tities may be useful:
cos(A + B) = (cos A cos B) - (sin A sin B),
cos 2A = -(1 + cos 2A),
sin 2A = -(1 - cos 2A),
sin 2A = 2 cos A sin A

P13.6

A single-sideband modulation system with carrier frequency we, is shown in Figure

P13.6-1.
 Continuous-Time Modulation / Problems
P13-5

x(t) y(t)

cos Cos + c t

Figure P13.6-1

Sketch the Fourier transform of si(t), s2(t), sA(t), s4(t), sr(t), s,(t), and y(t), thus
showing that y(t) is x(t) single-sideband-modulated on the carrier We. Assume that
x(t) has the real Fourier transform shown in Figure P13.6-2 and that H(w) is a low-
pass filter as shown in Figure P13.6-3.

X(O) H(w)

2 2

Figure P13.6-2 Figure P13.6-3

P13.7
Consider the system in Figure P13.7, which can be used to transmit two real signals
over a single transmission channel.

Y1 (t)

sin wo0 t sin ot

Figure P13.7

For y 1(t) to be the same as s,(t), and y 2 (t) to be the same as s2(t), choose the proper
filter H(w) and place the proper restrictions on the bandwidth of si(t) and s2(t).
Signals and Systems
P13-6

P13.8
A commonly used system to maintain privacy in voice communications is a speech
scrambler. As illustrated in Figure P13.8-1, the input to the system is a normal
speech signal x(t) and the output is the scrambled version y(t). The signal y(t) is
transmitted and then unscrambled at the receiver.

x(t) --- Scrambler Y~)YM-Unscrambler x(t)

(Normal
speech)

Figure P13.8-1

We assume that all inputs to the scrambler are real and bandlimited to fre­
quency wM; that is, X(w) = 0 for Iw I > WM. Given any such input, our proposed
scrambler permutes different bands of the input signal spectrum. In addition, the
output signal is real and bandlimited to the same frequency band; that is, Y(w) = 0
for IwI > wM. The specific permuting algorithm for our scrambler is
Y(M) = X(W - WM), 0 <CO <COM,

Y(M) = X(O + WM), -OM < CO <0

(a) If X(w) is given by the spectrum shown in Figure P13.8-2, sketch the spectrum
of the scrambled signal y(t).

X(O)

1 -­

-WM WM
Figure P13.8-2

(b) Using amplifiers, multipliers, adders, oscillators, and whatever ideal filters you
find necessary, draw the block diagram for such an ideal scrambler.
(c) Again, using amplifiers, multipliers, adders, oscillators, and ideal filters, draw
a block diagram for the associated unscrambler.
MIT OpenCourseWare
[Link]

Resource: Signals and Systems

Professor Alan V. Oppenheim

The following may not correspond to a particular course on MIT OpenCourseWare, but has been
provided by the author as an individual learning resource.

For information about citing these materials or our Terms of Use, visit: [Link]
14 Demonstration of
Amplitude Modulation
Recommended
Problems
P14.1

Consider the AM modulation system in Figure P14.1-1.

x(t) x + y (t)= [(A+ +x(t)] cos we t

A
Cos Oct

Figure P14.1-1

K/A is called the modulation index, where K is the maximum amplitude of x(t).
Parts (a)-(c) contain plots of y(t) versus t for several different modulation indices,
with x(t) = B cos o0t. Find the modulation index for each signal.
(a)

0 .2 .4 .6 .8
Figure P14.1-2

P14-1
Signals and Systems
P14-2

(b)

0 .2 .4 .6 .8
Figure P14.1-3

(c)

.2 .4 .6 .8 1
Figure P14.1-4

P14.2

(a) Consider the signal x(t) in Figure P14.2-1.

 Demonstration of Amplitude Modulation / Problems
P14-3

x(t)

1- -

t
1 2 3 4 5 6

Figure P14.2-1

Draw y(t) for each of the following systems.

(i)

x(t)

Fi

(ii)
x(t) X )-* y(t)

cos. rt

Figure ?14.2-3

(iii)|
x(t) N y(t)

p(t)

p(t)

-- P t

--1 1 2

-1

Figure P14.2-4
Signals and Systems
P14-4

(b) Suppose that x(t) has the Fourier transform shown in Figure P14.2-5. Find
Y(w) for each case in part (a).

X(W)

ri P
2 2

Figure P14.2-5

P14.3
For each of the time waveforms (a)-(j) (Figures P14.3-1 to P14.3-10), match its pos-
sible spectrum (i)-(x) (Figures P14.3-11 to P14.3-20).
(a)
x(t)

t
T

Figure P14.3-1

(b)

(c)
x (t)

| 9 t

I gu PT

Figure P14.3-3
 Demonstration of Amplitude Modulation / Problems
P14-5

(d)
x(t)

T
10 - -4

Figure P14.3-4

(e)

(f)

x(t)

Figure P14.3-6
Signals and Systems
P14-6

(g)

x(t)

/ A
2 A\ VL~4TYV

Figure P14.3-7

(h)

x(t)

Figure P14.3-8

(1)

x(t)

A +-T

-A u2

Figure P14.3-9
 Demonstration of Amplitude Modulation / Problems
P14-7

x(t)

r~f\R(Th
00

<RI
/ xT
7
/ T
A

Figure P14.3-10

(i)

IX(f dc value contribution

decay

flf

5 3 11 35
T' T T T TT

Figure P14.3-11

(ii)

IX(f)I

decay

1F P3 13

Figure P14.3-12
Signals and Systems
P14-8

(iii)

IX(f)

T T

Figure P14.3-13

(iv)

IX(f)

3 2 1 1 2 3
T T T T T T
Figure P14.3-14

(v)

IX(f)

10 10
T T
Figure P14.3-15

(vi)
 Demonstration of Amplitude Modulation / Problems
P14-9

(vii)

IX(f)

decay
n2

5
f 1 1 3 5
T T T T T T

Figure P14.3-17

(viii)

IX(f) I

_10 10
T T
Figure P14.3-18

(ix)

|X(f

t it
11
i i I i t i
T T

Figure P14.3-19

(x)

IX(f)|
1n decay

6 2 2 6
T T T T

Figure P14.3-20

P14.4
The spectrum analyzer discussed in the lecture computed the estimate of the mag-
nitude of the Fourier transform of x,(t) by taking samples of x,(t) at equally spaced
intervals T, stopping after N samples, and computing the discrete-time Fourier
transform of the N-point sequence.
Signals and Systems
P14-10

Thus,
N-1

X(Q) =
n=O
3 x[n]e -'", where x[n] = x,(nT)

(a) Suppose x,(t) = cos w0t. Find and sketch IX(Q) 1.

(b) In any practical system, X(Q) can be explicitly calculated only at a finite set of
Q. A common choice is
2rk
=k N for K = 0,. .. ,N - 1

For the following situations, ske tch

X(27rk
X ) for K = 0, ... ,N - 1

if x,(t) = cos Wot.

(i) N = 5, wo
27r

_2r
T (2)
(3)
(ii) N = 5, wo
T S1-0
MIT OpenCourseWare
[Link]

Resource: Signals and Systems

Professor Alan V. Oppenheim

The following may not correspond to a particular course on MIT OpenCourseWare, but has been
provided by the author as an individual learning resource.

For information about citing these materials or our Terms of Use, visit: [Link]
15 Discrete-Time Modulation
Recommended
Problems
P15.1

In the system shown in Figure P15.1-1, x(t) is used to modulate an impulse train
carrier. The signal x,(t) then corresponds to an impulse train of samples of x(t).
Under appropriate conditions, x(t) can be recovered from x,(t) with an ideal low-
pass filter.

x(t) x rx(t)

pt)= F, (t-nT)

0
xp~t H(w) Wxr t)

Figure P15.1-1

ForX(w) and H(w) as indicated in Figure P15.1-2, sketch X,(w) and X,(W). Indicate
specifically whether in this case x,(t) is equal to (or proportional to) x(t).

X(w)

2 2T
H~(o)

H(wj)

T T

Figure P15.1-2

P15-1
Signals and Systems
P15-2

P15.2

Consider the discrete-time modulation system in Figure P15.2-1. Let X(Q) be given
as in Figure P15.2-2. Sketch Y(Q) for 00 = 7r/2 and for Qo = r/4.

X(2)
1
x[n] lop y[n]

IT 7T
c[n] =cos[i2n] 4 1 . 4

Figure P15.2-1 Figure P15.2-2

P15.3

The system in Figure P15.3 is equivalent to a linear, time-invariant system with

frequency response G(Q). Determine and sketch G(Q).

x (n) y[n]

H(Q)

37T 3n
4 4

Figure P15.3

P15.4
A discrete-time pulse amplitude modulation system is shown in Figure P15.4, where
p[n] and X(Q) are as indicated.
 Discrete-Time Modulation / Problems
P15-3

x[n] x y[n]

p [n]
p [n]

-2 -1 0 1 2

X(2)

7T I
4 4

Figure P15.4

(a) Sketch P(Q) and Y(Q).

(b) Describe a system to recover x[n] from y[n].
(c) Discuss how this system could be used to time-division-multiplex two signals
x1[n] and x 2 [n].

P15.5
In the system in Figure P15.5, s(t) is a rectangular pulse train as indicated.

x(t) X H(w) y(t)

s(t)
s(t)

t
-T 2 2 T

Figure P15.5

Determine H(w) so that y(t) = x(t), assuming that no aliasing has occurred.
Signals and Systems
P15-4

Optional
Problems
P15.6

Consider the discrete-time system shown in Figure P15.6. The input sequence x[n]
is multiplied by # 1[n], and the product is taken as the input to an LTI system. The
final output y[n] is then obtained as the product of the output of the LTI system
multiplied by # 2[n].

x [n] :X l-h [n] X y[n]

y

#1[n] $2y[n]
Figure P15.6

(a) In general, is the overall system linear? Is it time-invariant? (Consider, for

example, #1 = 6[n]).
(b) If # 1[n] = z-" and # 2[n] = z", where z is any complex number, show that the
overall system is time-invariant.

P15.7
In the system in Figure P15.7, two time functions xi(t) and x 2 (t) are multiplied, and
the product w(t) is sampled by a periodic impulse train. xi(t) is bandlimited to wi,
and x 2 (t) is bandlimited to 02:
X = 0, |1oI > wi,
X2(W) 0, IwI > 'W2
Determine the maximum sampling interval T such that w(t) is recoverable from
w,(t) through the use of an ideal lowpass filter.

p(t) = 5(t - nT)

x, (t)

w (t)

x 2 (t)

X 2 (o)

r/0 N_1 -
_CCo

Figure P15.7
 Discrete-Time Modulation / Problems
P15-5

P15.8
A discrete-time filter bank is to be implemented by using a basic lowpass filter and
appropriate complex exponential amplitude modulation as indicated in Figure
P15.8-1.
(a) With H(Q) an ideal lowpass filter, as shown in Figure P15.8-2, the ith channel
of the filter bank is to be equivalent to a bandpass filter with frequency response
shown in Figure P15.8-2. Determine the values of ai and f#i to accomplish this.

e j21aN - 1n ej2ngN - 1n

YN -1 [n

yi [n]

x[n]

v1 [n]

yo[n]

Figure P15.8-1

(b) Again with H(Q) as in Figure P15.8-2 and with Q, = 27ri/N, determine the value
of 00 in terms of N so that the filter bank covers the entire frequency band with­
out any overlap.
Signals and Systems
P15-6

H(92)

-1 ­

F 1 -I I I

220

Figure P15.8-2

P15.9
Consider the modulation system in Figure P15.9.

x(t) P.X
r4y(t)

s(t)
Figure P15.9

Suppose we know that X(w) is bandlimited to + wc and that s(t) is an arbitraryperi­

odic function with period T.
(a) Draw a possible Fourier transform of s(t). Consider the case when S(W)I=o
is zero and the case when it is not zero.
(b) What is the range of T such that Y(o) will have regions equal to zero?
(c) For a typical value of T found in part (b), determine how to recover x(t) from
y(t).

P15.10
Consider the modulation system in Figure P15.10-1.
 Discrete-Time Modulation / Problems
P15-7

X(Q)

y[n]
x[n] X
x [Xn z[n]

cos cos4n 5 5
5 5
Figure P15.10-1

(a) Sketch Y(2).

(b) Sketch Z(Q).
(c) Suppose that we want to recover x[n] from z[n] using the system in Figure
P15.10-2.

z[n] a H(2) X x [n]

55

cos Oon
Figure P15.10-2

Determine two distinct combinations of H(Q) and go that will recover x[n] from
z[n].
MIT OpenCourseWare
[Link]

Resource: Signals and Systems

Professor Alan V. Oppenheim

The following may not correspond to a particular course on MIT OpenCourseWare, but has been
provided by the author as an individual learning resource.

For information about citing these materials or our Terms of Use, visit: [Link]
16 Sampling
Recommended
Problems
P16.1
The sequence x[n] = (-1)' is obtained by sampling the continuous-time sinusoidal
signal x(t) = cos oot at 1-ms intervals, i.e.,
cos(oonT) = (-1)", T = 10-' s
Determine three distinct possible values of wo.

P16.2
Consider the system in Figure P16.2.

cos wot ux xp(t)

p(t)= ( (t-nT), T= 3
n=-oo

Figure P16.2

(a) Sketch X,(w) for -97 : co 5 9x for the following values of wo.
(i) o = 7r
(ii) wo = 27
(iii) wo = 3w
(iv) wo = 57
(b) For which of the preceding values of wo is x,(t) identical?

P16.3
In the system in Figure P16.3, x(t) is sampled with a periodic impulse train, and a
reconstructed signal x,(t) is obtained from the samples by lowpass filtering.

T
xPX(t)
x (t) x rx(t)

T T

H(w)
p(t)=( 5(t - nT)
n=- F P
Figure P16.3

P16-1
Signals and Systems
P16-2

The sampling period T is 1 ms, and x(t) is a sinusoidal signal of the form x(t)
cos(27rfot + 0). For each of the following choices of fo and 0, determine x,(t).
(a) fo = 250 Hz, 0 = 7r/4
(b) fo = 750 Hz, 0 = -r/2
(c) fo = 500 Hz, 0 = r/2

P16.4

Figure P16.4 gives a system in which the sampling signal is an impulse train with
alternating sign. The Fourier transform of the input signal is as indicated in the
figure.

p(t)

x(t) 1 X xp (t) H(w) y (t)

p(t)

2A
--1

X(w)

WCM W

1--)

-37r -7r 7r 37r

i A

Figure P16.4
(a) For A < 7r/ 2 wm, sketch the Fourier transform of x,(t) and y(t).
(b) For A < r/ 2 M, determine a system that will recover x(t) from xp(t).
(c) For A < 7r/ 2wM, determine a system that will recover x(t) from y(t).
(d) What is the maximum value of A in relation to oM for which x(t) can be
recovered from either x,(t) or y(t).
 Sampling / Problems
P16-3

P16.5
Consider the system in Figure P16.5-1.

x(t) X t
xr (t)

2T
( 6(t -nT)
n=F-r P

Figure P16.5-1

Figures P16.5-2 and P16.5-3 contain several Fourier transforms of x(t) and x,(t).
For each input spectrum X(w) in Figure P16.5-2, identify the correct output spec­
trum X,(w) from Figure P16.5-3.

X(w)

-W

X(w)

-2W 2W
X(w)
(c)

-W w
X(W)

-2W -W W 2W

Figure P16.5-2
Signals and Systems
P16-4

Xr(Wj)

(i)

-W W
Xr(W)
(ii)

Ci
-W W

Xr(W)
(iii)

Co
-2W 2W

(iv) Xr(W)

-W w
Figure P16.5-3

P16.6
Suppose we sample a sinusoidal signal and then process the resultant impulse train,
as shown in Figure P16.6-1.

cos Wt Q(W)

p(t) = 6(t -nT)

n= ­

Figure P16.6-1
 Sampling / Problems
P16-5

The result of our processing is a value Q. As w changes, Q may change. Determine

which of the plots in Figures P16.6-2 and P16.6-3 are possible candidates for the
variation of Q as a function of w.

Q(M)

27n
T

Figure P16.6-2

Q ()

Figure P16.6-3

Optional
Problems
P16.7
The sampling theorem as we have derived it states that a signal x(t) must be sam­
pled at a rate greater than its bandwidth (or, equivalently, a rate greater than twice
its highest frequency). This implies that if x(t) has a spectrum as indicated in Figure
P16.7-1, then x(t) must be sampled at a rate greater than 2W2. Since the signal has
most of its energy concentrated in a narrow band, it seems reasonable to expect that
a sampling rate lower than twice the highest frequency could be used. A signal
whose energy is concentrated in a frequency band is often referred to as a bandpass
[Link] are a variety of techniques for sampling such signals, and these tech­
niques are generally referred to as bandpasssampling.
Signals and Systems
P16-6

X(Co)

-W2 1Cj 1 2

Figure P16.7-1

To examine the possibility of sampling a bandpass signal at a rate less than the
total bandwidth, consider the system shown in Figure P16.7-2. Assuming that wi >
(W 2 - wi), find the maximum value of T and the values of the constants A, W", and
Wb such that x,(t) = x(t).

p(t) = 1 6(t - nT)

.n =--o

XrM

p(t)

- Wb -Ca a jb

Figure P16.7-2

P16.8
In Problem P16.7 we considered one procedure for bandpass sampling and recon­
struction. Another procedure when x(t) is real consists of using complex modulation
followed by sampling. The sampling system is shown in Figure P16.8-1.
 Sampling / Problems
P16-7

x (t) iX x(t) HI X2( r iN x, (t)

elwot P (t)
Figure P16.8-1

With x(t) real and with X(w) nonzero only for wi < wI < W2 , the modulating fre­
quency wo is chosen as wo = 2(w1 + W2 ), and the lowpass filter H1 (w) has cutoff fre­
quency 1(c2 -2
(a) For X(w) as shown in Figure P 16.8-2, sketch X,(w).
(b) Determine the maximum sampling period T such that x(t) is recoverable from
xP(t).
(c) Determine a system to recover x(t) from x,(t).

X(W)

-W2 -Wi I
Figure P16.8-2
1 W2

P16.9
Given the system in Figure P16.9-1 and the Fourier transforms in Figure P16.9-2,
determine A and find the maximum value of T in terms of W such that y(t) = x(t)
if s(t) is the impulse train

s(t) = (t - nT)
n = -o

s(t)

x (t) yAt)

n (t)

Figure P16.9-1
Signals and Systems
P16-8

X(W) N(w)

-W W -2W -W W 2W

Figure P16.9-2

P16.10

Consider the system in Figure P16.10-1.

x(t) 10X 1 X1
xr(t)
-W R4

F ~W="-r
T
p(t)= E (t-nT)

Figure P16.10-1

Given the Fourier transform of x,(t) in Figure P16.10-2, sketch the Fourier trans­
form of two different signals x(t) that could have generated x,(t).

Xr(G)

1­

-W _5W
6
_W
6
W
6 5W
6
w

Figure P16.10-2

P16.11
Consider the system in Figure P16.11.
 Sampling / Problems
P16-9

H(w)

x(t) Xr (t)

p(t) (t-nT)
nl-oo

Figure P16.11

(a) If X(w) = 0 for I o > W, find the maximum value of T, We, and A such that x,(t)
= x(t).

(b) Let X 1 (w) = 0 for I > 2W and X 2(w) = 0 for Iw > W. Repeat part (a) for the
following.
(i) x(t) = x 1 (t) * x 2 (t)
(ii) x(t) = x 1 (t) + x 2(t)
(iii) x(t) = x 1(t)x 2(t)
(iv) x(t) = x 1 (lOt)
MIT OpenCourseWare
[Link]

Resource: Signals and Systems

Professor Alan V. Oppenheim

The following may not correspond to a particular course on MIT OpenCourseWare, but has been
provided by the author as an individual learning resource.

For information about citing these materials or our Terms of Use, visit: [Link]
17 Interpolation
Recommended
Problems
P17.1

Suppose we have the system in Figures P17.1-1 and P17.1-2, in which x(t) is sam­
pled with an impulse train. Sketch xP(t), y(t), and w(t).

xP (t) y~t
r (t) h (t) M.w(t)

p(t) 6(t - nT)

n=-F

Figure P17.1-1

h (t) x(t)

0 T -5T 10 5T

Figure P17.1-2

P17.2
Consider the signal x(t) = b(t - 1) + ib(t - 2), which we would like to interpolate
using the system given in Figure P17.2-1.

P17-1
Signals and Systems
P17-2

x(t) --- h(t) - y(t)

x(t)

1 -­

.5­ 1
1
t
2
t

Figure P17.2-1

For the following choices of h(t), sketch y(t).

(a) (b)

hl(t) h 2 (t)

t
t -1

Figure P17.2-3
Figure P17.2-2

(c)'h,(t) =sin rtr

irt

P17.3
Consider the system in Figure P17.3-1, with p(t) an impulse train with period T.

X(O)

x(t) xP(t)

-orm 0 Wm
p (t)

Figure P17.3-1
 Interpolation / Problems
P17-3

(a) Sketch P(w) and X,(w), assuming that no aliasing is present. What is the relation
between T and the highest frequency present in X(w) to guarantee that no alias­
ing occurs?
(b) Consider recovering x(t) from x,(t), assuming that no aliasing has occurred. For
example, assume that T = 21/4wm. We know that x(t) is recovered by interpo­
lating x,(t), as shown in Figure P17.3-2.

xp (t) h
h(t) ---- x(t)

Figure P17.3-2

Is the specification of h(t) unique so that x(t) can be exactly recovered from
x,(t)? Why not?
(c) Using the convolution integral, show that if the original sampling period was T
and if the filter is an ideal lowpass filter with cutoff we, then the recovered signal
X(t) is

Xr(t) x(nT) sinc W,(t-nT)

P17.4
In the system in Figure P17.4, p(t) is an impulse train with period A, and the impulse
response g(t) is as indicated. Determine H(w) so that y(t) = x(t), assuming that no
aliasing has occurred.

x(t) y (t)

p(t)

g(t)

Figure P17.4
Signals and Systems
P17-4

P17.5

Figure P17.5-1 shows the overall system for filtering a continuous-time signal using
a discrete-time filter. If Xe(w) and H1 (Q) are as shown in Figure P17.5-2, with 1/T =
20 kHz, sketch X,(x), X(Q), Y(Q), and Y(w).

xc (t) xp t) Conversion x[n] = xc(nT) h [] y[n] =Vc (nT) Conversion Yp(t) T yct)
X to a
sequence
1
H, (1
- - - --. to an-...
impulse train
ir/T
F
-T/T

pMt = 1 5(t - nT)

Figure P17.5-1 HCo

XC (W) H ( )

-IrX10 4 7rX104 ­

Figure P17.5-2

Optional
Problems
P17.6

We consider a sequence x[n] to which discrete-time sampling, as illustrated in Fig­

ure 8.32 of the text, has been applied. We assume that the conditions of the discrete-
time sampling theorem are satisfied; that is, Qs > 2 0,, where Qs is the sampling fre­
quency and X(Q) = 0, U < I I - r. The original sequence x[n] is then recoverable
from x,[n] by ideal lowpass filtering, which, as discussed in Section 8.6 of the text,
corresponds to bandlimited interpolation.
The ZOH represents an approximate interpolation whereby each sample value
is repeated (or held) N - 1 successive times, as illustrated in Figure P17.6-1 for
N = 3. The FOH represents a linear interpolation between samples, as illustrated in
Figure P17.6-1.
 Interpolation / Problems
P17-5

x[n]

XP [n]

n
xO[n]

ZOH

- FOH

Figure P17.6-1

(a) The ZOH can be represented as an interpolation in the form of eq. (8.51) of the
text (page 545) and the system in Figure P17.6-2. Determine and sketch ho[n]
for the general case of a sampling period N.

ZOH

xP [n] ----- ho [n] xO [n]

Figure P17.6-2

(b) x[n] can be exactly recovered from the ZOH sequence xo[n] using an appropriate
LTI filter H(G), as indicated in Figure P17.6-3. Determine H(Q).

xo [n] x[nI

Figure P17.6-3
Signals and Systems
P17-6

(c) The FOH (linear interpolation) can be represented as an interpolation in the

form of eq. (8.51) of the text and, equivalently, the system in Figure P17.6-4.
Determine and sketch hi[n] for the general case of a sampling period N.

FOH

xP [n] h1 [n] -- *-x1 [n]

Figure P17.6-4

(d) x[n] can be exactly recovered from the FOH sequence x 1 [n] using an appropriate
LTI filter with frequency response H(Q), as illustrated in Figure P17.6-5. Deter­
mine H(Q).

xI [n]- H(2) x[n]

Figure P17.6-5

P17.7

Figure P17.7 shows a system consisting of a continuous-time linear time-invariant

system followed by a sampler, conversion to a sequence, and a discrete-time linear
time-invariant system. The continuous-time LTI system is causal and satisfies the
LCCDE:
dyc(t) + y(t = X(0
dt

Ye (t Conversion of y [n]
x (t) LTI X impulse train w[n]
sequence

p(t) = X 6(t - nT) y[n] =yc(nT)

n =­

Figure P17.7

The input xc(t) is a unit impulse 6(t).

(a) Determine yc(t).
(b) Determine the frequency response H(Q) and the impulse h[n] such that w[n] =
b[n].
 Interpolation / Problems
P17-7

P17.8
Consider a signal x(t) that is nonzero only in an interval [-T, T]. This problem deals
with the sampling of the Fourier transform of x(t).
(a) Suppose that we consider sampling the Fourier transform with an impulse train

P(w) = ( - nos),
n= -OO

as shown in Figure P17.8-1.

X(Co) fXP(O)
P(W)

Figure P17.8-1

Sketch X,(w) if X(w) is as given in Figure P17.8-2.

X(W)

Figure P17.8-2

(b) Determine an expression for x,(t), the inverse Fourier transform of X,(w). How
does x,(t) relate to x(t)?
(c) Determine a relation between w, and T such that x(t) is recoverable.
(d) Assuming that w, satisfies the condition in part (c), how is x(t) recovered from
x,( t)?
MIT OpenCourseWare
[Link]

Resource: Signals and Systems

Professor Alan V. Oppenheim

The following may not correspond to a particular course on MIT OpenCourseWare, but has been
provided by the author as an individual learning resource.

For information about citing these materials or our Terms of Use, visit: [Link]
18 Discrete-Time Processing of
Continuous-Time Signals
Recommended
Problems

P18.1
Consider the system in Figure P18.1-1 for discrete-time processing of a continuous-
time signal using sampling period T, where the C/D operation is as shown in Figure
P18.1-2 and the D/C operation is as shown in Figure P18.1-3.

Xc(t) C/D xG(9) ----- D/C yPc (t)

Figure P18.1-1

H(w)
Discrete-time sequence T
y [n] to impulse train with - (t)
spacing T Co 1
T T

Figure P18.1-3

P18-1
Signals and Systems
P18-2

The filter G(Q) is the lowpass filter shown in Figure P18.1-4.

G(Gi)

IT iT

Figure P18.1-4

The Fourier transform of xc(t), Xc(w) is given in Figure P18.1-5.

Xc(w)

-2nTX 4kHz 27X 4kHz

Figure P18.1-5

The sampling frequency is 8 kHz. Sketch accurately the following transforms.

(a) X,(w)
(b) X(Q)
(c) Y(Q)
(d) Ye(w)

P18.2
Consider the continuous-time frequency response in Figure P18.2.

H(o)

-500T 500n7r

Figure P18.2

We want to implement this continuous-time filter using discrete-time processing.

(a) What is the maximum value of the sampling period T required?
 Discrete-Time Processing of Continuous-Time Signals / Problems
P18-3

(b) What is the required discrete-time filter G(Q) for T found in part (a)?
(c) Sketch the total system.

P18.3
The system in Figure P18.3 is similar to that demonstrated in the lecture. Note that,
as in the lecture, there is no anti-aliasing filter.

xc (t) C|D x[]YLj D|C Noy(t)

20000 20000

Figure P18.3

For the following signals, draw Xc(w), X(Q), and Yc(w).

(a) xc(t) = cos(2r - 5000t)
(b) xc(t) = cos(2xr 27000t)
(c) xc(t) = cos(2r - 17000t)

P18.4
Suppose we want to design a variable-bandwidth, continuous-time filter using the
structure in Figure P18.4-1.

Find, in terms of wc, the value of the sampling period To and the corresponding value
co, such that the total continuous-time filter has the frequency response shown in
Figure P18.4-2.

Figure P18.4-2
Signals and Systems
P18-4

P18.5
Consider the system in Figure P18.5-1.

x(t) C/D H(Q) -- ]

y-[nw

T = To

Figure P18.5-1

Let H(Q) be as given in Figure P18.5-2 and X(co) as given in Figure P18.5-3.

H(Q)
1

IT it
3 -3

Figure P18.5-2

X(W)

7T 7T
To To
Figure P18.5-3

(a) Sketch X(Q) and Y(Q).

(b) Suppose we replace the system in Figure P18.5-1 by the system in Figure
P18.5-4. Find G(w) such that y[n] = z[n].

x(t) G(w) C/D P z [n]

T = To

Figure P18.5-4
 Discrete-Time Processing of Continuous-Time Signals / Problems
P18-5

Optional
Problems

P18.6

Suppose we are given the system in Figure P18.6-1.

yyIn] Convert K
x In]
x / to y(t)
-1007r 1007r T= To train -100ir 100T
T = To

Figure P18.6-1

(a) Find the appropriate values of the sampling period To to avoid aliasing. Also
find the proper value for K so that the overall system has a gain of unity at
w = 0 (i.e., no overall dc gain).
(b) Suppose To is halved, but the anti-aliasing and reconstruction filters are not
modified.
(i) If X(w) is as given in Figure P18.6-2, find Y(Q).

X(W)

-1007r 100ff

Figure P18.6-2

(ii) If Y(Q) is as given in Figure P18.6-3, find Y(w).

Y(92)

1­

IT I
3 3

Figure P18.6-3
Signals and Systems
P18-6

P18.7
Figure P18.7 shows a system that processes continuous-time signals using a digital
filter. The digital filter h[n] is linear and causal with difference equation
y[n] = -y[n - 1]+x[n]
For input signals that are bandlimited so that Xe(w) = 0 for I l > ir/T, the system
is equivalent to a continuous-time LTI system. Determine the frequency response
He(w) of the equivalent overall system with input xc(t) and output yc(t).

Conversion of a y I(t) l owpass

x,( Conversion of x[]yn
_n] y t) fequtenr / yt)
xc t) X impulse train x-n h sequence

sequence impulse train and gain

p(t) = I (t - nT)

Figure P18.7

P18.8
Figure P18.8-1 depicts a system for which the input and output are discrete-time
signals. The discrete-time input x[n] is converted to a continuous-time impulse train
x,(t). The continuous-time signal x,(t) is then filtered by an LTI system to produce
the output yc(t), which is then converted to the discrete-time signal y[n]. The LTI
system with input xc(t) and output yc(t) is causal and is characterized by the linear
constant-coefficient difference equation

+4 C + 3yc(t) = x(t)
dt2 dt

I 5(t-nT)

xP(t = x[n] 6It - nT)

yp(t) = y,(t) 6(t - nT)

y[n] = y,(nT)

Figure P18.8-1
 Discrete-Time Processing of Continuous-Time Signals / Problems
P18-7

The overall system is equivalent to a causal discrete-time LTI system, as indicated

in Figure P18.8-2. Determine the frequency response H(Q) of the equivalent LTI
system.

h[n]; H(92)
x[n] 1 equivalent P y[n]
LTI system

Figure P18.8-2

P18.9
We wish to design a continuous-time sinusoidal signal generator that is capable of
producing sinusoidal signals at any frequency satisfying wi : W 5 W2, where w, and
W2 are positive numbers.
Our design is to take the following form. We have stored a discrete-time cosine
wave of period N; that is, we have stored x[O], . . . , x[N - 1], where

x[k] = cos (271.k

N)

Every T seconds we output an impulse weighted by a value of x[k], where we pro­

ceed through the values of k = 0, 1, ... , N - 1 in a cyclic fashion. That is,

y,(t) = ( x[k modulo N] S(t - kT)

k= ­

= cos N ­
(ttkT )

(a) Show that by adjusting T we can adjust the frequency of the cosine signal being
sampled. Specifically, show that

y,(t) = (cos wat) ('btt - kT),

k= -o

where wo = 21r/NT. Determine a range of values for T so that y,(t) can represent
samples of a cosine signal with a frequency that is variable over the full range

(b) Sketch Y,(w).

The overall system for generating a continuous-time sinusoid is depicted in Figure
P18.9-1. H(w) is an ideal lowpass filter with unity gain in its passband:

H(w) = othwis
0, otherwise
Signals and Systems
P18-8

x[O]

e T yp(t) H(w) y(t)

x[N -1] D

Figure P18.9-1

The parameter we is to be determined such that y(t) is a continuous-time cosine

signal in the desired frequency band.
(c) Consider any value of T in the range determined in part (a). Determine the min­
imum value of N and some value for w,such that y(t) is a cosine signal in the
range wi : O o02.
(d) The amplitude of y(t) will vary depending on the value of w chosen between wi
and W2. Determine the amplitude of y(t) as a function of w and as a function of
N.

P18.10
In many practical situations, a signal is recorded in the presence of an echo, which
we would like to remove by appropriate processing. For example, Figure P18.10-1
illustrates a system in which a receiver receives simultaneously a signal x(t) and
an echo represented by an attenuated delayed replication of x(t). Thus, the receiver
output is s(t) = x(t) + ax(t - TO), where ja l < 1. The receiver output is to be
processed to recover x(t) by first converting to a sequence and using an appropriate
digital filter h[n] as indicated in Figure P18.10-2.

/( X
0at - TO)

x(t)

Receiver output
s(t) = x(t) + a x(t - TO)

Figure P18.10-1
 Discrete-Time Processing of Continuous-Time Signals / Problems
P18-9

Ideal lowpass
filter

se(tM x(t) + ax(t - TO) H

SP(t) Conversion of s[n] y[n] Conversion of y,(t) A Y, t)

X impulseatrain - ---0 h[n] sequence ---- 0

sequence impulse train 7

T T

p(t) = I (t - kT)
k =­

Figure P18.10-2

Assume that x(t) is bandlimited, i.e., X(w) = 0 for IwI > wm, and that al < 1.
(a) If To < lr/M and the sampling period is taken equal to To (i.e., T = TO), deter­
mine the difference equation for the digital filter h[n] so that yc(t) is propor­
tional to x(t).
(b) With the assumptions of part (a), specify the gain A of the ideal lowpass filter
so that yc(t) = x(t).
2
(c) Now suppose that 2r/wm < To < 7r/WM. Determine a choice for the sampling
period T, the lowpass filter gain A, and the frequency response for the digital
filter h[n] such that yc(t) is equal to x(t).
MIT OpenCourseWare
[Link]

Resource: Signals and Systems

Professor Alan V. Oppenheim

The following may not correspond to a particular course on MIT OpenCourseWare, but has been
provided by the author as an individual learning resource.

For information about citing these materials or our Terms of Use, visit: [Link]
19 Discrete-Time Sampling
Recommended
Problems
P19.1

Consider Figures P19.1-1 and P19.1-2, and determine X(Q), P(Q), x,[n), and X,(Q).

[ n] xr xP [n]

p [nJ

Figure P19.1-1

x [n]

-3 -2 -1 0 1 2 3
Figure P19.1-2

P19.2
x[n] has a transform X(Q). Determine in terms of X(Q) the transforms of the signals
in parts (a) and (b).

x[n/2]1, n even,
(a)
(a)
x[n]=0,) n odd
(b) xd[n] = x[2n], i.e., xd[n] is x[n] decimated.

P19-1
Signals and Systems
P19-2

(c) If X(Q) is as given in Figure P19.2, sketch X,(Q) and Xd(Q) for parts (a) and (b).

X( 2)

Figure P19.2

P19.3
Consider the system in Figure P19.3-1.

[fn] x.0 xp[n]

p [n]

Figure P19.3-1

(a) If p[n] is given by Figure P19.3-2 sketch P(Q) for N =1, 2, and L, an arbitrary
integer.

p [n]

. . -N
...
0
... 1
N
..
Figure P19.3-2

(b) For each of the discrete-time spectra in Figures P19.3-3 and P19.3-4, determine
the maximum sampling period N such that x[n] is reconstructible from its sam­
ples x,[n] using an ideal lowpass filter.
 Discrete-Time Sampling / Problems
P19-3

In each case, specify the associated cutoff frequencies for the lowpass filter.

X(E2)

37T
10

Figure P19.3-3

(ii)
X(C )

j 27T

5 5

Figure P19.3-4

P19.4
Suppose the signal x(t) is processed as shown in Figure P19.4-1.

x(t) rx~t) C/D

C/D1:3 Decimation b y[n] =x[3n]

T, sampling period = 1 ms
Figure P19.4-1

(a) The system in Figure P19.4-1 can be replaced by the one in Figure P19.4-2.
Find T1 .

x(t) C/D 7 y[n]

Sampling period = T,

Figure P19.4-2
Signals and Systems
P19-4

(b) Let X(w) be given as in Figure P19.4-3. Find X(Q) and Y(Q).

IX(<)I

iT
2T 2T

L Figure P19.4-3

P19.5
As discussed in Section 8.7 and illustrated in Figure 8.40 of the text as well as in
Figure P19.5-1 below, the procedure for interpolation or upsampling by an integer
factor N can be thought of as the cascade of two operations. The first system, system
A, corresponds to inserting (N - 1) zero sequence values between each sequence
value of x[n], so that

= 0, +N, i2N, . . .
x,,[n] =<
x
N-n], n
0o, otherwise

System A

Conversion of x, [n Lowpass filter

Xd [n]----O decimated sequence L fite x [n
to sampled sequence

Figure P19.5-1
For exact bandlimited interpolation, H(Q) is an ideal lowpass filter.
(a) Determine whether system A is linear.
(b) Determine whether system A is time-invariant.
(c) For Xd(2) as sketched in Figure P19.5-2, with N = 3, sketch X(Q).
(d) For N = 3, Xd(Q) as in Figure P19.5-2, and H(Q) appropriately chosen for exact
bandlimited interpolation, sketch X(Q).

Xd (E2)

Figure P19.5-2
 Discrete-Time Sampling / Problems
P19-5

P19.6
Consider the sampling systems in Figure P19.6-1.

x(t) b. X P = x tp (t)
: x(t)

p(t)= , 6(t - kTO)

Y(w) X 1Y,(w)= Y(o)Q(w)

Q~) ~o- o
k=-oo

Figure P19.6-1

Let x(t) and X(w) be given as in Figure P19.6-2.

x(t)

X(w)

2To0 TO

Figure P19.6-2
Signals and Systems
P19-6

Let y(t) and Y(w) be given as in Figure P19.6-3.

y (t)
I
21T

0001,

r_
7ri
2 w0o
2w0

Y(co)

Co

Figure P19.6-3
(a) Draw x,(t) and Y,(w).
(b) Find X,(w) and y,(t).
(c) Is y,(t) periodic? Does Y,(w) reflect this property?

Optional
Problems
P19.7

Consider a discrete-time sequence x[n] from which we form two new sequences,
x,[n] and xd[n], where x,[n] corresponds to sampling x[n] with sampling period 2
and xd[n] corresponds to decimatingx[n] by a factor of 2, so that
n = 0, 2, 4,
x,[n] =
10, n = 1, ±3,..,
and
Xd[n] = x[2n]
(a) If x[n] is as illustrated in Figure P19.7-1, sketch the sequences x,[n] and xd[n].

0 P.

Figure P19.7-1
 Discrete-Time Sampling / Problems
P19-7

(b) If X(Q) is as shown in Figure P19.7-2, sketch X,(Q) and Xd(Q).

X(R)
1

31r 31r
4 4
Figure P19.7-2

P19.8
Consider the system in Figure P19.8-1, where X(Q) is as shown in Figure P19.8-2.

X(Q)

3-n 3n7

Figure P19.8-2

There is a range of values for N such that, with an appropriate choice for H(Q), y[n]
will equal x[n]. For each allowable positive integer value of N,
(a) Draw X,(Q).
(b) Find an appropriate H(Q) such that y[n] = x[n].

P19.9

Consider the system with input x[n] and output y[n] related by
x[3n] + x[3n + 1] + x[3n + 2]
y~n] =
Signals and Systems
P19-8

(a) For the sequence x[n] in Figure P19.9, sketch y[n].

x [un

-5 -4 -3 -2 -1 0 1 2 3 4 5 6

-214

Figure P19.9

(b) Express the system as a combination of filtering and decimation.

P19.10
Consider the system in Figure P19.10, where

x0[n] =
n,
x4 N) n=kN
0, n # kN, k an integer
Find a constraint on h[n] such that y[kN] = x[k], for all k.

x[n] padding y[n]

Figure P19.10
MIT OpenCourseWare
[Link]

Resource: Signals and Systems

Professor Alan V. Oppenheim

The following may not correspond to a particular course on MIT OpenCourseWare, but has been
provided by the author as an individual learning resource.

For information about citing these materials or our Terms of Use, visit: [Link]
20 The Laplace Transform
Recommended
Problems
P20.1
Consider the signal x(t) = 3e 2'u(t) + 4e 3 'u(t).
(a) Does the Fourier transform of this signal converge?
(b) For which of the following values of a does the Fourier transform of x(t)e -"
converge?
(i) a= 1
(ii) a = 2.5
(iii) a = 3.5
(c) Determine the Laplace transform X(s) of x(t). Sketch the location of the poles
and zeros of X(s) and the ROC.

P20.2
Determine the Laplace transform, pole and zero locations, and associated ROC for
each of the following time functions.
(a) e -"u(t), a > 0
t
(b) e ~au(t), a< 0
t
(c) -e -a u(- t), a< 0

P20.3
Shown in Figures P20.3-1 to P20.3-4 are four pole-zero plots. For each statement in
Table P20.3 about the associated time function x(t), fill in the table with the cor­
responding constraint on the ROC.
(a) (b)

Im

s plane

O - Re
-2 2

Figure P20.3-2

P20-1
Signals and Systems
P20-2

(c) (d)

Im

s plane

X O Re
-2 2

Figure P20.3-3

Constraint on ROC for Pole-Zero Pattern

x(t) (a) (b) (d)
(i) Fourier
transform
of x(t)e -'
converges

(ii) x(t) = 0,
t > 10

(iii)x(t) = 0,
t < 0

Table P20.3

P20.4
Determine x(t) for the following conditions if X(s) is given by
1
X(s) = _____
(s + 1)(s + 2)
(a) x(t) is right-sided
(b) x(t) is left-sided
(c) x(t) is two-sided

P20.5

An LTI system has an impulse response h(t) for which the Laplace transform H(s)
is

H(s) = h(t)e -dt =s 1' Re{s} > -1

Determine the system output y(t) for all t if the input x(t) is given by
x(t) = e t ' 2 + 2e -t 3
for all t.
 The Laplace Transform / Problems
P20-3

P20.6
(a) From the expression for the Laplace transform of x(t), derive the fact that the
Laplace transform of x(t) is the Fourier transform of x(t) weighted by an
exponential.
(b) Derive the expression for the inverse Laplace transform using the Fourier
transform synthesis equation.

Optional
Problems
P20.7
Determine the time function x(t) for each Laplace transform X(s).
1
(a) , Re'{s}> -1
s+ 1
1
(b) , R {s) < -1
s +1

(c) Refs) > 0

Rs
s2 + 4

(d) s +1 Refs) > -2

s2 + 5s + 6'
()2s + 1
Refs) < -3
s2 + 5s + 6'
(f) - s + 1
o < Refs) < 1
s 2(s - 1)
s- s + 1
(g) 2 S+1)2 , -1 < Refs)
(S +

s + 1
(h) () + 1 Refs) > - 1
(s ± 1)2 + 4'
Hint: Use the result from part (c).

P20.8
The Laplace transform X(s) of a signal x(t) has four poles and an unknown number
of zeros. x(t) is known to have an impulse at t = 0. Determine what information, if
any, this provides about the number of zeros.

P20.9

Determine the Laplace transform, pole-zero location, and associated ROC for each
of the following time functions.
(a) e a< 0
atu(t),

(b) -e"'u(-t), a > 0

(c) e atu(t), > 0
(d) e *t. a > 0
Signals and Systems
P20-4

(e) u(t)
(f) b(t - to)

(g) L
k=o
ako(t - kT), a > 0

(h) cos (wot + b)u(t)

(i) sin (oot + b)e --a t u(t), a> 0

P20.10
(a) If x(t) is an even time function such that x(t) = x(- t), show that this requires
that X(s) = X(-s).
(b) If x(t) is an odd time function such that x(t) = -x(--t), show that X(s) =
-X(-s).
(c) Determine which, if any, of the pole-zero plots in Figures P20.10-1 to P20.10-4
could correspond to an even time function. For those that could, indicate the
required ROC.
(i) (ii)

.x (Re C x (Re
-1 1 -1 1

Figure P20.10-1 Figure P20.10-2

(iii) (iv)

x 0 Re
1e -1 1

Figure P20.10-3 Figure P20.10-4

(d) Determine which, if any, of the pole-zero plots in part (c) could correspond to
an odd time function. For those that could, indicate the required ROC.
MIT OpenCourseWare
[Link]

Resource: Signals and Systems

Professor Alan V. Oppenheim

The following may not correspond to a particular course on MIT OpenCourseWare, but has been
provided by the author as an individual learning resource.

For information about citing these materials or our Terms of Use, visit: [Link]
21 Continuous-Time
Second-Order Systems
Recommended
Problems
P21.1

Consider the following four impulse responses:

(i) hi(t) = e-"'u(t), a > 0,
(ii) h 2(t) =-e t
-a u(- a > 0,
t),
(iii) t
h3 (t) = e -a u(t), a < 0,
t
(iv) h 4(t) = -e -a u(-t), a < 0
(a) Verify that
1
H 2 (s) = s+a'
with an ROC corresponding to Re{s} < - a.
(b) It is easily verified that in all four cases the system functions (i.e., the Laplace
transforms of the impulse responses) have the same algebraic expression. How­
ever, the ROC is different. For each of the four systems, identify which of the s
plane diagrams in Figures P21.1-1 to P21.1-6 correctly represents the associated
Laplace transform, including the ROC. Also indicate which diagrams corre­
spond to stable systems.
(A) (B)

Im Im
s plane s plane

Re Re

Figure P21.1-1 Figure P21.1-2

(C) (D)

Figure P21.1-3 Figure P21.1-4

P21-1
Signals and Systems
P21-2

(E) (F)

s plane s plane

Re

Figure P21.1-5 Figure P21.1-6

P21.2

Consider the LTI system with input x(t) = e u(t) and impulse response h(t) =
e-2t
u(t).
(a) Determine X(s) and H(s).
(b) Using the convolution property of the Laplace transform, determine Y(s), the
Laplace transform of the output, y(t).
(c) From your answer to part (b), find y(t).

P21.3

As indicated in Section 9.5 of the text, many of the properties of the Laplace trans­
form and their derivation are analogous to corresponding properties of the Fourier
transform developed in Chapter 4 of the text. In this problem you are asked to out­
line the derivation for some of the Laplace transform properties in Section 9.5 of
the text.
By paralleling the derivation for the corresponding property for the Fourier
transform in Chapter 4 of the text, derive each of the following Laplace transform
properties. Your derivation must include a consideration of the ROC.
(a) Time-shifting property (Section 9.5.2)
(b) Convolution property (Section 9.5.5)

P21.4

Consider a continuous-time LTI system for which the input x(t) and output y(t) are
related by the differential equation
d 2 y(t) dy(t) _ 2y(t) = x(t)
2 dt
dt
Let X(s) and Y(s) denote the Laplace transforms of x(t) and y(t), and let H(s)
denote the Laplace transform of the impulse response h(t) of the preceding system.
(a) Determine H(s). Sketch the pole-zero plot.
 Continuous-Time Second-Order Systems / Problems
P21-3

(b) Sketch the ROC for each of the following cases:

(i) The system is stable.
(ii) The system is causal.
(iii) The system is neither stable nor causal.
(c) Determine h(t) when the system is causal.

P21.5
Consider the following system function H(s) and its corresponding pole-zero plot in
Figure P21.5.

H(s) = ~
s +1

Im
s plane

Re

Figure P21.5

Using the graphical method discussed in the lecture, find IH(O) |, 4 H(O), IH(jl) 1,
4H(jl), |H(joo) , and 4 H(joo). Sketch the functions IH(jw) I and 4H(jw).

P21.6
For the following system function, plot the pole-zero diagram and graphically deter­
mine the approximate locations of the peaks and valleys of IH(jW) 1:

H(s) =(s
+ 5)s
[s - (-0.1 + j2)][s - (-0.1 - j2)]

Optional
Problems
P21.7
(a) Draw the block diagram for the following second-order system in terms of inte­
grators, coefficient multipliers, and adders.
2

H(s) = ~2 +2Wn + W2
Signals and Systems
P21-4

(b) Sketch the pole-zero plot of H(s) and plot IH(jw)| under the following con­
ditions.
(i) w. is kept constant, but r is varied from close to 0 to close to 1.
(ii) r is kept constant, but w, is varied from about 0 to infinity.
You don't have to be precise but show how the bandwidth and location of the
peak changes for the two cases above.

P21.8
(a) Consider the following system function H(s).

H(s) = +
s2 +s+1 s 2 +2s-+2
Draw the block diagram for H(s) implemented as
(i) a parallel combination of second-order systems,
(ii) a cascade combination of second-order systems.
(b) Is implementation (ii) unique?

P21.9
Complete the Laplace transform pairs given in Table P21.9. You may use the follow­
ing transform of e -au(t) as well as general properties of Laplace transforms to
derive your answer. You may also use earlier entries in the table to help you derive
later results.

e -"u(t) .a 1
-, ROC: Re{s} > Re{a}
s + a'

x(t) X(s) ROC

(a) sin(cot)u(t)

(b) e -2'sin(cot)u(t)

(c) te- 2'u(t)

(d) + Re{s}> - 2
(s + 2)(s + 3)

1 - e-2
(e) - Re{s}> -3
s + 3
Table P21.9
 Continuous-Time Second-Order Systems / Problems
P21-5

P.21.10
Figures P21.10-1, P21.10-2, and P21.10-3 contain impulse responses h(t), frequency
responses H(jw), and pole-zero plots, respectively.
(a) For each h(t), find the best matching IH(jw) |.
(b) For each |H(jw)|, find the best matching pole-zero plot of H(s).
Consider entries with references to other plots, such as (3) and (5) or (b) and (h),
as a pair.

Compare Compare
with (2) with (1)

Compare
Compare
with (5) with (7)

t .h2 t

Compare area = K
with (3)

area = K

V K Compare
with (4)

Figure P21.10-1
Signals and Systems
P21-6

Compare
with (d)

".."0/1

(c) (d)
Compare
with (a)

(e)

(g) Compare
with (b)
Compare
wit (fW
J1'***o co)

Figure P21.10-2
 Continuous-Time Second-Order Systems / Problems
P21-7

Im

double
pole

(iii) (iv) Im

double
pole --. j,

(vi)
double 'j
zero
double
zero

(vii) (Viii)

Figure P21.10-3
MIT OpenCourseWare
[Link]

Resource: Signals and Systems

Professor Alan V. Oppenheim

The following may not correspond to a particular course on MIT OpenCourseWare, but has been
provided by the author as an individual learning resource.

For information about citing these materials or our Terms of Use, visit: [Link]
22 The z-Transform
Recommended
Problems
P22.1
An LTI system has an impulse response h[n] for which the z-transform is

H(z) = h[n]z- z-1 , |

(a) Plot the pole-zero pattern for H(z).
(b) Using the fact that signals of the form z" are eigenfunctions of LTI systems,
determine the system output for all n if the input x[n] is
x[n] = (?)" + 3(2)n

P22.2
Consider the sequence x[n] = 2"u[n].
(a) Is x[n] absolutely summable?
(b) Does the Fourier transform of x[n] converge?
(c) For what range of values of r does the Fourier transform of the sequence
r-"x[n] converge?
(d) Determine the z-transform X(z) of x[n], including a specification of the ROC.
(e) X(z) for z = 3e'a can be thought of as the Fourier transform of a sequence x,[n],
i.e.,
2"u[n] X(z),
xi[n] 5 X(3ei") = XI(ei")
Determine x1 [n].

P22.3
Shown in Figure P22.3 is the pole-zero plot for the z-transform X(z) of a sequence
x[nI.

Im
z plane

X X XRe
1 2 2
3 3

Figure P22.3
P22-1
Signals and Systems
P22-2

Determine what can be inferred about the associated region of convergence from
each of the following statements.
(a) x[n] is right-sided.
(b) The Fourier transform of x[n] converges.
(c) The Fourier transform of x[n] does not converge.
(d) x[n] is left-sided.

P22.4
(a) Determine the z-transforms of the following two signals. Note that the z-trans­
forms for both have the same algebraic expression and differ only in the ROC.
(i) xi[n] = (i)"u[n]
(ii) X2[n] = -(I)"n[-U - 1]
(b) Sketch the pole-zero plot and ROC for each signal in part (a).
(c) Repeat parts (a) and (b) for the following two signals:
(i) x 3[n] = 2u[n]
(ii) x 4 [n] = -(2)"u[-n - 1]
(d) For which of the four signals xi[n], x 2[n], xs[n], and x 4[n] in parts (a) and (c)
does the Fourier transform converge?

P22.5

Consider the pole-zero plot of H(z) given in Figure P22.5, where H(a/2) = 1.

z plane

Figure P22.5
(a) Sketch |H(e'")I as the number of zeros at z = 0 increases from 1 to 5.
(b) Does the number of zeros affect <'H(ej")? If so, specifically in what way?
(c) Find the region of the z plane where IH(z) I = 1.
 The z-Transform / Problems
P22-3

P22.6
Determine the z-transform (including the ROC) of the following sequences. Also
sketch the pole-zero plots and indicate the ROC on your sketch.
(a) (I)"U [n]
(b) 6[n + 1]

P22.7
For each of the following z-transforms determine the inverse z-transform.
1 1
(a) X(z) = 1± zIZ Iz| >
1 -z 1z 1
(b) X(z) = 1 z-2 , Iz| >

1 -az- 1
(c) X(z)= z _-' IzI > iaa

Optional
Problems
P22.8
In this problem we study the relation between the z-transform, the Fourier trans­
form, and the ROC.
(a) Consider the signal x[n] = u[n]. For which values of r does r-"x[n] have a con­
verging Fourier transform?
(b) In the lecture, we discussed the relation between X(z) and I{r-"x[nj}. For each
of the following values of r, sketch where in the z plane X(z) equals the Fourier
transform of r-x[n].
(i) r=1
(ii) r=
(iii) r= 3
(c) From your observations in parts (a) and (b), sketch the ROC of the z-transform
of u[n].

P22.9
(a) Suppose X(z) on the circle z = 2e'j is given by
1
X(2ej') = 1 _

Using the relation X(reju) = 5{r~"x[n]}, find 2-x[n] and then x[n], the inverse
z-transform of X(z).
(b) Find x[n] from X(z) below using partial fraction expansion, where x[n] is
known to be causal, i.e., x[n] = 0 for n < 0.
3 + 2z~1
X(z) -2 + 3z-' + z- 2
Signals and Systems
P22-4

P22.10
A discrete-time system with the pole-zero pattern shown in Figure P22.10-1 is
referred to as a first-order all-pass system because the magnitude of the frequency
response is a constant, independent of frequency.

Urn
Unit circle

I ROC: I z I > a

Figure P22.10-1

(a) Demonstrate algebraically that IH(eu) I is constant.

(b) To demonstrate the same property geometrically, consider the vector diagram
in Figure P22.10-2. Show that the length of v 2 is proportional to the length of vi
independent of 0 by following these two steps:
(i) Express the length of v, using the law of cosines and the fact that it is
one leg of a triangle for which the other two legs are the unit vector and
a vector of length a.
(ii) In a manner similar to that in step (i), determine the length of v 2 and show
that it is proportional in length to vi independent of Q.

9JM

Unit circle

Figure P22.10-2

P22.11

Parts (a)-(e) (Figures P22.11-1 to P22.11-5) give pole-zero plots, and parts (i)-(iv)
(Figures P22.11-6 to P22.11-9) give sketches of possible Fourier transform magni­
tudes. Assume that for all the pole-zero plots, the ROC includes the unit circle. For
each pole-zero plot (a)-(e), specify which one if any of the sketches (i)-(iv) could
represent the associated Fourier transform magnitude. More than one pole-zero plot
may be associated with the same sketch.
 The z-Transform / Problems
P22-5

(a) (b)

Im z plane

F PRer

Figure P22.11-1

(c) (d)

(e)

(ii)

7T

Figure P22.11-7
Signals and Systems
P22-6

(iii) (iv)

iTV

ir
Figure P22.11-8 Figure P22.11-9

P22.12
Determine the z-transform for the following sequences. Express all sums in closed
form. Sketch the pole-zero plot and indicate the ROC. Indicate whether the Fourier
transform of the sequence exists.
(a) (i)"{u[n] - u[n - 10]}
(b) (i)II

(c) 7 cos [ + r u[n]

0, n < 0
(d) x[n]= 1, 0 s n - 9
0, 9< n

P22.13

Using the power-series expansion

log(1 - W) = - Wi , w l< 1,

determine the inverse of the following z-transforms.

(a) X(z) = log(1 - 2z), IzI < i
(b) X(z) = log(1 - z- 1), |z > i
MIT OpenCourseWare
[Link]

Resource: Signals and Systems

Professor Alan V. Oppenheim

The following may not correspond to a particular course on MIT OpenCourseWare, but has been
provided by the author as an individual learning resource.

For information about citing these materials or our Terms of Use, visit: [Link]
23 Mapping Continuous-Time Filters
to Discrete-Time Filters
Recommended
Problems
P23.1
For each of the following sequences, determine the associated z-transform, includ­
ing the ROC. Make use of properties of the z-transform wherever possible.
(a) x1[n] = (I)"u[n]
(b) X2[n] = (-3)"u[-n]
(c) xsn] = xi[n] + x2[n]
(d) X4[n] = xl[n - 5]
(e) x5[n] = xi[n + 51
(f) eX[n] = (I)"u[n]
(g) x 7[n] = xi[n] * x 6 [n]

P23.2
A causal LTI system is described by the difference equation
y[n] - 3y[n - 1] + 2y[n - 2] = x[n]
(a) Find H(z) = Y(z)/X(z). Plot the poles and zeros and indicate the ROC.
(b) Find the unit sample response. Is the system stable? Justify your answer.
(c) Find y[n] if x[n] = 3"u[n].
(d) Determine the system function, associated ROC, and impulse response for all
LTI systems that satisfy the preceding difference equation but are not causal.
In each case, specify whether the corresponding system is stable.

P23.3
Carry out the proof for the following properties from Table 10.1 of the text (page
654).
(a) 10.5.2
(b) 10.5.3
(c) 10.5.6 [Hint: Consider
dX(z) x n
dz dzn=-_ ]

P23.4

Consider the second-order system with the pole-zero plot given in Figure P23.4. The
poles are located at z = re", z = re- ", and H(1) = 1.

P23-1
Signals and Systems
P23-2

Im
z plane

/ r \
Re

Figure P23.4

(a) Sketch |H(e'")I as 0 is kept constant at w/ 4 and with r = 0.5, 0.75, and 0.9.
(b) Sketch H(ein) as r is kept constant at r = 0.75 and with 0 = r/4, 27r/4, and
37r/ 4.

P23.5

Consider the system function

H(z) =z
(Z - = )(z - 2)

(a) Sketch the pole-zero locations.

(b) Sketch the ROC assuming the system is causal. Is the system stable?
(c) Sketch the ROC assuming the system is stable. Is the system causal?
(d) Sketch the remaining possible ROC. Is the corresponding system either stable or
causal?

P23.6
Consider the continuous-time LTI system described by the following equation:
d 2 y(t) dy(t) dx(t)
+5 +6ytt)=x(t)+2 d
dt2 dt dt
(a) Determine the system function He(s) and the impulse response he(t).
(b) Determine the system function Hd(z) of a discrete-time LTI system obtained
from He(s) through impulse invariance.
(c) For T = 0.01 determine the impulse response associated with Hd(z), hd[n].
(d) Verify that hd[n] = hc(nT) for all n.
 Mapping Continuous-Time Filters to Discrete-Time Filters / Problems
P23-3

Optional
Problems
P23.7

Consider a continuous-time LTI filter described by the following differential

equation:
dy(t) + 0.5y(t) = x(t)
dt
A discrete-time filter is obtained by replacing the derivative by a first forward dif­
ference to obtain the difference equation
y[n + 1] - y[n] + 0.5y[n] = x[n]
T
Assume that the resulting system is causal.
(a) Determine and sketch the magnitude of the frequency response of the continu­
ous-time filter.
(b) Determine and sketch the magnitude of the frequency response of the discrete-
time filter for T = 2.
(c) Determine the range of values of T (if any) for which the discrete-time filter is
unstable.

P23.8
Consider an even sequence x[n] (i.e., x[n] = x[-n]) with rational z-transform X(z).
(a) From the definition of the z-transform show that

Xz) = X (1)
(b) From your result in part (a), show that if a pole (or a zero) of X(z) occurs at
z = zo, then a pole (or a zero) must also occur at z = 1/zo.
(c) Verify the result in part (b) for each of the following sequences:
(i b[n + 1] + b[n - 11
(ii) 6[n + 1] - !Rkn] + b[n - 1]
(d) Consider a real-valued sequence y[n] with rational z-transform Y(z).
(i) Show that Y(z) = Y*(z*).
(ii) From part (i) show that if a pole (or a zero) of Y(z) occurs at z = zo, then
a pole (or a zero) must also occur at z = z*.
(e) By combining your result in part (b) with that in part (d), show that for a real,
even sequence, if there is a pole (or a zero) of H(z) at z = pej", then there is also
a pole (or a zero) of H(z) at z = pe-j", at z = (1/p)e'", and at z = (1/p)e -".
Signals and Systems
P23-4

P23.9

In Section 10.5.5 of the text we stated the convolution property for the z-transform.
To prove this property, we begin with the convolution sum expressed as

x 3 [n] = x 1 [n] * x 2[n] = E x 1 [k]x2[n - k] (P23.9-1)

k= -o

(a) By taking the z-transform of eq. (P23.9-1) and using eq. (10.3) of the text (page
630), show that

XA(z) = x1[k]X 2(z),

k= -o0

where X 2(z) = Z{x2[n - k]}.

(b) Using the result in part (a) and the time-shifting property of z-transforms, show
that

XA(z) = X 2(z) x1[k]z­

k=-o

(c) From part (b), show that

XA(z) = X 1(z)X 2(z).

P23.10

Consider a signal x[n] that is absolutely summable and its associated z-transform
X(z). Show that the z-transform of y[n] = x[n]u[n] can have poles only at the poles
of X(z) that are inside the unit circle.

P23.11

Let hc(t), sc(t), and He(s) denote the impulse response, step response, and system
function, respectively, of a continuous-time, linear, time-invariant filter.
Let hd[n], sd[n], and Hd(z) denote the unit sample response, step response, and
system function, respectively, of a discrete-time, linear, time-invariant filter.

(a) If hd[n] = hc(nT), does Sd[n] = 2 hc(kT)?

k=

(b) If Sd[n] = sc(nT), does hd[n] = hc(nT)?

P23.12

Consider a continuous-time filter with input xc(t) and output yc(t) that is described
by a linear constant-coefficient differential equation of the form
N dk yc(0 M dkxc(t) (P23.12-i)
=0
a k (k=O bk dtk
 Mapping Continuous-Time Filters to Discrete-Time Filters / Problems
P23-5

The filter is to be mapped to a discrete-time filter with input x[n] and output y[n]
by replacing derivatives with central differences. Specifically, let V("k{x[n]} denote
the kth central difference of x[n], defined as follows:
V(43{x[n]} = x[n]
V('[x[n]} x[n + 1] - x[n - 1]

V(k){xln]} V(l){V(k-'){x[nJ}}

The difference equation for the digital filter obtained from the differential equation
(P23.12-1) is then
N M

E
k=O
akV k){y[n]) = Z bkV(k){x[nI}
k=0

(a) If the transfer function of the continuous-time filter is Hc(s) and if the transfer
function of the corresponding discrete-time filter is Hd(z), determine how Hd(z)
is related to He(s).
(b) For the continuous-time frequency response He(jw), as indicated in Figure
P23.12-1, sketch the discrete-time frequency response Hd(ej0 ) that would result
from the mapping determined in part (a).

- 2 4

Figure P23.12

(c) Assume that Hc(s) corresponds to a causal stable filter. If the region of conver­
gence of Hd(z) is specified to include the unit circle, will Hd(z) necessarily cor­
respond to a causal filter?

P23.13
In discussing impulse invariance in Section 10.8.1 of the text, we considered Hc(s)
to be of the form of eq. (10.84) of the text with only first-order poles. In this problem
we consider how the presence of a second-order pole in eq. (10.84) would be
reflected in eq. (10.87) of the text. Toward this end, consider Hc(s) to be
A
He(s) (S - so)2
(a) By referring to Table 9.2 of the text (page 604), determine hc(t). (Assume
causality.)
Signals and Systems
P23-6

(b) Determine hd[n] defined as hd[n] = hc(nT).

(c) By referring to Table 10.2 of the text (page 655), determine Hd(z), the z-trans­
form of hd[n].
(d) Determine the system function and pole-zero pattern for the discrete-time sys­
tem obtained by applying impulse invariance to the following continuous-time
system:
1
He(s) =
(s + 1)(s + 2)2
MIT OpenCourseWare
[Link]

Resource: Signals and Systems

Professor Alan V. Oppenheim

The following may not correspond to a particular course on MIT OpenCourseWare, but has been
provided by the author as an individual learning resource.

For information about citing these materials or our Terms of Use, visit: [Link]
24 Butterworth Filters
Recommended
Problems
P24.1

Do the following for a fifth-order Butterworth filter with cutoff frequency of 1 kHz
and transfer function B(s).
(a) Write the expression for the magnitude squared of the frequency response.
(b) Sketch the locations of the poles of B(s)B(-s).
(c) Indicate the locations of the poles of B(s), assuming that B(s) represents a
causal and stable filter.
(d) Indicate the locations of the poles of B(-s).

P24.2

Figure P24.2-1 shows the frequency response of a discrete-time filter.

Hl(eE )

1­

~
- 2,T IT _ 27T 7
3 3 3 3

Figure P24.2-1

(a) Determine and sketch the analog frequency response characteristic that
(assuming no aliasing) will map to the discrete-time frequency response given
in the figure when the impulse invariance method is used.
(b) Sketch the analog frequency response that will map to the discrete-time fre­
quency response in Figure P24.2-1 when the bilinear transformation is applied.
(c) Repeat parts (a) and (b) for the discrete-time frequency response characteristic
in Figure P24.2-2.

21T
- 27r iTr7

3 3 3

Figure P24.2-2

P24-1
Signals and Systems
P24-2

P24.3
Consider the system function
__1

He(s) +
(a) Determine the discrete-time transfer function Hd(z) obtained by mapping Hc(s)
to Hd(z) using the bilinear transformation with T = 2.
(b) Find the range of the constant a for which H(s) is stable and causal.
(c) Verify that if H(s) is stable and causal, then H(z) is also stable and causal.

P24.4

Consider the following discrete-time filter specifications:

1 Hd(e")| 0.8 for 0 Q

Q<­
4,
0.2 |Hd(ej")| :0 for 3 Q 7r
4
To design Hd(eju) using the impulse invariance method or the bilinear transfor­
mation, we need first to specify a continuous-time filter B(jw). Assume that we will
use a Butterworth filter.
(a) Set up the proper equations for the order N and the cutoff frequency we of the
continuous-time filter B(jw) that will map to Hd(e'n) when the impulse invari­
ance method is used. Set T = 1.
(b) Set up the proper equations for N and w, of B(jo) when the bilinear transfor­
mation is used. Set T = 1.

P24.5
Consider the system in Figure P24.5, which implements a continuous-time filter by
discrete-time processing.

Xc(t) -- > C/D x ilOH(eln) ----- l D/C ye?(t)

Figure P24.5

The sampling frequency is 15 kHz. The continuous-time filter must satisfy the
following specifications:
1 1
|Hc(jw)| 1 0.9 for 0 o 5 (27r)3000,
0.1 IHe(jw)| 1 0 for (27)4500 w < oo
(a) Determine the appropriate specifications for H(ej"), the frequency response of
the discrete-time filter.
(b) Suppose that to design H(ej"), we use the impulse invariance method. We need
to introduce a second continuous-time filter, G(jw). Using T = 3 for the value
of the parameter T in the impulse invariance design procedure, determine the
filter specifications of G(jw).
 Butterworth Filters / Problems
P24-3

(c) Suppose that we now use the bilinear transformation to design H(ej). Using
T = 2 for the value of the parameter T in the bilinear transformation method,
determine the filter specification of G(jw).
(d) With either the impulse invariance or bilinear design procedure, is H(eju) or
H,(jw) dependent in any way on the parameter T?

Optional
Problems
P24.6
In this problem we consider more closely the design procedure for continuous-time
Butterworth filters.
Suppose that we are to design a filter B(jw) such that
20 logioIB(j2r)I > -1 dB,
20 logio IB(j3r)I -15 dB
(a) There are two unknown parameters, the order N of B(s) and the cutoff fre­
quency co,. Set up the two simultaneous equations for N and co, and verify that
N = 5.88 and w, = 7.047 satisfy the equations.
(b) Since N is not an integer, we must choose N to be the next higher integer. We
can now pick whether to meet exactly the stopband specification and exceed the
passband specification or vice versa. Find w, such that the passband specifica­
tion is met exactly, and verify that the stopband specification is exceeded.
(c) What would happen if we picked N = 5?

P24.7
We want to design a discrete-time lowpass filter with a passband magnitude char­
2
acteristic that is constant to within 0.75 dB for frequencies below 0 = 0. 613x and
that has a stopband attenuation of at least 20 dB for frequencies between Q =
0.4018x and r. Determine the poles of the lowest-order Butterworth continuous-
time transfer function that, when mapped to a discrete-time filter using the bilinear
transformation with T = 1, will meet the specifications. If possible, exceed the stop-
band specifications. Indicate also how you would proceed to obtain the transfer
function of the discrete-time filter.

P24.8

Suppose that we want to design a discrete-time filter using the impulse invariance
method. The filter specifications are given by
1 |l Hd(e 0 )I a, 0 s 0 ! 0. 2 r,
b Hd(e")| 0, 0.3r ! Q - r

Using T = 3, we obtain the corresponding filter specifications for the associated

continuous-time filter Hb(s):
3 |Hb(jo)I 3a, 0 : o 5 0.27r/3,
3b |Hb(jo)I -0 0.3r/3 5 w < oo
Signals and Systems
P24-4

Assume that a filter H,(s) satisfies the specifications exactly; thus,

H(O.2) =3a, H,(03) -3b

3 3
The designed discrete-time filter is given by

Hu
H(e") = -1 E Go H, j (11 _2 k)
3 k- _Go 3 3

Study the case of T = 2.

(a) For T = 2, give the filter specifications for the associated continuous-time filter
I(s).
(b) Verify that the continuous-time filter given by
2
He(s) = -H( 2s3)­
3H
satisfies part (a) exactly.
(c) Substitute 11,(s) from part (b) to solve for H(eja) and verify that H(ei") =
H(eju). Thus, the value of T does not affect the final discrete-time filter designed.

P24.9

As mentioned in Section 10.8.3 of the text, the bilinear transformation map from the
s plane to the z plane can be interpreted as arising from the use of the trapezoidal
rule in numerically integrating differential equations.
(a) Consider a continuous-time system for which the differential equation is
dy(t)
dt) = x(t) (P24.9-1)
dt
or, equivalently,

y(t) - x(r) dr (P24.9-2)

Determine the system function H(s) for this continuous-time system.

In numerical analysis the procedure known as the trapezoidal rule for inte­
gration proceeds by approximating the continuous-time function as a set of con­
tiguous trapezoids, as illustrated in Figure P24.9(a), and then adding their areas
to compute the total integral. The areaA of an individual trapezoid, with dimen­
sions shown in Figure P24.9(b), is

A = (b + a) h
2
 Butterworth Filters / Problems
P24-5

(b) What is the area An in the trapezoidal approximation between x[(n - 1)T] and
x(nT)?
(c) From eq. (P24.9-2), y(nT) denotes the area under x(t) up to time t = nT. Let
y[n] denote the approximation to y(nT) obtained using the trapezoidal rule for
integration, that is,

Q[n] = L
k=-0
A

Show that
9[n] = y[n - 1] + An

(d) With -V[n] defined as &[n] = x(nT), show that the trapezoidal rule approxima­
tion to eq. (P24.9-2) becomes

9[n] = P[n - 1] + {2[n - 11 + t[n]} (P24.9-3)

(e) Determine the system function corresponding to the difference equation in part
(d). Demonstrate, in particular, that it is the same as would be obtained by
applying the bilinear transformation to the continuous-time system function
corresponding to eq. (P24.9-1).
MIT OpenCourseWare
[Link]

Resource: Signals and Systems

Professor Alan V. Oppenheim

The following may not correspond to a particular course on MIT OpenCourseWare, but has been
provided by the author as an individual learning resource.

For information about citing these materials or our Terms of Use, visit: [Link]
25 Feedback
Recommended
Problems
P25.1
Consider the two feedback systems in Figures P25.1-1 and P25.1-2.

x(t) H(s) y(t)

K(s)­

Figure P25.1-1

(a) Express the overall system functions

Q Y(s) Q(z) = Y(z)
Q(s) - X(S) and
X(z)

in terms of the system functions in the forward and feedback paths.

For the remainder of the problem, assume that K(s) and K(z) are both constant val­
ues K and that
2 2 2z-' 1
H(s) =s and H(z) = z-2 - 2-
- 2

(b) Plot the pole-zero pattern of Q(s) and Q(z) for K = 0, -1, and 1.
(c) Indicate the locus in the s plane of the poles of Q(s) and in the z plane the poles
of Q(z) as K increases from K = 0 toward large positive values and as K
decreases from K = 0 toward large negative values.
(d) Specify the range of values of K for which each of the systems is stable.

P25-1
Signals and Systems
P25-2

P25.2

For the systems in Figures P25.2-1 to P25.2-3, derive the transfer functions
indicated.

(a)
+ e(t)
x(t) + H(s) y~t

r(t) Gs

Figure P25.2-1

(i Y(s)
X(s)
E(s)
(ii) X(s)
X(s)

(iii) Y(s)
E(s)

(iv) R(S)

(b) Y(z)
X(z)

+w[n] +
x [n] x.+ H, (z) + y[n]

G(z)

H 0 (z)

Figure P25.2-2

(C) Y(s)
X(s)
 Feedback / Problems
P25-3

P25.3

Consider the system G(s) with frequency response G(jw) indicated in Figure
P25.3-1.

G(jw)

100
5 __
-wc (oc

Figure P25.3-1

(a) Sketch the approximate overall frequency response of each of the feedback sys­
tems in Figure P25.3-2.

x (t) y(t)

SYSTEM 1

x(t) y(t)

SYSTEM 2
Figure P25.3-2
Signals and Systems
P25-4

(b) Determine which of the two systems in part (a) attempts to equalize the gain
G(jw), i.e., to make it more constant over all frequencies, and which system is
intended as an approximation to a system that is the inverse of G(jw).

P25.4
Consider the basic feedback systems of Figure P25.4. Determine the closed-loop sys­
tem impulse response for each of the following specifications of the system func­
tions in the forward and feedback paths.

x(t) ±
+y v(t)

r(t)
G

(a)

x[n] + e[n] y[
v[n]

Figure P25.4

1
(a) H(s) = G(s) = 1
(s + 1)(s + 3)
1
(b) H(s) = G(s) = s + 1
s + 3
(c) H(s) = 1, G(s) = e -s/3
-1
(d) H(z) = 1 Z 1' G(z) = 1 e
-2z

(e) H(z) = 2 - iz-, G(z) =

1 ­
 Feedback / Problems
P25-5

Optional
Problems
P25.5

For the system given in Figure P25.5, find Y(s) in terms of X(s) and W(s).

w(t)

x (t) y(t)

Figure P25.5

P25.6
Consider the causal discrete-time system depicted in Figure P25.6.

x[n] yln]

Figure P25.6
(a) Show that this is not a stable system.
(b) Suppose that we allow feedback with one unit of delay. That is, suppose that
x[n] = xZ,[n] - Ky[n - 1],

where xe[n] is an externally applied signal now regarded as the input to the
overall closed-loop system. Is it possible to stabilize the system with feedback
of this form? If so, find the maximum range of values of K for which the system
is stable.
Signals and Systems
P25-6

(c) Suppose that instead of using feedback with one unit of delay as in part (b), we
allow feedback with two units of delay, so that
x[n] = x,[n] - Ky[n - 21

Specify the full range of values of K (if any such values exist) for which this
system is stable.

P25.7
In this problem we illustrate how feedback can be used to increase the bandwidth
of an amplifier. Consider an amplifier whose gain falls off at high frequencies. Spe­
cifically, suppose that the system function of this amplifier is
Ga
H(s) = +a
s + a

(a) What is the dc gain of the amplifier (i.e., the magnitude of its frequency
response to zero frequency)?
(b) What is the system time constant?
(c) Suppose that we define bandwidth as the frequency at which the magnitude of
the amplifier frequency response is 1/ \2 times its magnitude of dc. What is the
bandwidth of this amplifier?
(d) Suppose that we place this amplifier in a feedback loop as depicted in Figure
P25.7. What is the dc gain of the closed-loop system? What are the time constant
and bandwidth of the closed-loop system?

x(t) + ayMt

Figure P25.7

(e) Find the value of K that leads to a closed-loop bandwidth that is exactly double
the bandwidth of the open-loop amplifier. What are the corresponding closed-
loop system time constant and dc gain?
MIT OpenCourseWare
[Link]

Resource: Signals and Systems

Professor Alan V. Oppenheim

The following may not correspond to a particular course on MIT OpenCourseWare, but has been
provided by the author as an individual learning resource.

For information about citing these materials or our Terms of Use, visit: [Link]
26 Feedback Example:
The Inverted Pendulum
Recommended
Problems
P26.1
Consider an inverted pendulum mounted on a movable cart as depicted in Figure
P26.1. We have modeled the pendulum as consisting of a massless rod of length L,
with a mass m attached at the end. The variable 0(t) denotes the pendulum's angular
deflection from the vertical, g is gravitational acceleration, s(t) is the position of the
cart with respect to some reference point, a(t) is the acceleration of the cart, and
x(t) represents the angular acceleration resulting from any disturbances, such as
gusts of wind.

NIM

L L I x(t)

0(t)
N9

a(t)

s(t)

Figure P26.1

Our goal in this problem is to analyze the dynamics of the inverted pendulum
and more specifically to investigate the problem of balancing the pendulum by judi­
cious choice of the cart acceleration a(t). The differential equation relating 0(t),
a(t), and x(t) is
d 20(t)
L dt = g sin [(t)] - a(t) cos [(t)] + Lx(t) (P26.1-1)

This relation merely equates the actual acceleration of the mass along a direction
perpendicular to the rod and the applied accelerations (gravity, the disturbance
acceleration due to x(t), and the cart acceleration) along this direction.
Note that eq. (P26.1-1) is a nonlinear differential equation relating 0(t), a(t),
and x(t). The detailed, exact analysis of the behavior of the pendulum therefore
requires that we examine this nonlinear equation; however, we can obtain a great
deal of insight into the dynamics of the inverted pendulum by performing a linear
analysis. Specifically, we will examine the dynamics of the pendulum when it is
nearly vertical (i.e., when 0(t) is small). In this case, we can make the approx­
imations
sin [(t)] 0(t),
0 cos [(t)] ~ 1 (P26.1-2)

P26-1
Signals and Systems
P26-2

(a) Suppose that the cart is stationary (i.e., a(t) = 0) and consider the causal LTI
system with input x(t) and output 0(t) described by eq. (P26.1-1) together with
the approximations given in eq. (P26.1-2). Find the system function for this sys­
tem and show that it has a pole in the right half-plane, implying that the system
is unstable.
The result of part (a) indicates that if the cart is stationary, any minor angular
disturbance caused by x(t) will lead to growing angular deviations from the verti­
cal. Clearly, at some point these deviations will become sufficiently large that the
approximations of eq. (P26.1-2) will no longer be valid. At this point the linear anal­
ysis is no longer accurate, but the fact that it is accurate for small angular displace­
ments allows us to conclude that the vertical equilibrium position is unstable, as
small angular displacements will grow rather than diminish.
(b) We now wish to consider the problem of stabilizing the vertical position of the
pendulum by moving the cart in an appropriate fashion. Suppose that we try
proportional feedback, with a(t) = KO(t). Assume that 0(t) is small so that the
approximations in eq. (P26.1-2) are valid. Draw a block diagram of the linear
system with 0(t) as output, x(t) as the external input, and a(t) as the signal
that is fed back. Show that the resulting closed-loop system is unstable. Find a
value of K such that if x(t) = 5(t), the pendulum will sway back and forth in
an undamped oscillatory fashion.
(c) Consider using proportional plus derivative (PD) feedback:
d0(t)
a(t) = K 10(t) + K 2 dt
dt
Show that we can find values of Ki and K 2 that do stabilize the pendulum. In
fact, using
g = 9.8 m/s', L = 0.5 m, (P26.1-3)
choose values of Ki and K 2 so that the damping ratio of the closed-loop system
is 1 and the natural frequency is 3 rad/s.

P26.2

Consider the feedback system given in Figure P26.2.

x(t) : s2+(Jo 2 0 10 y (t

K1

Figure P26.2

(a) The closed-loop system is a new second-order system. Find the new natural fre­
quency ,, and the new damping factor k for K = 1.
(b) Find the poles of the closed-loop transfer function.
(c) Sketch the trajectory of the poles as K goes from negative infinity to positive
infinity.
 Feedback Example: The Inverted Pendulum / Problems
P26-3

P26.3
Consider the system in Figure P26.3.

w (t)

x (t) y W)

Figure P26.3

(a) Find X = H1(s).

X(S)

(b) Find = H 2 (s).

W(s)
(c) For what range of values of a, 0, and r will both H1 (s) and H2 (s) be stable?

P26.4
Consider the feedback system in Figure P26.4.

x(t ) +K .y (t)

S +1
s + 100

Figure P26.4

Find the closed-loop poles and zeros of this system for the following values of K:
(a) K = 0.01
(b) K = 1
(c) K = 10
(d) K = 100
Signals and Systems
P26-4

P26.5

Sketch the root loci for K > 0 and K < 0 for each of the following systems.
(a)

x(t) y (t)

K 1

Figure P26.5-1

(b)

x (t) S y(t)

s+ 3

Figure P26.5-2

Optional
Problems
P26.6

Consider again the discrete-time feedback system of Example 11.3 of the text (page
712):

G(z)H(z) = ______

1)(z - 4)
(z -
The root loci for K > 0 and K < 0 are depicted in Figure P26.6.
(a) Consider the root locus for K > 0. In this case the system becomes unstable
when one of the closed-loop poles is less than or equal to -1. Find the value of
K for which z = -1 is a closed-loop pole.
(b) Consider the root locus for K < 0. In this case the system becomes unstable
when one of the closed-loop poles is greater than or equal to 1. Find the value
of K for which z = 1 is a closed-loop pole.
(c) What is the full range of values of K for which the closed-loop system is stable?
 Feedback Example: The Inverted Pendulum / Problems
P26-5

K>O Unit circle

1/
(a)

K<0 - Unit circle

4 2 /

(b)

Figure P26.6

P26.7

In this problem we consider several examples of the design of tracking systems.

Consider the system depicted in Figure P26.7. Here H,(s) is the system whose output
is to be controlled, and He(s) is the compensator to be designed. Our objective in
choosing H,(s) is that we would like the output y(t) to follow the input x(t). Spe­
cifically, in addition to stabilizing the system, we would also like to design the sys­
tem so that the error e(t) decays to zero for certain specified inputs.

(a) Suppose that

H,(s) = a a =A0 (P26.7-1)

S + a

Show that if He(s) = K (which is known as proportional or P control), we can

choose K to stabilize the system so that e(t) goes to 0 if x(t) = b(t). Show that
we cannot get e(t) - 0 if x(t) = u(t).
Signals and Systems
P26-6

+ et) dt)
x(t) + He,(s) HP(s) ,, W
t

Figure P26.7

(b) Again, let H,(s) be as in eq. (P26.7-1) and suppose that we use proportionalplus
integral(PI) control:

He(s) = K 1 + K2
S

Show that we can choose K, and K2 to stabilize the system, and we can also get
e(t) - 0 if x(t) = u(t). Thus, this system can track a step. In fact, this illus­
trates a basic and important fact in feedback system design: To track a step
(X(s) = 1/s), we need an integrator (1/s) in the feedback system.

P26.8

Consider a system with system function

(P26.8-1)
(s + 1)(s - 2)
This system is unstable, and we would like to devise some method for its sta­
bilization.
(a) Consider first a series compensation scheme as illustrated in Figure P26.8-1.

x(t) C(s) H(s) y(t)

Figure P26.8-1

Show that the overall system of this figure is stable if the system function C(s)
is taken as

C(s) = 2
s + 3
In practice, this is not considered a particularly useful way to attempt to sta­
bilize a system. Explain why.
(b) Suppose that instead we use a feedback system as depicted in Figure P26.8-2.
Is it possible to stabilize this system using a constant gain for the compensating
element, i.e., C(s) = K? Justify your answer.
 Feedback Example: The Inverted Pendulum / Problems
P26-7

x(t) + C(s) - H(s) y(t)

Figure P26.8-2

(c) Show that the system of Figure P26.8-2 can be stabilized if C(s) is a proportional
plus derivative system:
C(s) = K(s + a)
Consider both 0 < a < 1 and a > 1.
(d) Suppose that C(s) = K(s + 2). Choose the value of K so that the closed-loop
system has a pair of complex poles with a damping ratio of = 1.[Hint: In this
case the denominator of the closed-loop system function must have the form
s2+ S + W2.}

P26.9
In Problem P26.7 we discussed how the presence of an integrator in a feedback sys­
tem can make it possible for the system to track a step input with zero error in
steady state. This problem extends this idea. Specifically, consider the feedback sys­
tem depicted in Figure P26.9, and assume that the overall closed-loop system is sta­
ble. Suppose also that H(s) has the form

K f (s - Ok)
H(s) =
s' 11 (s - ak)
k=1

where a and #kare given, nonzero numbers and I is a positive integer. The feedback
system of Figure P26.9 is often referred to as a Type 1feedback system.

x(t) F P26
H(s)

Figure P26.9
Signals and Systems
P26-8

(a) Use the final value theorem (Section 9.5.9 of the text) to show that a Type 1
feedback system can track a step, i.e., that
e(t) - 0 if x(t) = u(t)
(b) Similarly, show that a Type 1 system cannot track a ramp but rather that
e(t) - a finite constant if x(t) = U 2 (t)
(c) Show that for a Type 1 system, unbounded errors result if x(t) = uk(t), with
k > 2.
(d) More generally, show that for a Type 1system,
(i) e(t) - 0 if x(t) = U k(t), with ks 1
(ii) e(t) - a finite constant if x(t) = Um e(t)
(iii) e(t) -- oc if x(t) = uk(t), with k > 1+ 1

P26.10

(a) Consider the discrete-time feedback system of Figure P26.10. Suppose that
1
H(z) =
(z - 1)(z + i)
Show that this system can track a unit step in the sense that if x[n] = u[n], then
lim e[n] = 0 (P26.10-1)

(b) More generally, consider the feedback system of Figure P26.10 and assume that
the closed-loop system is stable. Suppose that H(z) has a pole at z = 1. Show
that the system can track a unit step. [Hint: Express the transform E(z) of e[n]
in terms of H(z) and the transform of u[n]; explain why all the poles of E(z) are
inside the unit circle.]
The results of parts (a) and (b) are discrete-time counterparts of the results for con­
tinuous-time systems discussed in Problems P26.7 and P26.9. In discrete time, we
can also consider the design of systems that track specified inputs perfectly after a
finite number of steps. Such systems are known as deadbeatfeedback systems, sev­
eral examples of which are illustrated in the remainder of this problem.
(c) Consider the discrete-time system of Figure P26.10, with
z1
H(z) = Z­
1- z
Show that the overall closed-loop system has the property that it tracks a step
input exactly after one step. That is, if x[n] = u[n], then e[n] = 0, n > 1.
 Feedback Example: The Inverted Pendulum / Problems
P26-9

(d) Show that the feedback system of Figure P26.10, with

a + .1z2
H(z) =
-1iz-1)(1 - Z-')'

has the property that the output tracks a unit step perfectly after a finite num­
ber of steps. At what time step does the error e[n] first settle to zero?
(e) More generally, consider the feedback system of Figure P26.10. Find H(z) such
that y[n] perfectly tracks a unit step for n > N and in fact such that
N-1

e[n] = ak 6[n - k], (P26.10-2)

k=O

where the ak are specified constants. [Hint: Use the relation between H(z) and
E(z) when the input is a unit step and e[n] is given by eq. (P26.10-2).]
(f) Consider the system of Figure P26.10, with

H(z) =
(1 + z-1)(1 - z-1)2
Show that this system tracks a ramp x[n] = (n + 1)u[n] exactly after two time
steps.
MIT OpenCourseWare
[Link]

Resource: Signals and Systems

Professor Alan V. Oppenheim

The following may not correspond to a particular course on MIT OpenCourseWare, but has been
provided by the author as an individual learning resource.

For information about citing these materials or our Terms of Use, visit: [Link]