You are on page 1of 6

# THE DISCRETE FOURIER TRANSFORM

Solution 9.1
(a) X(k) =
=

N-1
x(n) WNkn
n=0

RN(k)

=RN (k)

X (k) (c) X(k)

= WNkn0 RN (k) N-1
= n=0 1-aN W kN

an WNkn

RN(k)

1 Nk -a .- a WN

RN (k )

1-a WN)

Solution 9.2

x( (n)) 4

0

T?

x((n-2)) 4

0 x((n-2)) 4 R4 (n)

0

x( ) (-n)
x((-n)) R (n)

0

0
Figure S9.2-1
S9.1

Solution 9.3
Since X1(0)
=

X -

(k) X

=

-

( (N

-

k) )N RN (k)
=
-

((N))N (0) = 0

RN(k)

X

(0)

Therefore X

Also, for N even,

--

))N

RN(k)

NN

=e -

X

(

)

Therefore X()

=

0.

Solution 9.4

x (M) 2
x2

()N

9

*1 Figure 59.4-1 Solution 9.5

7

1

1

x 1(M)

0

1

2

3

4

5 x 2 ( (-M))
6

R 6 (M)

0
Ix 2 ( (2-m)M

m

0

m

x 1(n)

x 2 (n)

Figure S9.5-1 S9.2

Note that this corresponds to x1 (n) circularly shifted to the right by
two points.
Solution 9.6
We wish to compute X1 (k) given by

9
X1 (k) =a n=0 x(n) zk-n R1 0 (k) j27k ­ where zk = 0.5 e 10 ej0 so that

9

j2Trk

j-n 1
R10 (k)

X (k) =
n=0
9

x (n)

e

10 e 10
-n

jn Tk

x(n)

e 10

e

10

R1 0 (k)

Thus X (k) is the 10-point DFT of the sequence

x (n)

=

x(n)[j e 10

Solution 9.7
In all of the following equations the DFT computed is valid only in the range O<k<N-1 and is zero outside that range. RN (k). N-1
G1 (k) = n=0
N-1
= x(m) WNk(N-1-m)

This permits us to keep

the equations somewhat cleaner by suppressing the use of the function

x(N -1-n)

WN

m=0
N-1 =

j27k j2rr x(m) e N e N -j27r k X(e

m=0 27rk
= ej N

)

= H7 (k)

N-1 G2 (k) = (- )n x(n) WNkn 1
=

N-1 x(n) WN n=0

N
2

kn WN

S9.3

N-1

= x(n)

-j2 e N

(k+ N)n 2=

X (e

j

N

(k+

2

)

n=0

= H8 (k)

N-1 G 3 (k) Z n= 0 N-1
x (n)

2N-l
W2Nnk + Zx(n N)

W 2nk

n=N
x (n)

=
n=0
N-1
= n=0

w7Nnk + W 2 N (n+N) k 2

x (n)

W2 N

[1 + W2 NNk

j2
1+(-1) ]X (e

r
k kN = H 3 (k)

N
-1
2
+ x(n + N ))WNnk
2 k 4 = x (n) N f-
n=0
x (n)

N- 1 WNk
2

=

+ N
n=f

x (n)

WN(n- N) k wN 2

n=0 N-i
G = x (n) n=0 2N-1

G 5 (k)
=

nk
= X(e r

N

=

H 6 (k)

x (n)

W2Nnk =X(e

N

=

H 2 (k)

n= 0 N-i G 6 (k)
=

j 2rR
w2nk 2N

X X(e

a x (n)

N )

=

Hi(k)

n= 0
N_

G 7 (k)
=

x (2n)

WNk
f

n= 0 N-1

x2 E
S9.4

x()1+ (-1)n

nk/2

n=0
N-1 nk + W n(k+N/2) 2 n= x (n)WWN NI

=

j sk X (e3 2

j
+ Xe

2x

(k+N/

(k+N/2))]

(k

H5 (k)

All of the above properties can alternatively be obtained from the basic
DFT properties of sections 8.7 and 8.8, or the z-transform properties of
section 4.4. Many of the properties used in this problem have important
practical applications. g5 (n), for example, corresponds to augmenting a
finite length sequence with zeros so that a computation of the DFT for
this augmented sequence provides finer spectral sampling of the Fourier
transform.

S9.5

MIT OpenCourseWare http://ocw.mit.edu

Resource: Digital Signal Processing
Prof. 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.