DISCRETE FOURIER TRANSFORM

AND SIGNAL SPECTRUM

BAU Dr. Abdul Rahman El Falou

 Discrete Fourier Transform
2

 In time domain,
representation of
digital signals
describes the signal
amplitude versus the
sampling time
instant or the sample
number
 The representation
of the digital signal
in terms of its
frequency
component in a
frequency domain is 1,000-Hz sinusoid with 32 samples
very useful at a sampling rate of 8,000 Hz

Dr. Abdul Rahman El Falou

 Fourier Series Coefficients of Periodic
Digital Signals
3

 we want to estimate the spectrum of a periodic digital

signal x(n)
 sampled at a rate of fs Hz
 fundamental period T0 = N.T where T=1/ fs
 we assume that the periodic digital signal is band limited
such that all harmonic frequencies are less than the folding
frequency fs/2 so that aliasing does not occur

Dr. Abdul Rahman El Falou

 Fourier Series Coefficients of Periodic
Digital Signals
4

 The coefficients of the Fourier series expansion of the

periodic signal x(n) in a complex form are:

 The Fourier series coefficient ck is periodic of N:

Dr. Abdul Rahman El Falou

 Fourier Series Coefficients of Periodic
Digital Signals
5

 Only the line spectral portion between the frequency -fs/2 and
frequency fs/2 (folding frequency) represents frequency information
of the periodic signal

Dr. Abdul Rahman El Falou

 Fourier Series Coefficients of Periodic
Digital Signals
6

 The spectral portion from fs/2 to fs is a copy of the

spectrum in the negative frequency range from -fs/2 to
0 Hz
 For convenience, we compute the spectrum over the
range from 0 to fs Hz with nonnegative indices

 For the kth harmonic, the frequency is

 f0 = 1/T0 is the fundamental frequency in Hz

Dr. Abdul Rahman El Falou

 EXAMPLE
7

Dr. Abdul Rahman El Falou

 Solution
8

Dr. Abdul Rahman El Falou

 Solution
9

Dr. Abdul Rahman El Falou

 Solution
10

Dr. Abdul Rahman El Falou

 Discrete Fourier Transform Formulas
11

 We assume that a periodic signal x(n) is obtained by

copying the acquired N data samples with the duration
of T to itself repetitively.
x(n) ----- DFT -----> X(k)
 X(k) constitutes the DFT coefficients

 Notice that the factor of N is a constant and does not

affect the relative magnitudes of the DFT coefficients
X(k)

Dr. Abdul Rahman El Falou

 Development of DFT formula
12

Dr. Abdul Rahman El Falou

 DFT definition
13

 Given a sequence x(n), 0 ≤ n ≤ N-1

 Where WN is defined as (twiddle factor)

Dr. Abdul Rahman El Falou

 Inverse DFT
14

 The inverse DFT is given by

the expansion is:

Dr. Abdul Rahman El Falou

 Matlab FFT functions
15

Dr. Abdul Rahman El Falou

 Example
16

Dr. Abdul Rahman El Falou

 Solution
17

Dr. Abdul Rahman El Falou

18

Dr. Abdul Rahman El Falou

19

Dr. Abdul Rahman El Falou

 DFT definition
20

 Given a sequence x(n), 0 ≤ n ≤ N-1

 The inverse DFT is given by

Dr. Abdul Rahman El Falou

 Frequency components
21

 The calculated N DFT coefficients X(k) represent the frequency

components ranging from 0 Hz (or radians/second) to fs Hz (or ωs
radians/second)

 Frequency resolution: the frequency step between two consecutive

DFT coefficients to measure how fine the frequency domain
presentation is

Dr. Abdul Rahman El Falou

 Example
22

Dr. Abdul Rahman El Falou

 Solution
23

Dr. Abdul Rahman El Falou

 Amplitude Spectrum and Power
Spectrum
24

 One of the DFT applications is transformation of a finite-length

digital signal x(n) into the spectrum in frequency domain

Dr. Abdul Rahman El Falou

 Amplitude Spectrum and Power
Spectrum
25

 we achieve the digital sequence x(n) by sampling the analog signal

x(t)
 Truncating the sampled signal with a data window with a length T0
= NT
 we get
 we apply the DFT to the obtained sequence

 each calculated DFT coefficient is a complex number, We define the

amplitude spectrum as:

Dr. Abdul Rahman El Falou

 Amplitude Spectrum and Power
Spectrum
26

 We can modify the amplitude spectrum to a one-sided amplitude

spectrum by doubling the amplitudes, keeping the original DC term
at k = 0.

 the phase spectrum is given by

Dr. Abdul Rahman El Falou

 Amplitude Spectrum and Power
Spectrum
27

 The DFT power spectrum is defined as:

 Similarly, for a one-sided power spectrum, we get:

Dr. Abdul Rahman El Falou

 Example
28

 Consider the sequence

Dr. Abdul Rahman El Falou

 Solution
29

Dr. Abdul Rahman El Falou

 Solution
30

Dr. Abdul Rahman El Falou

 Solution
31

 The one-sided amplitude spectrum and one-sided power spectrum

Dr. Abdul Rahman El Falou

 Example
32

Dr. Abdul Rahman El Falou

 Zero padding effect by using FFT
33

Dr. Abdul Rahman El Falou

 Spectral Estimation Using Window
Functions
34

 When we apply DFT to the sampled data, we

theoretically imply the following assumptions:
 the sampled data are periodic to themselves (repeat
themselves)
 the sampled data are continuous to themselves and band
limited to the folding frequency

Dr. Abdul Rahman El Falou

 Spectral Estimation Using Window
Functions
35

Consider the pure 1-Hz sine wave with 32 samples

Dr. Abdul Rahman El Falou

 Spectral Estimation Using Window
Functions
36

 if we use a window size of N = 16 samples, which is a

multiple of the two waveform cycles, the second window
repeats with continuity.

 when the window size is chosen to be 18 samples, which

is not a multiple of the waveform cycles (2.25 cycles), the
second window repeats the first window with
discontinuity

Dr. Abdul Rahman El Falou

 Spectral Estimation Using Window
Functions
37

 The first spectral plot contains a single frequency, while the second
spectrum has the expected frequency component plus many
harmonics: spectral leakage

Dr. Abdul Rahman El Falou

 Spectral Estimation Using Window
Functions
38

 The amount of spectral leakage shown in the second

plot is due to amplitude discontinuity in time domain
 The bigger the discontinuity, the more the leakage.
 To reduce the effect of spectral leakage, a window
function can be used whose amplitude tapers smoothly
and gradually toward zero at both ends
 Applying the window function w(n) to a data sequence
x(n) to obtain a windowed sequence xw(n)

Dr. Abdul Rahman El Falou

 Spectral Estimation Using Window
Functions
39

Dr. Abdul Rahman El Falou

 Common window functions
40

 The common window functions are listed as follows:

Dr. Abdul Rahman El Falou

 Plots of window sequences.
41

Dr. Abdul Rahman El Falou

 Spectral Estimation Using Window
Functions
42

Dr. Abdul Rahman El Falou

 Fast Fourier Transform
43

 The DFT coefficients may be computed via a fast Fourier

transform (FFT) algorithm
 The FFT is a very efficient algorithm for computing
DFT coefficients
 The FFT algorithm requires the time domain sequence
x(n) to have a length of data points equal to a power of
2;
 For example, the number of samples in x(n) can be N =
2, 4, 8, 16, etc.

Dr. Abdul Rahman El Falou

 Fast Fourier Transform
44

 In the case of using the FFT algorithm to compute DFT

coefficients, where the length of the available data is not
equal to a power of 2  we can pad the data sequence
with zeros to create a new sequence
 The modified data sequence for applying FFT, therefore,
is:

Dr. Abdul Rahman El Falou

 Fast Fourier Transform
45

 we consider the digital sequence x(n) consisting of 2m

samples
 If x(n) does not contain 2m samples, then we simply
append it with zeros
 radix-2 FFT algorithms:
 Decimation in-frequency algorithm
 decimation-in-time algorithm

Dr. Abdul Rahman El Falou

 Method of Decimation-in-Frequency
46

 Using the definition of DFT:

 The equation can be expanded as:

Dr. Abdul Rahman El Falou

 Method of Decimation-in-Frequency
47

 we can rewrite as a sum of the following two parts:

 Modifying the second term:

Dr. Abdul Rahman El Falou

 Method of Decimation-in-Frequency
48

 Now letting k = 2m as an even number achieves

 while substituting k = 2m + 1 as an odd number yields

Dr. Abdul Rahman El Falou

 Method of Decimation-in-Frequency
49

 where a(n) and b(n) are introduced and expressed as

 We can summarize by:

Dr. Abdul Rahman El Falou

 Method of Decimation-in-Frequency
50

First
Iteration

Dr. Abdul Rahman El Falou

 Method of Decimation-in-Frequency
51

Second
Iteration

Block
Diagram
of 8-point
FFT

Dr. Abdul Rahman El Falou

 Method of Decimation-in-Frequency
52

 the number of complex multiplications for DFT and FFT,

respectively, are determined by

 In the 8-point FFT, there are 12 complex multiplications

compared with the eight-point DFT with 64 complex
multiplications
 a sequence with 1,024 data points. Applying DFT will
require 1,024 x 1,024=1,048,576 complex multiplications
 applying FFT will need only (1,024/2) x log2 (1,024) = 5,120
complex multiplications
Dr. Abdul Rahman El Falou
 Method of Decimation-in-Frequency
53

 Next, the index (bin number) of the eight-point DFT

coefficient X(k) becomes 0, 4, 2, 6, 1, 5, 3, and 7,
respectively, which are not in the natural order.

Dr. Abdul Rahman El Falou

 Method of Decimation-in-Frequency
54

Dr. Abdul Rahman El Falou

 Example 4.12
55

Dr. Abdul Rahman El Falou

 Solution
56

Dr. Abdul Rahman El Falou

 Example 4.13
57

Dr. Abdul Rahman El Falou

 Method of Decimation-in-Time
58

 In this method, we split the input sequence x(n) into the

even indexed x(2m) and x(2m+1), each with N data
points

Dr. Abdul Rahman El Falou

 Method of Decimation-in-Time
59

 Define new functions as

 Note that

Dr. Abdul Rahman El Falou

 Method of Decimation-in-Time
60

 We can write:

 Considering that:

 the second half of frequency bins can be computed as

follows:

Dr. Abdul Rahman El Falou

 Method of Decimation-in-Time
61

Dr. Abdul Rahman El Falou

 Method of Decimation-in-Time
62

Dr. Abdul Rahman El Falou

 Method of Decimation-in-Time
63

Dr. Abdul Rahman El Falou

 Example 4.14
64

Dr. Abdul Rahman El Falou

 Example 4.15.
65

Dr. Abdul Rahman El Falou