Scribd
Upload
38 views

Week 14

Uploaded by

gghous611
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
38 views

Week 14

Uploaded by

gghous611
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 34

DIGITAL SIGNAL PROCESSING

Prof. Dr. Jameel Ahmed


Week 14
Faculty of Engineering & Applied Sciences Riphah International University
Content
01 Introduction

02 DFT and IDFT

03 FFT
Introduction

3
The Discrete Fourier

Transform (DFT)
The Discrete Fourier
Transform (DFT)
• The DTFT yields a continuous and periodic spectra of a
aperiodic signal.
• So here the problem arises, when we want to process, estimate,
plot or compute this continuous signal.

𝐿−1

𝑋 𝜔 = ෍𝑥 𝑛 𝑒 −𝑗𝜔𝑛 0 ≤ 𝜔 ≤ 2𝜋
𝑛=0
The Discrete Fourier Transform
(DFT)
• When x(𝜔) is sampled at equally spaced frequencies. The resultant
samples are:
𝑁−1

𝑋 𝑘 = ෍𝑥 𝑛 𝑒 −𝑗2𝜋𝑘𝑛/𝑁 𝑘 = 0,1,2, … 𝑁 − 1
𝑛=0

𝑁−1
1 𝑗2𝜋𝑘𝑛/𝑁 𝑛 = 0,1,2, … 𝑁 − 1
𝑥 𝑛 = ෍𝑋 𝑘 𝑒
𝑁
𝑘=0

DFT is Sampling in Frequency Domain


The Discrete Fourier
Transform (DFT)
• DFT and IDFT expresses a finite-length signal as one period of a
periodic sequence.
• DFT assumes aperiodic signal 𝑥(𝑛) to be periodic for the purpose
of computation.
• The two equations differs by the sealing factor 1/N and a sign
change in exponent. This leads to almost a same algorithm for the
calculation of both equations.
The Discrete Fourier
Transform (DFT)
• DFT equation can be treated as DTFS as well. If the signal 𝑥(𝑛) is
periodic, the equation exhibits as Discrete-time Fourier series, but if
x(n) is aperiodic, the equation forms as DFT.
𝑁−1

𝑋 𝑘 = ෍𝑥 𝑛 𝑒 −𝑗2𝜋𝑘𝑛/𝑁
𝑘 = 0,1,2, … 𝑁 − 1
𝑛=0
The Discrete Fourier
Transform (DFT)
Discrete Fourier Transform
(DFT)
Example
• Find the 4-point DFT of the sequence 𝑥(𝑛) = cos 𝑛𝜋/4.
Solution:
Since N=4 𝑛 = {0,1,2,3}

𝑥(𝑛) = 1, 0.707, 0, −0.707


DFT will be: 3
−𝑗2𝜋𝑘𝑛/𝑁
𝑋 𝑘 = ෍𝑥 𝑛 𝑒
𝑛=0
Discrete Fourier Transform
(DFT)
• For k=0 3

𝑋 0 = ෍𝑥 𝑛 =1
𝑛=0
• For k=1
𝑋 1 = 1 − 𝑗1.414

𝑋 𝑘 = {1, 1 − 𝑗1.414, 1,1 + 𝑗1.414}


Inverse Discrete Fourier
Transform (IDFT)
Example
• Find the IDFT of 𝑋 𝑘 = {3, 2 + 𝑗 , 1, (2 − 𝑗)}
Solution:
Since N=4 k = {0,1,2,3}
IDFT will be: 3
1 𝑗2𝜋𝑘𝑛/𝑁
𝑥 𝑛 = ෍𝑋 𝑘 𝑒
4
𝑘=0
Discrete Fourier Transform
(DFT)
• For n=0 3

𝑥 0 = ෍𝑋 𝑘
𝑛=0
1
𝑥 0 = {3 + 2 + 𝑗 + 1 + (2 − 𝑗)}
4
𝑥 0 =2
• For n=1
𝑥 1 =0

𝑥 𝑛 = {2,0,0,1}
Fast Fourier Transform

(FFT)
Computational Complexity of
DFT
• A major practical consideration when computing the
DFT is its speed.
• The most obvious approach to compute the DFT & IDFT
is to implement equations directly, but this way is very
slow.
Computational Complexity of
DFT
• N Point DFT Requires: 𝑁−1

𝑋 𝑘 = ෍𝑥 𝑛 𝑒 −𝑗2𝜋𝑘𝑛/𝑁
• 𝑁 complex multiplies
2
𝑛=0
• 𝑁(𝑁 − 1) complex additions
DFT Decomposition

• DFT Decomposition:
i. Decimation in time
ii. Decimation in frequency
The Discrete Fourier Transform
(DFT)
• The DFT and IDFT can also be expressed in terms of 𝑊𝑁 as:
𝑁−1

𝑋 𝑘 = ෍ 𝑥 𝑛 𝝎𝑘𝑛
𝑁 0≤𝑘 ≤𝑁−1
𝑛=0

𝑁−1
1 −𝑘𝑛
𝑥 𝑛 = ෍ 𝑋 𝑘 𝝎𝑁 0≤𝑛 ≤𝑁−1
𝑁
𝑘=0

Where 𝜔𝑁 = 𝑒 −𝑗2𝜋/𝑁
Decimation in time

𝑁−1

𝑋 𝑘 = ෍ 𝑥 𝑛 𝝎𝑘𝑛
𝑁
𝑛=0
• Put n=2r and 2r+1
𝑁/2−1 𝑁/2−1
2𝑟𝑘 (2𝑟+1)𝑘
𝑋 𝑘 = ෍ 𝑥 2𝑟 𝝎𝑁 + ෍ 𝑥 2𝑟 + 1 𝝎𝑁
𝑛=0 𝑛=0
𝑁/2−1 𝑁/2−1
2 𝑟𝑘 2𝑟𝑘 𝑘
𝑋 𝑘 = ෍ 𝑥 2𝑟 (𝝎𝑁 ) + ෍ 𝑥 2𝑟 + 1 𝝎𝑁 𝝎𝑁
𝑛=0 𝑛=0
Decimation in time

𝑁/2−1 𝑁/2−1

𝑋 𝑘 = ෍ 𝑥 2𝑟 (𝝎2𝑁 )𝑟𝑘 + 𝝎𝑘𝑁 ෍ 𝑥 2𝑟 + 1 𝝎2𝑟𝑘


𝑁
𝑛=0 𝑛=0
𝑗2π
2 −2∗
• Since 𝝎𝑁 =𝑒 𝑁 = 𝝎𝑁/2
𝑁/2−1 𝑁/2−1
𝑟𝑘 𝑘 𝑟𝑘
𝑋 𝑘 = ෍ 𝑥 2𝑟 𝝎𝑁/2 + 𝝎𝑁 ෍ 𝑥 2𝑟 + 1 𝝎𝑁/2
𝑛=0 𝑛=0

𝑋 𝑘 = 𝐺 𝑘 + 𝝎𝑘𝑁 𝐻(𝑘)
Decimation in time

X [k ] = G
[ k ] +W 
k
N H
 [ k ]
N/2 DFT N/2 DFT
of even samples of odd samples

• Then repeat decomposition of N/2 to N/4 DFTs, etc.


Recursive Divide & Conquer
Recursive Divide & Conquer
Bit Reversal in DIT
Detail of “Butterfly”

• Cross feed of G[k] and H[k] in flow diagram is called a “butterfly”,


due to shape
Four Point DIT FFT Graph
Eight Point DIT FFT Graph
Four Point DIF FFT Graph
The DIF is the transpose of
the DIT
DIT FFT structure DIF FFT structure
Examples
Decimation in time
Decimation in Frequency
Decimation in time

• FFT is a technique to solve DFT efficiently.


Example:
Let a signal 𝑥(𝑛) = 1.5𝑛 is defined for N=8. Then compute FFT using decimation
in time
Solution:
THE END!

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy