Tutorial on DFS DTFT and DFT ECC3852 Digital Signal Processing

1. Consider the periodic sequence 𝑥 𝑛 defined as follows:

1, 0 𝑛 4
𝑥𝑛
0, 4 𝑛 8

The sequence 𝑥 𝑛 is periodic with a period 𝑁 8

(a) Find the Discrete Fourier Series (DFS) coefficients 𝑥 𝑘 for 𝑘 1,2, . . ,7.

(b) Write the explicit expression for the periodic sequence 𝑥 𝑛 in terms of its DFS coefficients.

(c) Plot the magnitude and phase of the DFS coefficients 𝑥 𝑘 for 𝑘 1,2, . . ,7.

2. Consider the discrete-time signal 𝑥 𝑛 defined as:

𝑗𝜋𝑛
𝑥𝑛 𝑛 2 exp 𝑢𝑛 1
4

where 𝑢 𝑛 is the unit step function. Find the DTFT X 𝜔 of the signal 𝑥 𝑛 .

3. By means of DFT and IDFT, determine the response of FIR filter with impulse response

ℎ𝑛 1, 2,3
to the input sequence
𝑥𝑛 1,2, 2, 1

4. Assume we want to compute an N-point FFT of an 𝑥 𝑛 audio signal from a compact disc (CD), with
the FFT’s output frequency-domain sample spacing no greater than 1 Hz. If 𝑥 𝑛 ’s sample rate is 𝑓 =
44.1 kHz, what is the number of necessary time samples, N, applied to the FFT?

Given two sequences

5.
𝑥 2,1,2,1
𝑥 1,2,3,4

(a) Find 𝑥 which is a circular convolution of the two sequences.

(b) By means of DFT and IDFT, determine the sequence 𝑥 .
QI

[O
x[n) =

period N = 8

a) Find the Discrete Fourier Series Th

& = cost + join It
]2
= x[K)
=

↓ [K) e

* [n) = %Y1X2yY4X5
1 = 0 : xto) = [1 eT210 1 20 . + . + 1 le+ 0 + ...
)
=
1 + 1 + 1 + 1 + 0

= 4

1 = 1
,
XID = 1 go .
+ 1 .
e **C(1) JEXjs + 0 .
eTI
...
+

=
1+ e + e
+
+ ej + 0 + 0 + 0 + 0

I -j +
0-j--j
=
1 -

j() + (2)
j27(2)(3)
j2x(2( e-j22(12
-

[1 8
]
*
K = 2
x(2) =
.
e # 1 .
e + 1 . + 10e + 0 + ...

5 j35
eT
=
-

=
[1 + + e + e +

o]
= 1 + 0 -

j -

1 +
j8 +0 +
j
k =
3 + x(3] =
71e + 1 .
e
-

g24(3)1) + 1 .
e-Jex(l()-JxCl + 0 + ...
)
53 j
eT3
-
-

=
1 + + e +e + 0

=
1 +
j()
-
j)
- 25(4)(2) 825(4)(3)
-T25(4)()
-

k = 4
+
x(4) =

[1 .
e + 1
.

e 1 .
e
8
+ 1 -
2
g + 0 + ...
]
-

5
-

82x -

T35)
=
1 + e + e + 2

=
1 -

1 +
j0 + 1 +
y0
-

1 +7)

=
O
 j25(5)(3)
zisID(1)-CTM)
-

)
=

k = 5
+
x[5) =
[1 + 1 .
e 102 # + 1 .
e
G
+ Ot ...

o -j5
-5 -
=
1 + e + e + e

1 E +
j +
= -

=
1 +
j)
-

1 + jz)
j2T(6)(2) j25(b)(3)
-10
-

]
-

k 6 =

,
x[6] =

[l + le + 1 e
.
g
+ 0 + ...

839 -j9
e-
-

-
1 + + e +

2
1 + 0 T + j0 + 0 j
=
+
-

+
= O

j25(7)(1)-JI-52Cl )
-

k = 7
,
x(7) =
1 + 1 .

e 8 + 0 +...

jE-TE
-

= 1 + e

1# +
5 + 0 +
j -j
=

=
1 + j(1 + Jz)

~ Ek] = [4 ( -

j(( +() 0 ( +
T (1-2) 0 ( + H +)0 ItiCIIt]
b) Write the explicit expression for periodic sequence in]

x[n] = (keT27/N
=> 4 [ ill 5)]Te]+ + + + 0 + [Hill-5)][en] + 0 +

[1 + 5 - + /z][ei2ny + 0 + [ + j(1 + ][e]

x [n] = 4 + [1 j( + J)]eT
-

+ [1 +
j() - 5)]e5 + [1 + j( -
1 + ()]e + 5)]e
[1 +j(1 +

a) Plot the magnitude and Phase of DFS

magnitude
&

4 *

3
-
X

2 -
X

1 -

X X

X(X)
* 13
01234567
Q2 .
Discrete-time signal xiny =
(n-2(Juin-1)
where uin] is the unit step function .
Find DAT X [W]

x (w)
= In]e-inw
= (n-2)(e * (UIn-17 - e-inc
Ne-co

-
new
= ne -
5(4 -

w]n

let = ei -w

= Ens
(
= I So 1 + (1 s)
-

+
=

1 -
S 1-S
(1
-
S

1 -
S
n
( 2)i = s)
= -

w) 5(G 2)
ei(
-
-

[reT -jate
=
2

[l-eit-w)]2
Q3 .

M 4 N = 3
=

c = m+ n -
1

I 4 + 3 -1

= 6

x(1))
= (n)
= H+ 2 - Yk/ + 2- + e -j3Nk14k = 0
,
1, 0007

+ -j(4 +35
~ (a) = 6 x(1) = 2 5
,

x(3)
-5 j(4 -35)
x(2) =
-

1 -j =
2
, +

x(x) 0 x(5) (
j(4 -352)
= = 2
=
-

x(6)
j(4 3)
( +j x(z) 2 5
+
= =
+
-

+
,

8 point of DFT

H(k) = h(n)ejk
-Fik/4 -J5k/2
- 1 + 2e + 3e

H(0) =
G

H(l) = 1+ Je -

j(3 + (2)

H(2) =
-
2 -

j2

H(3) = 1 -
(2 +
j(3 j) -

H(4) = 2

H(5) (z)
=
1 5
j(3
- -
-

H(6) = -
2+
j2
H(z) = 1 + ( + j(3 + Jz)

Y(0) =
36 y(2) =
j4 u(4) =
g y(6) = -

54

Y() 50 515495
14
%17 484(3) 0 07 0 07
58 515y(7) 48
=
07 14 07 +
517
= =
+
-
-
- .

=
. .
. . -
. .
.
Q3 .

122(
I 1221

22442

33663

[1 + 4 + 9 +
11 +
8 3) +

y [n] (9
=

,
z
+
9 +
11]

Y[K] = y[ne- =
4

= yn) e-
o -
JE -

j57k -

73x
=
x[0]e + y(1)e + y(2]e + y[3]e

7 3
=
9 + 7 + 99
- Mk
+ /y

0
y [8] =
9 #720 + 9-0 + 11e =
36

75 geim + =
=

y (1)
=
9 + + 11e =
9 -

j7 -

9 + j1) =
j4

53x
↑ (2) = 9 + 7e
-
+ get2n + 11e = 9 - 7 + 9 -

11 = 0

5
y [33 =
9 + 7e + ge 3x +
11e
-

=
q + j7
-

q -

j1)
=
-

j4

Q4
fs = 44 /HE
.

N = As = 44100
IItz I

N = 44108

Since 44100 is not the integer power of two , so the next N will be larger integer of two

N =
65536 samples
 .
Q5 X1 =
[212 1]

x2 = [1234]

2 12

12 12/

2424 2

363 6 3

48 4 84

[2
=

=
5
+
10
+
16
+,
12
+
11
+ 4)
k
+
11
+
4 4
Xs[n] =
[14 +
16 +
14 +
16]

b) x[k] = s[n]e-jinkN =
4

= s In) e-T
Jak
Xs[2]e-54 Xs[3]em
-

&
= Xs [0)e= + Xs[]e* + +

53
16e 5 142-544 16e
=

= 14 + + +

0 0
8) [0)
=
14 + 16 + 14 + 162 =
14 + 16 + 14 + 16 = 60

xs[1] =
14 + 16 + Lin + 16 J =
14
-j16 -
14 +
j66 = 0

-%
149* + 162
539
-

x3[2] =
14 + 162 + = 14 -

16 + 14 -
16 = -

* *
& [3] = 14 + 16 + 14 + 165 = 14 + 516 -

14 -

j6 =
0