220 Croft,Davison and Hargreaves,Engineering Mathematics,Solutions Manual

Chapter 24. The Fourier transform

Exercises 24.2
1. (a)

 3
1 −jωt
F (ω) = e dt
−3 4
1  −jωt 3
= − e −3
4jω
1  −3jωt 
= − e − e3jωt
4jω
1
= sin 3ω
2ω

(b)

 0 
  2 
t t
F (ω) = e−jωt dt +
1+ 1− e−jωt dt
−2 2 0 2
 
0  0
t e−jωt 1 e−jωt
= 1+ − dt
2 −jω −2 −2 2 −jω
 
2  2
t e−jωt 1 e−jωt
+ 1− + dt
2 −jω 0 0 2 −jω

0
2
1 1 e−jωt 1 1 e−jωt
= − + +
−jω 2 −ω 2 −2 jω 2 −ω 2 0
1 1
= − 2 cos 2ω
ω2 ω

(c)

 0  ∞
αt −jωt
F (ω) = e e dt + e−αt e−jωt dt
−∞ 0

∞ −(α+jω)t
∞
e(α−jω)t e
= +
α − jω −∞ −(α + jω) 0
1 1
= +
α − jω α + jω
2α
=
α2 + ω 2

(d)  ∞
F (ω) = e−(1+jω)t cos t dt
0
Let
 ∞
I = e−(1+jω)t cos t dt
0

Pearson
c Education Limited 2001
Croft,Davison and Hargreaves,Engineering Mathematics,Solutions Manual 221

 ∞  ∞
−(1+jω)t
= e (1 + jω) sin t e−(1+jω)t dt
sin t +
0
 ∞  ∞
0

−(1+jω)t −(1+jω)t
= (1 + jω) −e cos t − (1 + jω) cos t e dt
0 0
= (1 + jω) {1 − (1 + jω)I}

from which

1 + jω
F (ω) = I =
2 − ω 2 + 2jω

(e)

 ∞
F (ω) = e−t/τ e−jωt dt
0

∞
e−(jω+1/τ )t
=
−(jω + 1/τ ) 0
1
=
jω + 1/τ
τ
=
τ jω + 1

2. 
∞
∞
e−(α+jω)t 1
F{u(t)e−αt } = e−(α+jω)t dt = =
0 −(α + jω) 0 α + jω
 ∞ −(s+α)t
∞
e 1
L{u(t)e−αt } = e−αt e−st dt = =
0 −(s + α) 0 s + α
Setting s = jω in the Laplace transform results in the Fourier transform.

3.

 2
F{f (t)g(t)} = e(1−ω)jt dt
−2

2
e(1−ω)jt
=
j(1 − ω) −2
2(1−ω)j −2(1−ω)j
e e
= −
j(1 − ω) j(1 − ω)
2
= sin 2(1 − ω)
1−ω

Exercises 24.3
1. The first shift theorem.  ∞
F{f (t)} = f (t)e−jωt dt = F (ω)
−∞

Pearson
c Education Limited 2001
222 Croft,Davison and Hargreaves,Engineering Mathematics,Solutions Manual

 ∞
F{ejat f (t)} = f (t)ejat e−jωt dt
−∞
 ∞
= f (t)e−jt(ω−a) dt
−∞
= F (ω − a)

2.
 1
F (ω) = (1 − t2 )e−jωt dt
−1

1  1 −jωt
(1 − t2 )e−jωt e
= − (−2t)dt
−jω −1 −1 −jω
 1
2
= − te−jωt dt
jω −1

1  1 −jωt 
2 te−jωt e
= − − dt
jω −jω −1 −1 −jω

2 2 cos ω e−jω − ejω
= − +
jω −jω ω2
4 sin ω 4 cos ω
= −
ω3 ω2

(a) Let a = 3 in the first shift theorem:

4 sin(ω − 3) 4 cos(ω − 3)
F (ω − 3) = −
(ω − 3)3 (ω − 3)2

(b) Let a = j:
4 sin(ω − j) 4 cos(ω − j)
F (ω − j) = −
(ω − j)3 (ω − j)2

3. (a) Noting that

1
F{u(t)e−t } = = F (ω), say
jω + 1
then
1
F (ω + 7) = = F{e−7jt f (t)}
j(ω + 7) + 1
via the first shift theorem.
(b) Noting that
1
F{u(t)e−t/2 } = = F (ω), say
ωj + 1/2
we have
2 1
= = F (ω − 1)
1 + 2(ω − 1)j (ω − 1)j + 1/2
Thus the required inverse is
u(t)e−t/2 ejt

Pearson
c Education Limited 2001
Croft,Davison and Hargreaves,Engineering Mathematics,Solutions Manual 223

4. The second shift theorem:  ∞

F{f (t)} = f (t)e−jωt dt = F (ω)
−∞
 ∞
F{f (t − α)} = f (t − α)e−jωt dt
−∞
Let t − α = z say, then

 ∞
F{f (t − α)} = f (z)e−jω(z+α) dz
z=−∞
 ∞
= e−jωα f (z)e−jωz dz
−∞
= e−jωα F (ω)

5.
1
f (t) = u(t)e−t , F (ω) =
1 + jω
f (t + 4) = u(t + 4)e−(t+4)
By the second shift theorem
e4jω
F{f (t + 4)} = e4jω F (ω) =
1 + jω

6. From the table 

sin 2ω 1/2 −2 ≤ t ≤ 2
F −1 =
ω 0 otherwise
Using the second shift theorem with α = 4 we find

−1 −4jω sin 2ω 3 2≤t≤6
F 6e =
ω 0 otherwise

Exercises 24.4

 0  1
F (ω) = (t + 1)e−jωt dt + e−jωt dt
−1 0

0  0 −jωt
1  1 −jωt
e−jωt e (1 − t)e−jωt e
= (t + 1) − dt + + dt
−jω −1 −1 −jω −jω 0 0 −jω
−jωt
0 −jωt
1
1 e 1 e
= + + −
−jω ω 2
−1 jω ω2 0
1 ejω e−jω 1
= 2
− 2 − 2 + 2
ω ω ω ω
2
= (1 − cos ω)
ω2

The pulse which would result if f (t) were passed through a filter which removed all angular frequencies greater
than 2π would be  2π
1 2
(1 − cos ω)ejωt dω
2π −2π ω 2

Pearson
c Education Limited 2001
224 Croft,Davison and Hargreaves,Engineering Mathematics,Solutions Manual

Exercises 24.8

1. Refer to Figure 24.8.1. Figure 24.8.1

4
g(λ)
a)

f (λ)
-4 -3 -2 -1 0 1 2 3 4 λ

g(−λ) 4

b)

-4 -3 -2 -1 0 1 2 3 4 λ

g(t − λ) when t < −1 4

c)

-4 -3 -2 -1 0 1 2 3 4 λ

4 g(t − λ) when − 1 < t < 2

t
d)

-4 -3 -2 -1 0 1 2 3 4 λ

4
g(t − λ) when 2 < t < 3

e)

-4 -3 -2 -1 0 1 2 3 4 λ

4
g(t − λ) when 3 < t < 6
t− 
f)

-4 -3 -2 -1 0 1 2 3 4
λ

4
g(t − λ) when t > 6
g)

-4 -3 -2 -1 0 1 2 3 4 λ

Pearson
c Education Limited 2001
Croft,Davison and Hargreaves,Engineering Mathematics,Solutions Manual 225

Figure 24.8.1(b) shows g(−λ). f (λ) is shown dashed.

When t ≤ −1, (c), the graphs do not overlap and hence f ∗ g = 0.

When −1 < t < 2, (d), graphs overlap between 0 and 1 + t. Why ? When t = 0 g(t − λ) = g(−λ) as in b). For
t > 0, g(t − λ) is the graph in (b) translated t units to the right as in (d). Hence the overlap is between λ = 0
and λ = 1 + t.
 1+t 1+t
2λ 4λ2  4(1 + t)2
f ∗g = 4 dλ = =
0 3 3 0 3

 3
2λ 3
4λ2 
for t > 2 but t < 3, (e), f ∗g = dλ =
4 = 12
0 3 3 0
 3 3
2λ 4λ2  4
for t > 3 but t < 6, (f), 4 dλ =  = 12 − (t − 3)2
t−3 3 3 t−3 3
When t > 6 the graphs do not overlap, (g), and f ∗ g = 0.

2. If t < −2, f ∗ f = 0.
If −2 < t < 0

 t+1
t+1
f ∗f = f dλ = [t]−1
−1
= (t + 1) − (−1)
= t+2

If 0 < t < 2

 1
1
f ∗f = f dλ = [t]−1+t
−1+t
= 1 − (−1 + t)
= 2−t

if t > 2, f ∗ g = 0.

Pearson
c Education Limited 2001
226 Croft,Davison and Hargreaves,Engineering Mathematics,Solutions Manual

3.

a) f(λ)
1

λ
b)
1
g(λ)

λ
c)
g( λ− t ) when t < 0 1

t −1 1 λ

Referring to the figure above

if t < −1

 1
f g = e−(λ−t) dλ
−1
= e [−e−λ ]1−1
t

= et (e1 − e−1 )

if −1 < t < 1

 1
f g = e−(λ−t) dλ
t
= et [−e−λ ]1t
= 1 − et−1

if t > 1 f  g = 0.

2 sin ω 1 2 sin ω(1 + jω)

F (ω) = , G(ω) = , F (ω)G(−ω) =
ω 1 + jω ω(1 + ω 2 )
 ∞
4. f ∗ g = f (λ)g(t − λ)dλ.
−∞

dµ
Let µ = t − λ, i.e. λ = t − µ, = −1.
dλ

Pearson
c Education Limited 2001
Croft,Davison and Hargreaves,Engineering Mathematics,Solutions Manual 227

Then
 µ=−∞
f ∗g = f (t − µ)g(µ)(−dµ)
µ=∞
 ∞
= + g(µ)f (t − µ)dµ
−∞
∞
= + g(λ)f (t − λ)dλ since µ is a dummy
−∞
= g∗f as required

5. If f (t) = 0 for t < 0, then f (λ) = 0 for λ < 0.

Then  ∞
f ∗g = f (λ)g(t − λ)dλ
0

Also if g(t) = 0 for t < 0

g(t − λ) = 0 for t − λ < 0
i.e. g(t − λ) = 0 if λ > t
and hence
 t
f ∗ g reduces to f ∗ g = f (λ)g(t − λ)dλ as required.
0

6. Convolution theorem: 

 ∞ ∞
F{f ∗ g} = f (λ)g(t − λ)dλ e−jωt dt
−∞ −∞

Interchange order of integration:

 ∞  ∞

−jωt
F{f ∗ g} = f (λ)g(t − λ)e dt dλ
−∞ −∞
 ∞  ∞

−jωt
= f (λ) g(t − λ)e dt dλ
−∞ −∞

Let t − λ = µ, for fixed λ

 
as t→∞ µ→∞
t → −∞ µ → −∞

 ∞  ∞

−jω(µ+λ)
F{f ∗ g} = f (λ) g(µ)e dµ dλ
−∞ −∞
 ∞  ∞
−jωλ
= f (λ)e dλ g(µ)e−jωµ dµ
−∞ −∞
= F (w)G(w) as required

Pearson
c Education Limited 2001
228 Croft,Davison and Hargreaves,Engineering Mathematics,Solutions Manual

7.  ∞
f g = f (λ)g(λ − t)dλ
−∞

Let g(t) = h(−t) g(λ − t) = h(t − λ)

 ∞
f (t)  g(t) = f (t)  h(−t) = f (λ)h(t − λ)dλ
−∞
= f (t) ∗ h(t)
= f (t) ∗ g(−t) as required

And hence the convolution of f (t) and h(t) can be found by calculating the correlation of f (t) and g(t) where
g(t) = h(−t) and vice versa.


f (t)  g(t) = f (λ)g(λ − t)dλ

f (t)  g(−t) = f (λ)g(t − λ)dλ

= f (t) ∗ g(t) as required

8.  ∞
f g = f (λ)g(λ − t)dλ
−∞

Let λ − t = µ, then
 ∞
f g = f (µ + t)g(µ)dµ (µ is a dummy variable)
−∞
 ∞
= g(λ)f (t + λ)dλ as required
−∞

9.
 ∞  ∞

F {f  g} = f (λ)g(λ − t)dλ e−jωt dt

−∞ −∞
 ∞  ∞
= f (λ) g(λ − t)e−jωt dt dλ
−∞ −∞

Let λ − t = µ, for fixed λ

 
t → −∞ µ → ∞
t → ∞ µ → −∞

 ∞  −∞
F {f  g} = g(µ)e−jω(λ−µ) (−dµ) dλ
f (λ)
−∞ ∞
 ∞  ∞
−jωλ
= f (λ)e dλ g(µ)ejωµ dµ
−∞ −∞

Pearson
c Education Limited 2001
Croft,Davison and Hargreaves,Engineering Mathematics,Solutions Manual 229

Now
 ∞
G(ω) = g(t)e−jωt dt
−∞
 ∞
G(−ω) = g(t)ejωt dt
−∞

So
F {f  g} = F (ω)G(−ω) as required.

Exercises 24.9.1
1.

f [n] = 1, 2, 0, −1. N =4

3
F [k] = f [n]e−jnkπ/2
n=0

3
F [0] = f [n] = 1 + 2 + 0 + (−1) = 2.
n=0


F [1] = f [n]e−jnπ/2 = 1 + 2e−jπ/2 + 0 − 1e−j3π/2
= 1 − 2j − j
= 1 − 3j


F [2] = f [n]e−jnπ = 1 + 2(−1) + 0 − 1(−1)
= 1−2+1
= 0


F [3] = f [n]e−jn3π/2 = 1 + 2j + 0 − 1(−j)
= 1 + 2j + j
= 1 + 3j

g[n] = 3, 1, −1, 1. N =4

3
G[k] = g[n]e−jnkπ/2
n=0
G[0] = 3+1−1+1=4
G[1] = 3 + e−jπ/2 − 1e−jπ + 1e−3jπ/2
= 3−j+1+j
= 4

Pearson
c Education Limited 2001
230 Croft,Davison and Hargreaves,Engineering Mathematics,Solutions Manual

G[2] = 3 + 1e−jπ − 1e−2jπ + e−3jπ

= 3−1−1−1
= 0
G[3] = 3 + 1e−3jπ/2 − 1e−3jπ + 1e−9jπ/2
= 3+j+1−j
= 4

2.

f [n] = 5, −1, 2 N =3

2
F [k] = f [n]e−2jnkπ/3
n=0

F [0] = f [n] = 6

F [1] = f [n]e−2jnπ/3
= 5 − 1e−2jπ/3 + 2e−4jπ/3
   
2π 2π 4π 4π
= 5 − cos − j sin + 2 cos − j sin
3 3 3 3
 √   √ 
1 3 1 3
= 5− − − j +2 − +j
2 2 2 2
√
1 3 √
= 5+ + j−1+j 3
2√ 2
9 3 3
= + j
2 2
= 4.5 + 2.5981j


F [2] = f [n]e−4jnπ/3
   
4π 4π 8π 8π
= 5 − 1 cos − j sin + 2 cos − j sin
3 3 3 3
 √   √ 
1 3 1 3
= 5− − +j +2 − −j
2 2 2 2
√
1 3 √
= 5+ −j −1−j 3
2√ 2
9 3 3
= − j
2 2
= 4.5 − 2.5981j


N −1
4. If F [k] = f [n]e−2jnkπ/n
n=0

rhen taking the complex conjugate of both sides gives

Pearson
c Education Limited 2001
Croft,Davison and Hargreaves,Engineering Mathematics,Solutions Manual 231


N −1
F [k] = f [n] e2jnkπ/N
n=0

(since if z = ejθ then z = e−jθ and z1 z2 = z1 · z2 ). If f [n] is real then f [n] = f [n]
and the result follows immediately.

5.


N −1
F [k] = f [n]e−2jnπk/N
n=0

N −1
F [k] = f [n]e2jnπk/N if f is real
n=0


N −1
F [N − k] = f [n]e−2jnπ[N −k]/N
n=0

= f [n]e−2jnπN/N e2jnπk/N

= f [n]e−2jnπ e2jnπk/N

= f [n]e+2jnπk/N since e−2jnπ = 1
= F [k] as required

Exercises 24.9.2

1.
N −1
1 
f [n] = F [k]e2jknπ/N n = 0, 1, 2, . . . N − 1
N
k=0

F [k] = 6, −2, 2, −2. N =4

1 
3
f [n] = F [k]ejknπ/2
4
k=0
1 1
f [0] = F [k] = (6 − 2 + 2 − 2) = 1.
4 4
1 jkπ/2
f [1] = F [k]e
4
1  
= 6 − 2ejπ/2 + 2ejπ − 2ej3π/2
4
1
= {6 − 2j − 2 + 2j}
4
1
= {4} = 1
4

Pearson
c Education Limited 2001
232 Croft,Davison and Hargreaves,Engineering Mathematics,Solutions Manual

1 1
f [2] = F [k]ejkπ = (6 − 2(−1) + 2(1) − 2(−1))
4 4
1
= (12) = 3
4

1 1 
f [3] = F [k]ejk3π/2 = 6 − 2ej3π/2 + 2ej3π − 2ej9π/2
4 4
1
= (6 + 2j − 2 − 2j) = 1
4

N −1
1 
3. Consider F [k]e2jknπ/N
N
k=0

N −1
N −1 
1  
−2jmkπ/N
= f [m]e e2jknπ/N
N m=0
k=0

Interchange the order

N −1 −1
1  
N
= f [m] e−2jkπ(m−n)/N (∗)
N m=0
k=0

Now study this sum over m. As m varies between 0 and N − 1, for some value m will equal n. Consider this
first.
With m = n, this becomes


N −1
1 1
f [n] 1 = f [n]N (1)
N N
k=0
= f [n] (2)


N −1
Now consider any other value of m with m = n. Consider e−2jkπ(m−n)/N
k=0


N −1  k
= e−2jπ(m−n)/N
k=0

N −1
= rk where r = e−2jπ(m−n)/N
k=0
= 1 + r + r2 + . . . rN −1

Pearson
c Education Limited 2001
Croft,Davison and Hargreaves,Engineering Mathematics,Solutions Manual 233

This is a geometric progression with N terms, first term 1 and common ratio r.

   
1 1 − rN 1 1 − e−2jπ(m−n)
SN = =
1−r 1 − e−2jπ(m−n)/N
= 0 since e−2jπ(m−n) = 1 for m, n integers

Therefore (∗) reduces to f [n] as required.

Exercises 24.12
1. N = 4, W = e−jπ/2

W1 = e−jπ/2 = −j
W2 = j 2 = −1
W3 = j
..
.
W 9 = jjj = j 3 = −j

 
W0 W0 W0 W0
W0 W1 W2 W 
3
W = 0 
W W2 W4 W6
W0 W3 W6 W 9

 
1 1 1 1
1 −j −1 j 
= 
1 −1 1 −1
1 j −1 −j

    
1 1 1 1 1 3 2 4
 1 −j −1 j   2 1   1 − 3j 4
F [k] =   = 
1 −1 1 −1 0 −1 0 0
1 j −1 −j −1 1 1 + 3j 4

Exercises 24.13
1. (a)
    
1 1 1 1 1 2 6 6
 1 −j −1 j   2 1   −1 − j 1+j
  = 
1 −1 1 −1 2 1 0 0
1 j −1 −j 1 2 −1 + j 1−j

(b)

3
| f [n] |2 = 1 + 4 + 4 + 1 = 10
n=0

1 1 1
| F [k] |2 = {36 + (1 + 1) + 0 + (1 + 1)} = 40 = 10
4 4 4

Pearson
c Education Limited 2001
234 Croft,Davison and Hargreaves,Engineering Mathematics,Solutions Manual

2.
f = 3 1 5 4
g = 2 −1 9 5
fg = 6 −1 45 20


(a) f g = 70
(b)
    
1 1 1 1 3 2 13 15
 1 −j −1 j   1 −1   −2 + 3j −7 + 6j 
F, G =   = 
1 −1 1 −1 5 9 3 7
1 j −1 −j 4 5 −2 − 3j −7 − 6j

(c) (With multiplication term by term)

      
13 15 195 195
 −2 + 3j   −7 − 6j   14 − 21j + 12j + 18   32 − 9j 
F Ḡ =   = = 
3 7 21 21
−2 − 3j −7 + 6j 14 + 21j − 12j + 18 32 + 9j

(d)
1  1 280
F Ḡ = (195 + 32 + 21 + 32) = = 70
N 4 4

3.


N −1
F [k] = f [n]e−2jnkπ/N = D{f [n]}
n=0

N −1
D{f [n − i]} = f [n − i]e−2jnπk/N
n=0

N −1
= f [n − i]e−2jπk(n−i)π/N e−2jπki/N
n=0
−2jkπi/N
= e F [k]


N −1
This development assumes that F [k] = f [n − i]e−2jπk(n−i)/N .
n=0

This is true. Note that by letting n − i = p this becomes

−1−i
N
F [k] = f [p]e−2jπkp/N
p=−i

Because of periodicity in f we are summing over a complete cycle.

Pearson
c Education Limited 2001
Croft,Davison and Hargreaves,Engineering Mathematics,Solutions Manual 235

4. Parseval’s theorem:

N −1 N −1
N −1 
1  1  
−2πjnk/N
F [k]G[k] = f [n]e G[k]
N N n=0
k=0 k=0
N −1 −1
1  
N
= f [n] G[k]e−2πjnk/N (∗)
N n=0
k=0
N −1
1 
but g[n] = G[k]e2πjkn/N
N
k=0
1 
therefore g[n] = G[k]e−2πjkn/N
N

N −1
therefore (∗) becomes f [n]g[n]
n=0

Hence
 N −1
1 
f [n]g[n] = F [k]G[k]
N
k=0

(Parseval)

5. Let f = g, F = G in Parseval’s theorem.

 1 
f [n]f [n] = F [k]F [k]
N
 1 
⇒ | f [n] |2 = | F [k] |2
N

Exercises 24.14.1

1. The solution is developed in the following table.

m 0 1 2 3
f [m] 1 2 2 1
n=0 g[−m] 2 1 1 2 (2)(1) = 2
n=1 g[1 − m] 2 1 1 2 (1)(1) + (2)(2) = 5
n=2 g[2 − m] 2 1 1 2 (1)(1) + (1)(2) + (2)(2) = 7
n=3 g[3 − m] 2 1 1 2 (2)(1) + (1)(2) + (1)(2) + (2)(1) = 8
n=4 g[4 − m] 2 1 1 2 (2)(2) + (1)(2) + (1)(1) = 7
n=5 g[5 − m] 2 1 1 2 (2)(2) + (1)(1) = 5
n=6 g[6 − m] 2 1 1 2 (2)(1) = 2
Hence f ∗ g = 2, 5, 7, 8, 7, 5, 2.
2. (a)
(3 + 9x + 2x2 − x3 )(−4 + 8x + 5x2 ) = −12 +24x + 15x2
−36x +72x2 +45x3
−8x2 +16x3 +10x4
4x3 −8x4 −5x5
= −12 −12x +79x 2
+65x3 +2x4 −5x5

Pearson
c Education Limited 2001
236 Croft,Davison and Hargreaves,Engineering Mathematics,Solutions Manual

Note f ∗ g = −12, −12, 79, 65, 2, −5, the coefficients of this polynomial.
(b)
(9 − 8x)(1 + 2x − 4x2 ) = 9 +18x −36x2
−8x −16x2 +32x3
= 9 +10x −52x2 +32x3
f ∗ g = 9, 10, −52, 32.
3.
(1 − x + x2 + 3x3 )(7 + 2x + x3 ) = 7 +2x + 0x2 +x3
−7x −2x2 +0x3 −x4
7x2 +2x3 +0x4 +x5
21x3 +6x4 +0x5 +3x6
= 7 −5x +5x 2
+24x3 +5x4 +x5 +3x6
Hence f ∗ g = 7, −5, 5, 24, 5, 1, 3.

4.
∞
 
f ∗g = f [m]g[n − m] =
m=−∞ all m

Let n − m = p
  
f ∗g = f [n − p]g[p] = = g[p]f [n − p] = g ∗ f
all p

Hence linear convolution is commutative.

Exercises 24.14.2
1. The solution is developed in the following table.
f [m] 1 −1 1 3
n=0 g[−m] 1 0 2 7 1 0 2 12
n=1 g[1 − m] 1 0 2 7 1 0 −4
n=2 g[2 − m] 1 0 2 7 1 8
n=3 g[3 − m] 1 0 2 7 24
Hence f ∗ g = 12, −4, 8, 24.

Exercises 24.14.3
1.     
1 1 1 1 1 7 4 10
 1 −j −1 j   2   7−j 

F, G =    −1  =  4j 
1 −1 1 −1   1 0   0 4 
1 j −1 −j 3 1 −4j 7+j
F · G = 40, 4 + 28j, 0, 4 − 28j
Also     
1 1 1 1 12 40
 1 −j −1 j   −4   4 + 28j 
  = 
 1 −1 1 −1   8   0 
1 j −1 −j 24 4 − 28j

Pearson
c Education Limited 2001
Croft,Davison and Hargreaves,Engineering Mathematics,Solutions Manual 237

So
D{f ∗ g} = F [k]G[k]
as required.
2.
f = 9, 0, 1
g = 5, 4, 5, 2, 1.
The linear convolution has length 3 + 5 − 1 = 7. Pad to give
f = 9, 0, 1, 0, 0, 0, 0
g = 5, 4, 5, 2, 1, 0, 0.
f [m] 9 0 1 0 0 0 0
n=0 g[−m] 0 0 1 2 5 4 5 0 0 1 2 5 4 45
n=1 g[1 − m] 0 0 1 2 5 4 5 0 0 1 2 5 36
n=2 g[2 − m] 0 0 1 2 5 4 5 0 0 1 2 50
n=3 g[3 − m] 0 0 1 2 5 4 5 0 0 1 22
n=4 g[4 − m] 0 0 1 2 5 4 5 0 0 14
n=5 g[5 − m] 0 0 1 2 5 4 5 0 2
n=6 g[6 − m] 0 0 1 2 5 4 5 1
f ∗ g = 45, 36, 50, 22, 14, 2, 1.

3. not provided

Exercises 24.14.4
1.
f [m] 4 5 9
n=0 g[m] 3 1 1 26
n = −1 g[m + 1] 3 1 1 9
n = −2 g[m + 2] 3 1 1 4
n=1 g[m − 1] 3 1 1 24
n=2 g[m − 2] 3 1 1 27

f  g = 4, 9, 26, 24, 27.

2.
∞
f g = m=−∞ f [m]g[m − n].
Let m − n = p.
∞
f  g = p=−∞ f [p + n]g[p], p is a dummy variable.
∞
Hence, f  g = m=−∞ f [m + n]g[m], as required.

Pearson
c Education Limited 2001
238 Croft,Davison and Hargreaves,Engineering Mathematics,Solutions Manual

3.
m 0 1 2
g[m] 3 1 1
n=0 f [m] 4 5 9
n=1 f [m − 1] 0 4 5 9
n=2 f [m − 2] 0 0 4 5 9
n = −1 f [m + 1] 4 5 9 0
n = −2 f [m + 2] 4 5 9 0 0

Then using the above table


n = 0: f [m]g[m] = 12 + 5 + 9 = 26.

n = 1: f [m − 1]g[m] = 4 + 5 = 9.

n = 2: f [m − 2]g[m] = 4.

n = −1: f [m + 1]g[m] = 15 + 9 = 24.

n = −2: f [m + 2]g[m] = 27.

The required sequence is then 27, 24, 26, 9, 4, obtained by taking the previous results in order.

m 0 1 2
f [m] 3 2 1
n=0 f [m] 3 2 1
4. n=1 f [m − 1] 0 3 2 1
n=2 f [m − 2] 0 0 3 2 1
n = −1 f [m + 1] 3 2 1 0
n = −2 f [m + 2] 3 2 1 0 0

Then f  f = 3, 8, 14, 8, 3.

Exercises 24.14.6
1. f = 2, 3, −1, g = 8, 7, 1.
f [m] 2 3 −1
n=0 g[m] 8 7 1 36
n=1 g[m − 1] 1 8 7 1 19
n=2 g[m − 2] 7 1 8 7 1 9
f
 g = 36, 19, 9.
F = 4, 1 − 3.4641j, 1 + 3.4641j.
G = 16, 4 − 5.1962j, 4 + 5.1962j.
F G = 64, 22 − 8.6603j, 22 + 8.6603j.
D−1 {F G} = 36, 19, 9.
This verifies the circular convolution theorem.

Pearson
c Education Limited 2001
Croft,Davison and Hargreaves,Engineering Mathematics,Solutions Manual 239

2.


N −1 N
 −1
D{f 
 g} = f [m]g[m − n]e−2πjkn/N
n=0 m=0

N −1 
N −1
= f [m] g[m − n]e−2πjkn/N
m=0 n=0
let m − n = p
−1 m−(N −1)

N 
= f [m] g[p]e−2πjk(m−p)/N
m=0 p=m
−1 m−(N −1)

N 
−2πjkm/N
= f [m]e g[p]e2πpjk/N
m=0 p=m

= F [k]G[k]

as required. Note the sum over one period of a periodic sequence (and hence the second term is G[k]), and also


N −1
G[k] = g[n]e−2πjkn/N
n=0

and

N −1
G[k] = g[n]e2πjkn/N
n=0

since g is declared to be real.

3. If

N −1
f
g= f [m − n]g[m]
m=0


N −1 N
 −1
D{f 
 g} = f [m − n]g[m]e−2πjkn/N
n=0 m=0

N −1 
N −1
= g[m] f [m − n]e−2πjkn/N
m=0 n=0
let m − n = p
−1 m−(N −1)

N 
= g[m] f [p]e−2πjk(m−p)/N
m=0 p=m
−1 m−(N −1)

N 
= g[m]e−2πjkm/N f [p]e2πjkp/N
m=0 p=m

= G[k]F [k]

Pearson
c Education Limited 2001
240 Croft,Davison and Hargreaves,Engineering Mathematics,Solutions Manual

Review Exercises 24
1. (a)

 1
F (ω) = (1 − t2 )e−jωt dt
−1

−jωt 1  1
2 e e−jωt
= (1 − t ) − −2t dt
−jω −1 −1 −jω

1  1 −jωt 
2 te−jωt e
= − − dt
jω −jω −1 −1 −jω
 −jωt
1 
2 e−jω + ejω e
= − +
jω −jω ω 2 −1
4 cos ω 4 sin ω
= − +
ω2 ω3

(b)

 π
F (ω) = sin t e−jωt dt
−π
 π
1
= e(j−jω)t − e(−j−jω)t dt
2j −π

π
1 ej(1−ω)t e−j(1+ω)t
= −
2j j(1 − ω) −j(1 + ω) −π
j sin(1 − ω)π sin(1 + ω)π
= − +j
1−ω 1+ω
j sin ωπ sin ωπ
= − −j
1+ω 1−ω
2j sin ωπ
= −
1 − ω2

c)

 τ
F (ω) = e−jωt dt
0
−jωt
τ
e
=
−jω 0
cos ωτ − j sin ωτ − 1
=
−jω
sin ωτ (1 − cos ωτ )
= −j
ω ω

d)

 0  ∞
F (ω) = −e (α−jω)t
dt + e(−α−jω)t dt
∞ 0

Pearson
c Education Limited 2001
Croft,Davison and Hargreaves,Engineering Mathematics,Solutions Manual 241

0 (−α−jω)t
∞
−e(α−jω)t e
= +
α − jω −∞ −α − jω 0
1 1
= − +
α − jω α + jω
−2jω
=
α2 + ω 2

2. (a)
 2
F (ω) = 3te−jωt dt
−2

2 
3te−jωt 3e−jωt
2
= − dt
−jω −2 −2 −jω
−jωt
2
6e−2jω + 6e2jω 3e
= +
−jω ω 2 −2
6
= (2jω cos 2ω − j sin 2ω)
ω2

(b)
 2
F (ω) = 6 e−jωt dt
0

2
e−jωt
= 6
−jω 0
−2jω
(e − 1)
= 6
−jω
6
= (sin 2ω + j cos 2ω − j)
ω

3. (a)
 ∞

F{f (t)} = f  (t)e−jωt dt
−∞
 ∞
 ∞
= e−jωt f (t) −∞ − f (t)(−jω)e−jωt dt
−∞
= jωF (ω)

(b) not supplied

4. Let u = at:

 ∞
F{f (at)} = f (at)e−jωt dt
−∞

1 ∞
= f (u)e−jωu/a du
a −∞
1 ω 
= F
a a

Pearson
c Education Limited 2001
242 Croft,Davison and Hargreaves,Engineering Mathematics,Solutions Manual

5.
(i)  
∞ ∞
F (ω) = f (t)e−jωt dt F (0) = f (t)dt
−∞ −∞

(ii)  
∞ ∞
1 1
f (t) = F (ω)ejωt dω → f (0) = F (ω)dω
2π −∞ 2π −∞

(iii)

 ∞  ∞  ∞
F (ω) = f (t)e−jωt dt = f (t) cos ωt dt − j f (t) sin ωt dt
−∞ −∞ −∞
 
= even function − j odd function
 ∞
= 2 f (t) cos ωt dt if f is even
0

6. not supplied

7.
2α
f (t) = e−α|t| F (ω) =
α2 + ω 2
Then, from the t − −w duality principle,

1 π −α|ω|
F = e
α + t2
2 α

F{cos bt} = π(δ(ω + b) + δ(ω − b))


1 1  π −α|ω| 
F cos bt = e ∗ π(δ(ω + b) + δ(ω − b))
α 2 + t2 2π α
π  −α|ω| 
= e ∗ (δ(ω + b) + δ(ω − b))
2α 
∞
π
= δ(λ + b)e−α|ω−λ| dλ + δ(λ − b)e−α|ω−λ| dλ
2α −∞
π  −α|ω+b| 
= e + e−α|ω−b|
2α

Hence  ∞
cos bt π  −αb  πe−αb
2 2
dt = F (0) = e + e−αb =
−∞ α +t 2α α

Pearson
c Education Limited 2001
Croft,Davison and Hargreaves,Engineering Mathematics,Solutions Manual 243

8.

    
1 1 1 1 3 9
 1 −j −1 j   3   3 
  = 
 1 −1 1 −1   0   −3 
1 j −1 −j 3 3


|f [n]|2 = 9 + 9 + 9 = 27

1  1
|F [k]|2 = (81 + 9 + 9 + 9)
N 4
1
= (108)
4
= 27

9.
(3 − x − 7x2 )(4 + 0x + 12 x2 ) = 12 + 0x + 32 x2
−4x +0x2 − 12 x3
−28x2 +0x3 − 72 x4

12 −4x − 53
2 x
2
− 12 x3 − 72 x4
f ∗ g = 12, −4, − 53
2 , −2, −2.
1 7

10. (a) Consider the two integrals


0
0
−et(−jω)
−e t(−jω)
dt =
−∞ * − jω −∞

and
 ∞
∞
t(−−jω) et(−−jω)
e dt =
0 −* − jω 0

Then
 ∞
0
b
−et(−jω) et(−−jω)
sgn (t)e−jωt dt = lim +
−∞ →0,b→∞ * − jω −b −* − jω 0
1 e−b(−jω) eb(−−jω) 1
= lim − + + −
→0,b→∞ * − jω * − jω −* − jω −* − jω
 
1 1
= lim − +
→0 * − jω * + jω
 
−2jω
= lim
→0 *2 + ω 2
2j
= −
ω
2
=
jω

as required.

Pearson
c Education Limited 2001
244 Croft,Davison and Hargreaves,Engineering Mathematics,Solutions Manual

(b) We know that F{sgn (t)} = 2

jω . From the t–ω duality principle we have

2
2π sgn (−ω) is the Fourier transform of
jt
So  ∞
2 −jωt
e dt = 2πsgn (−ω)
−∞ jt
and by rearranging the constants  ∞
1 −jωt
e dt = jsgn (−ω)
−∞ πt
Also the signum function is odd and hence this last expression can be written −jsgn (ω).
(c)
From the convolution theorem
f (t) ∗ g(t) = F −1 {F (ω)G(ω)}
Take f (t) = 1
πt . Then F (ω) = −jsgn(ω). Take g(t) = cos αt. Then G(ω) = π(δ(ω + α) + δ(ω − α)).
Then

1
∗ cos αt = F −1 {−jsgn(ω)π(δ(ω + α) + δ(ω − α))}
πt

Then note that

sgn(ω)δ(ω − α) = δ(ω − α)
(sketch a graph to see this) and
sgn(ω)δ(ω + α) = −δ(ω + α)
Hence
1
∗ cos αt = F −1 {jπδ(ω + α) − jπδ(ω − α)}
πt
= sin αt

11.


N −1 
N −1
f [n] 
 g[−n] = f [m]g[−(m − n)] = f [m]g[n − m] = f [n] 
∗ g[n]
m=0 m=0

Pearson
c Education Limited 2001