Homework#5

5.23. Let 𝑋(𝑒 𝑗𝜔 ) denote the Fourier transform of the signal 𝑥[𝑛] depicted in Figure P5.23. Perform the
following calculations without explicitly evaluating𝑋(𝑒 𝑗𝜔 ):

(a) Evaluate𝑋(𝑒 𝑗0 ).
∞
𝑗0
𝑋(𝑒 ) = ∑ 𝑥[𝑛] = 6
𝑛=−∞

(b) Find∢ 𝑋(𝑒 𝑗𝜔 ).

Note that 𝑦[𝑛] = 𝑥[𝑛 + 2] is an even signal. Therefore 𝑌(𝑒 𝑗𝜔 ) is real and even signal. This
implies that∢ 𝑌(𝑒 𝑗𝜔 ) = 0. Furthermore, from the time shifting property of the Fourier transform we
have∢ 𝑌(𝑒 𝑗𝜔 ) = 𝑒 𝑗2𝜔 𝑋(𝑒 𝑗𝜔 ). Therefore,∢ 𝑋(𝑒 𝑗𝜔 ) = 𝑒 −𝑗2𝜔 .
𝜋
(c) Evaluate∫−𝜋 𝑋(𝑒 𝑗𝜔 )𝑑𝜔.
1
We have from equation (5.8) 𝑥[𝑛] = ∫ 𝑋(𝑒 𝑗𝜔 )𝑒 𝑗𝜔𝑛 𝑑𝜔
2𝜋 2𝜋
𝜋
2𝜋𝑥[0] = ∫ 𝑋(𝑒 𝑗𝜔 )𝑑𝜔
−𝜋

Therefore,

𝜋
∫ 𝑋(𝑒 𝑗𝜔 )𝑑𝜔 = 4𝜋
−𝜋

(d) Find𝑋(𝑒 𝑗𝜋 ).

We have from equation (5.9)

∞

𝑋(𝑒 𝑗𝜔 ) = ∑ 𝑥[𝑛]𝑒 −𝑗𝜔𝑛

𝑛=−∞
 ∞

𝑋(𝑒 ) = ∑ 𝑥[𝑛](−1)𝑛 = 2
𝑗𝜋

𝑛=−∞

(e) Determine and sketch the signal whose Fourier transform is𝑅𝑒{𝑥(𝜔)}.
𝐹𝑇
From Table 5.1, we have𝐸𝑣{𝑥[𝑛]} ↔ 𝑅𝑒{𝑋(𝑒 𝑗𝜔 )}. Therefore, the desired signal is

𝑥[𝑛] + 𝑥[−𝑛]
𝐸𝑣{𝑥[𝑛]} =
2

(f) Evaluate:

(i)
𝜋
2
∫ |𝑋(𝑒 𝑗𝜔 )| 𝑑𝜔
−𝜋

From Parseval’s theorem we have

∞ ∞
𝑗𝜔 2
∫ |𝑋(𝑒 )| 𝑑𝜔 = 2𝜋 ∑ |𝑥[𝑛]|2 = 28𝜋
−∞ 𝑛=−∞

(ii)
𝜋 2
𝑑𝑋(𝑒 𝑗𝜔 )
∫ | | 𝑑𝜔
−𝜋 𝑑𝜔

Using the differentiation property of the Fourier transform we obtain

𝐹𝑇 𝑑𝑋(𝑒 𝑗𝜔 )
𝑛𝑥[𝑛] ↔ 𝑗
𝑑𝜔
Again using Parseval’s theorem, we obtain
𝜋 2 ∞
𝑑𝑋(𝑒 𝑗𝜔 )
∫ | | 𝑑𝜔 = 2𝜋 ∑ |𝑛|2 |𝑥[𝑛]|2 = 316𝜋
−𝜋 𝑑𝜔
𝑛=−∞
5.24. Determine which, if any, of the following signals have Fourier transforms that satisfy each of the
following conditions:

1. 𝑅𝑒{𝑋(𝑒 𝑗𝜔 )} = 0.

2. 𝐼𝑚{𝑋(𝑒 𝑗𝜔 )} = 0.

3. There exists an integer α such that 𝑒 𝑗𝛼𝜔 𝑋(𝑒 𝑗𝜔 ) is real

4.
𝜋
∫ 𝑋(𝑒 𝑗𝜔 )𝑑𝜔 = 0
−𝜋

5. 𝑋(𝑒 𝑗𝜔 ) periodic

6. 𝑋(𝑒 𝑗0 ) = 0

(a) x[n] as in Figure P5.24(a)

(b) x[n] as in Figure P5.24(b)

1𝑛
(c) 𝑥[𝑛] = 𝑢[𝑛]
2

1|𝑛|
(d) 𝑥[𝑛] = 2

(e) 𝑥[𝑛] = 𝛿[𝑛 − 1] + 𝛿[𝑛 + 2]

(f) 𝑥[𝑛] = 𝛿[𝑛 − 1] + 𝛿[𝑛 + 3]

(g) x[n] as in Figure P5.24(c)

(h) x[n] as in Figure P5.24(d)

(i) 𝑥[𝑛] = 𝛿[𝑛 − 1] + 𝛿[𝑛 + 1]

Solutions:

(1) For 𝑅𝑒{𝑋(𝑒 𝑗𝜔 )} = 0 the signal must be real and odd. Only signals (b) and (i) are real and odd.
(2) For 𝐼𝑚{𝑋(𝑒 𝑗𝜔 )} = 0 the signal must be real and even. Only signals (d) and (h) are real and even.
(3) Assume 𝑌(𝑒 𝑗𝜔 ) = 𝑒 𝑗𝛼𝜔 𝑋(𝑒 𝑗𝜔 ) Using the time shifting property of the Fourier transform we
have 𝑦[𝑛] = 𝑥[𝑛 + α]. If 𝑌(𝑒 𝑗𝜔 ) is real, then 𝑦[𝑛] is real and even (assuming x[n] is real).
Therefore, x[n] is symmetric about α. This is true only for signals (a), (b), (d), (e), (f) and (h).
𝜋
(4) Since 2𝜋𝑥[0] = ∫−𝜋 𝑋(𝑒 𝑗𝜔 )𝑑𝜔 the given condition is satisfied only if x[0]=0. This is true for
signals (b), (e), (f), (h) and (i).
(5) 𝑋(𝑒 𝑗𝜔 ) is always periodic with period 2𝜋. Therefore, all signals satisfy this condition.
(6) 𝑋(𝑒 𝑗0 ) = ∑∞ 𝑛=−∞ 𝑥[𝑛] the given condition is satisfied only if the samples of the signal add up to
zero. This is true for signals (b), (g), and (i).
5.26 Let 𝑥1 [𝑛] be discrete-time signal whose Fourier transform 𝑋1 (𝑒 𝑗𝜔 ) is depicted in Figure P5.26(a)
(a) Consider the signal 𝑥2 [𝑛] with Fourier transform 𝑋2 (𝑒 𝑗𝜔 ), as illustrated in Figure P5.26(b).
Express 𝑥2 [𝑛] in terms of 𝑥1 [𝑛]

Solution:

𝑋2 (𝑒 𝑗𝜔 ) = 𝑅𝑒{𝑋1 (𝑒 𝑗𝜔 )} + 𝑅𝑒{𝑋1 (𝑒 𝑗(𝜔−2𝜋/3) )} + 𝑅𝑒{𝑋1 (𝑒 𝑗(𝜔+2𝜋/3) )}

Therefore
 𝑗2𝜋 −𝑗2𝜋
𝑥2 [𝑛] = 𝐸𝑣{𝑥1 [𝑛]}{1 + 𝑒 3 +𝑒 3 }

5.27 (a) Let x[n] be a discrete-time signal with Fourier transform 𝑋(𝑒 𝑗𝜔 ), which is illustrated in
Figure P5.27. Sketch the Fourier transform of

𝑤[𝑛] = 𝑥[𝑛]𝑝[𝑛]
For each of the following signal p[n]:

(i) 𝑝[𝑛] = cos(𝜋𝑛)

(ii) 𝑝[𝑛] = cos(𝜋𝑛/2)
(iii) 𝑝[𝑛] = sin(𝜋𝑛/2)
(iv) 𝑝[𝑛] = ∑∞𝑘=−∞ 𝛿[𝑛 − 2𝑘]
(v) 𝑝[𝑛] = ∑∞𝑘=−∞ 𝛿[𝑛 − 4𝑘]

(b) Suppose that the signal w[n] of part (a) is applied as the input to an LTI system with unit
sin(𝜋𝑛/2)
sample response ℎ[𝑛] = Determine the output y[n] for each of the choices of p[n] in
𝜋𝑛
part (a).
Solution:

(a)

(b) The Fourier transform of 𝑌(𝑒 𝑗𝜔 ) of y[n] is 𝑋(𝑒 𝑗𝜔 ) = 𝑃(𝑒 𝑗𝜔 )𝐻(𝑒 𝑗𝜔 ). The LTI system with unit
sample response h[n] is an ideal lowpass filter with cutoff frequency 𝜋/2. Therefore, y[n] in each case is:

(i) y[n]=0
sin(𝜋𝑛/2) 1− cos(𝜋𝑛/2)
(ii) 𝑦[𝑛] = −
2𝜋𝑛 𝜋2 𝑛 2
 sin(𝜋𝑛/2) cos(𝜋𝑛/2)
(iii) 𝑦[𝑛] = 𝜋2 𝑛 2
− 2𝜋𝑛

sin(𝜋𝑛/4) 2
(iv) 𝑦[𝑛] = 2 [ 𝜋𝑛
]

1 sin(𝜋𝑛/2) 2
(v) 𝑦[𝑛] = 4 [ 𝜋𝑛
]

5.33 Consider a casual LTI system described by the difference equation

1
𝑦[𝑛] + 𝑦[𝑛 − 1] = 𝑥[𝑛]
2
(a) Determine the frequency response 𝐻(𝑒 𝑗𝜔 ) of this system.
(b) What is the response of the system to the following inputs?
1
(i) 𝑥[𝑛] = (2)𝑛 𝑢[𝑛]
1
(ii) 𝑥[𝑛] = (− )𝑛 𝑢[𝑛]
2
1
(iii) 𝑥[𝑛] = 𝛿[𝑛] + 2 𝛿[𝑛 − 1]
1
(iv) 𝑥[𝑛] = 𝛿[𝑛] − 2 𝛿[𝑛 − 1]
(c) Find the response to the response to the inputs with the following Fourier transforms:
1
1− 𝑒 −𝑗𝜔
(i) 𝑋(𝑒 𝑗𝜔 ) = 4
1
1+ 𝑒 −𝑗𝜔
2
1
1+ 𝑒 −𝑗𝜔
(ii) 𝑋(𝑒 𝑗𝜔 ) = 2
1
1− 𝑒 −𝑗𝜔
4
1
(iii) 𝑋(𝑒 𝑗𝜔 ) = 1 1
(1+ 𝑒 −𝑗𝜔 )(1− 𝑒 −𝑗𝜔 )
2 4
(iv) 𝑋(𝑒 𝑗𝜔 ) = 1 + 2𝑒 −3𝑗𝜔

Solutions:

(a) Taking the Fourier transform of the given difference equation we have
𝑌(𝑒 𝑗𝜔 ) 1
𝐻(𝑒 𝑗𝜔 ) = =
𝑋(𝑒 ) 1 + 1 𝑒 −𝑗𝜔
𝑗𝜔
2
(b) The Fourier transform of the output will be 𝑌(𝑒 𝑗𝜔 ) = 𝑋(𝑒 𝑗𝜔 )𝐻(𝑒 𝑗𝜔 )
1
(i) 𝑋(𝑒 𝑗𝜔 ) = 1
1− 𝑒 −𝑗𝜔
2
1 1
Therefore 𝑌(𝑒 𝑗𝜔 ) = 1 1
1− 𝑒 −𝑗𝜔 1+ 𝑒 −𝑗𝜔
2 2
Taking inverse Fourier transform
1 1 𝑛 1 1 𝑛
𝑦[𝑛] = ( ) 𝑢[𝑛] + (− ) 𝑢[𝑛]
2 2 2 2
2
𝑗𝜔 1
(ii) 𝑌(𝑒 )= ( 1 )
1− 𝑒 −𝑗𝜔
2

Taking inverse Fourier transform

 1 1 𝑛
𝑦[𝑛] = (𝑛 + 1) (− ) 𝑢[𝑛]
2 2
1
(iii) In this case 𝑋(𝑒 𝑗𝜔 ) = 1 + 𝑒 −𝑗𝜔
2
Therefore 𝑌(𝑒 𝑗𝜔 ) = 1 Taking inverse Fourier transform
𝑦[𝑛] = 𝛿[𝑛]
1
(iv) In this case 𝑋(𝑒 𝑗𝜔 ) = 1 − 2 𝑒 −𝑗𝜔
1 1 2
Therefore 𝑌(𝑒 𝑗𝜔 ) = 1 − 2 𝑒 −𝑗𝜔 1 = −1 + 1
1+ 𝑒 −𝑗𝜔 1+ 𝑒 −𝑗𝜔
2 2
Taking inverse Fourier transform
1 𝑛
𝑦[𝑛] = −𝛿[𝑛] + 2 (− ) 𝑢[𝑛]
2
(c)
1 1 −𝑗𝜔
1− 𝑒 −𝑗𝜔 1 1 𝑒
(i) We have 𝑌(𝑒 𝑗𝜔 ) = 4
1 1 = 1 − 4
1 −𝑗𝜔 2
1+ 𝑒 −𝑗𝜔 1+ 𝑒 −𝑗𝜔 (1+ 𝑒 −𝑗𝜔 ) 2 (1+ 𝑒 )
2 2 2 2
Taking the inverse Fourier transform

1 𝑛 1 1 𝑛
𝑦[𝑛] = (𝑛 + 1) (− ) 𝑢[𝑛] − 𝑛 (− ) 𝑢[𝑛 − 1]
2 4 2
1
1+ 𝑒 −𝑗𝜔 1 1
(ii) We have 𝑌(𝑒 𝑗𝜔 ) = 2
1 1 = 1
1− 𝑒 −𝑗𝜔 1+ 𝑒 −𝑗𝜔 1− 𝑒 −𝑗𝜔
4 2 4
Taking the inverse Fourier transform
1 𝑛
𝑦[𝑛] = ( ) 𝑢[𝑛]
4
1 1 2/3 2/9 !/9
(iii) We have 𝑌(𝑒 𝑗𝜔 ) = 1 1 1 −𝑗𝜔 = 1 −𝑗𝜔 2 + 1 + 1
(1+ 𝑒 −𝑗𝜔 )(1− 𝑒 −𝑗𝜔 ) 1+ 𝑒 (1+ 𝑒 ) 1+ 𝑒 −𝑗𝜔 1− 𝑒 −𝑗𝜔
2 4 2 2 2 4
Taking the inverse Fourier transform

2 1 𝑛 2 1 𝑛 1 1 𝑛
𝑦[𝑛] = (𝑛 + 1) (− ) 𝑢[𝑛] + (− ) 𝑢[𝑛] + (− ) 𝑢[𝑛]
3 2 9 2 9 4
1 1 2𝑒 −3𝑗𝜔
(iv) We have 𝑌(𝑒 𝑗𝜔 ) = (1 + 2𝑒 −3𝑗𝜔 ) 1 = 1 + 1
1+ 𝑒 −𝑗𝜔 1+ 𝑒 −𝑗𝜔 1+ 𝑒 −𝑗𝜔
2 2 2
Taking the inverse Fourier transform

1 𝑛 1 𝑛−3
𝑦[𝑛] = (− ) 𝑢[𝑛] + 2 (− ) 𝑢[𝑛 − 3]
2 2
5.35 A casual LTI system is described by the difference equation
𝑦[𝑛] − 𝑎𝑦[𝑛 − 1] = 𝑏𝑥[𝑛] + 𝑥[𝑛 − 1]
Where a is real and less than 1 in magnitude.

(a) Find a value of b such that the frequency response of the system satisfies
|𝐻(𝑒 𝑗𝜔 )| = 1, 𝑓𝑜𝑟 𝑎𝑙𝑙 𝜔
This kind of system is called an all-pass system, as it does not attenuate the input for 𝑒 𝑗𝜔 any
value of 𝜔.
 Solution:
𝑌(𝑒 𝑗𝜔 ) 𝑏 + 𝑒 −𝑗𝜔
𝐻(𝑒 𝑗𝜔 ) = =
𝑋(𝑒 𝑗𝜔 ) 1 − 𝑎𝑒 −𝑗𝜔
𝑗𝜔
In order to |𝐻(𝑒 )| to be one, we must ensure that
|𝑏 + 𝑒 −𝑗𝜔 | = |1 − 𝑎𝑒 −𝑗𝜔 |
1 + 𝑏 2 + 2𝑏𝑐𝑜𝑠𝜔 = 1 + 𝑎2 − 2𝑎𝑐𝑜𝑠𝜔

This is possibly only if b=-a.

5.36 (c) Consider the casual, discrete-time LTI system described by the difference equation
1 1
𝑦[𝑛] + 𝑦[𝑛 − 1] + 𝑦[𝑛 − 2] = 𝑥[𝑛 − 1] − 𝑥[𝑛 − 2]
4 2
What is the inverse of this system? Show that the inverse is not casual. Find another casual LTI system
that is an “inverse with delay” of the system described by above equation. Specifically, find a casual LTI
system such that output w[n] in Figure P5.36 equals x[n-1].

Solution:

The frequency response of given system is

1
𝑌(𝑒 𝑗𝜔
) 𝑒 −𝑗𝜔 − 2 𝑒 −2𝑗𝜔
𝐻(𝑒 𝑗𝜔 ) = =
𝑋(𝑒 𝑗𝜔 ) 1 + 𝑒 −𝑗𝜔 + 1 𝑒 −2𝑗𝜔
4
Frequency response of inverse system is
1
1 1 + 𝑒 𝑗𝜔 + 4 𝑒 −𝑗𝜔
𝐺(𝑒 𝑗𝜔 ) = =
𝐻(𝑒 𝑗𝜔 ) 1
1 − 2 𝑒 −𝑗𝜔

Therefore,

1 𝑛+1 1 𝑛 1 1 𝑛−1
𝑔[𝑛] = ( ) 𝑢[𝑛 + 1] + ( ) 𝑢[𝑛] + ( ) 𝑢[𝑛 − 1]
2 2 4 2
Clearly, g[n] is not a casual impulse response.

If we delay this impulse response by 1 sample, then it becomes casual. Furthermore, the output of the
inverse system will be then x[n-1]. The impulse response of this casual system is
 1 𝑛 1 𝑛−1 1 1 𝑛−2
𝑔1 [𝑛] = 𝑔[𝑛 − 1] = ( ) 𝑢[𝑛] + ( ) 𝑢[𝑛 − 1] + ( ) 𝑢[𝑛 − 2]
2 2 4 2
5.39 Derive the convolution property
𝐹
𝑥[𝑛] ∗ ℎ[𝑛] ↔ 𝑋(𝑒 𝑗𝜔 )𝐻(𝑒 𝑗𝜔 )

Solution:
∞
𝑗𝜔
𝑌(𝑒 ) = ∑ {𝑥[𝑛] ∗ ℎ[𝑛]} 𝑒 −𝑗𝜔𝑛
𝑛=−∞
∞ ∞

= ∑ ∑ {𝑥[𝑘]ℎ[𝑛 − 𝑘]} 𝑒 −𝑗𝜔𝑛

𝑛=−∞ 𝑘=−∞
∞ ∞

= ∑ 𝑥[𝑘] ∑ ℎ[𝑛 − 𝑘]𝑒 −𝑗𝜔𝑛

𝑘=−∞ 𝑛=−∞
∞

= ∑ 𝑥[𝑘]𝑒 −𝑗𝜔𝑘 𝐻(𝑒 𝑗𝜔 )

𝑘=−∞
∞
𝑗𝜔
= 𝐻(𝑒 ) ∑ 𝑥[𝑘]𝑒 −𝑗𝜔𝑘
𝑘=−∞

= 𝐻(𝑒 𝑗𝜔 )𝑋(𝑒 𝑗𝜔 )

5.42 In this problem, we derive the frequency-shift property of the discrete-time Fourier transform as a
special case of the multiplication property. Let x[n] be any discrete-time signal with Fourier transform
𝑋(𝑒 𝑗𝜔 ) and let

𝑔[𝑛] = 𝑒 𝑗𝜔0 𝑛 𝑥[𝑛]

(a) Determine and sketch the Fourier transform of
𝑝[𝑛] = 𝑒 𝑗𝜔0 𝑛
(b) The multiplication property of the Fourier transform tells us that, since
𝑔[𝑛] = 𝑝[𝑛]𝑥[𝑛]
1
𝐺(𝑒 𝑗𝜔 ) = ∫ 𝑋(𝑒 𝑗𝜔 )𝑃(𝑒 𝑗(𝜔−𝜃) )𝑑𝜃
2
<2𝜋>
Evaluate this integral to show that
𝐺(𝑒 𝑗𝜔 ) = 𝑋(𝑒 𝑗(𝜔−𝜔0 ) )

Solutions:

(a) 𝑃(𝑒 𝑗𝜔 ) = 2𝜋𝛿(𝜔 − 𝜔0 ) for |𝜔| < 𝜋 This is as shown in Figure S5.42
 (b) From the multiplication property of the Fourier transform we have
1
𝐺(𝑒 𝑗𝜔 ) = ∫ 𝑋(𝑒 𝑗𝜔 )𝑃(𝑒 𝑗(𝜔−𝜃) )𝑑𝜃
2
<2𝜋>

1
= ∫ 𝑋(𝑒 𝑗𝜔 )2𝜋𝛿(𝜔 − 𝜃 − 𝜔0 ) 𝑑𝜃
2
<2𝜋>
= 𝑋(𝑒 𝑗(𝜔−𝜔0 ) )

5.47 Determine whether each of the following statements is true or false. Justify your answers. In each
statement, the Fourier transform of x[n] is denoted by 𝑋(𝑒 𝑗𝜔 )

(a) If 𝑋(𝑒 𝑗𝜔 ) = 𝑋(𝑒 𝑗(𝜔−1) ) then x[n]=0 for |𝑛| > 0

(b) If 𝑋(𝑒 𝑗𝜔 ) = 𝑋(𝑒 𝑗(𝜔−𝜋) ) then x[n]=0 for |𝑛| > 0

(c) If 𝑋(𝑒 𝑗𝜔 ) = 𝑋(𝑒 𝑗𝜔/2 ) then x[n]=0 for |𝑛| > 0

(d) If 𝑋(𝑒 𝑗𝜔 ) = 𝑋(𝑒 𝑗2𝜔 ) then x[n]=0 for |𝑛| > 0

Solutions:

(a) 𝑋(𝑒 𝑗𝜔 ) = 𝑋(𝑒 𝑗(𝜔−1) ) then 𝑋(𝑒 𝑗𝜔 ) is periodic with a period of 1. But we already know that
𝑋(𝑒 𝑗𝜔 ) is periodic with a period of 2𝜋. This is only possible if 𝑋(𝑒 𝑗𝜔 ) is a constant for all 𝜔. This
implies that x[n] is of the form 𝑘𝛿[𝑛] where k is a constant. Therefore, the given statement is
true.
(b) 𝑋(𝑒 𝑗𝜔 ) = 𝑋(𝑒 𝑗(𝜔−𝜋) ) then 𝑋(𝑒 𝑗𝜔 ) is periodic with a period of 𝜋. We also know that 𝑋(𝑒 𝑗𝜔 ) is
periodic with a period of 2𝜋. Both these conditions can be satisfied even if 𝑋(𝑒 𝑗𝜔 ) has some
arbitrary shape in the region 0 ≤ |𝜔| ≤ 𝜋/2. Therefore, 𝑋(𝑒 𝑗𝜔 ) need not necessarily be a
constant. Consequently, x[n] need not be just impulse. Therefore, given statement is false.
(c) The inverse Fourier transform of 𝑋(𝑒 𝑗𝜔/2 ) is the sequence 𝑣[𝑛] = (𝑥[𝑛] + 𝑒 𝑗𝜋𝑛 𝑥[𝑛])/2. The
even indexed samples of v[n] are identical even indexed samples of x[n]. The odd indexed
samples of v[n] are zeros. If 𝑋(𝑒 𝑗𝜔 ) = 𝑋(𝑒 𝑗𝜔/2 ) then x[n]=v[n]. This implies that the even
indexed samples of x[n] are zero. Consequently, x[n] need not be just impulse. Therefore, given
statement is false.
(d) The inverse Fourier transform of 𝑋(𝑒 𝑗2𝜔 ) is time expanded signal
𝑛
𝑥[ ], 𝑛 = 0, ±2, ±4, …
𝑥2 = { 2
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
 (e) If 𝑋(𝑒 𝑗𝜔 ) = 𝑋(𝑒 𝑗2𝜔 ), then 𝑥[𝑛] = 𝑥2 [𝑛]. This is possible only if x[n] is an impulse. Therefore,
the given statement is true.

5.49 (a) Particular discrete-time system has input x[n] and output y[n]. The Fourier transform of these
signals are related by the equation

𝑑𝑋(𝑒 𝑗𝜔 )
𝑌(𝑒 𝑗𝜔 ) = 2𝑋(𝑒 𝑗𝜔 ) + 𝑒 −𝑗𝜔 𝑋(𝑒 𝑗𝜔 ) −
𝑑𝜔
(I) Is the system linear? Clearly justify your answer.
(II) Is the system time invariant? Clearly justify your answer.
(III) What is y[n] if 𝑥[𝑛] = 𝛿[𝑛]?

Solutions:

(i) Consider the signal 𝑥[𝑛] = 𝑎𝑥1 [𝑛] + 𝑏𝑥2 [𝑛], where a and b are constants. Then 𝑋(𝑒 𝑗𝜔 ) =
𝑎𝑋1 (𝑒 𝑗𝜔 ) + 𝑏𝑋2 (𝑒 𝑗𝜔 ). Also let the response of the system to 𝑥1 [𝑛] and 𝑥2 [𝑛] be 𝑦1 [𝑛] and
𝑦2 [𝑛], respectively. Substituting for 𝑋(𝑒 𝑗𝜔 ) in the equation given in the problem and
simplifying we obtain 𝑌(𝑒 𝑗𝜔 ) = 𝑎𝑌1 (𝑒 𝑗𝜔 ) + 𝑏𝑌2 (𝑒 𝑗𝜔 ). Therefore system is linear.
(ii) Consider the signal 𝑥1 [𝑛] = 𝑥[𝑛 − 1]. Then 𝑋1 (𝑒 𝑗𝜔 ) = 𝑒 −𝑗𝜔 𝑋(𝑒 𝑗𝜔 ). Let the response of the
system to this signal be 𝑦1 [𝑛]. From the given equation
𝑑𝑋1 (𝑒 𝑗𝜔 )
𝑌1 (𝑒 𝑗𝜔 ) = 2𝑋1 (𝑒 𝑗𝜔 ) + 𝑒 −𝑗𝜔 𝑋1 (𝑒 𝑗𝜔 ) −
𝑑𝜔
𝑗𝜔
−𝑗𝜔 𝑗𝜔 −𝑗𝜔 𝑗𝜔
𝑑𝑋 1 (𝑒 )
=𝑒 [2𝑋1 (𝑒 ) + 𝑒 𝑋1 (𝑒 ) − ] + 𝑗𝑒 −𝑗𝜔 𝑋(𝑒 𝑗𝜔 )
𝑑𝜔
≠ 𝑒 −𝑗𝜔 𝑌(𝑒 𝑗𝜔 )
(iii) If 𝑥[𝑛] = 𝛿[𝑛] , 𝑋(𝑒 𝑗𝜔 ) = 1 Then
𝑌(𝑒 𝑗𝜔 ) = 2 + 𝑒 −𝑗𝜔

Therefore, 𝑦[𝑛] = 2𝛿[𝑛] + 𝛿[𝑛 − 1]