0% found this document useful (0 votes)
5 views22 pages

Eee-3217-L11 DFT FFT

Uploaded by

h01713717024
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
5 views22 pages

Eee-3217-L11 DFT FFT

Uploaded by

h01713717024
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 22

LECTURE 11: DFT and FFT

Digital Signal Processing I


Course ID: EEE 3217
Course Teacher: Prof. A K M Baki
Ahsanullah University of Science and Technology
Email: [email protected]

© A K M Baki
Discrete Fourier Transform

Following points should be remembered:


 Time domain signal should be sampled fast to avoid
aliasing.
 Sampling duration should be long enough in order to
estimate the frequency correctly.
 Frequency bins should be chosen correctly to avoid
wrong estimate and minimize spectral leakage.
 Increase in sampling rate and sampling duration will
cause computational complexities and cost.

© A K M Baki
Discrete Fourier Transform

Effect of sampling rate and duration:

𝑁 = 𝐹𝑠 × 𝐷
Reference: Chapter 7 (Introduction to Digital Signal Processing, Theory and
Applications Using Matlab by Kathleen Ossman)
© A K M Baki
Discrete Fourier Transform

Effect of sampling rate and duration:


𝑥 𝑡 = sin(2 × 𝜋 × 2500 × 𝑡)

Example 7-2

Sufficient sampling
rate and duration

Reference: Chapter 7 (Introduction to Digital Signal Processing, Theory and


Applications Using Matlab by Kathleen Ossman)
© A K M Baki
Discrete Fourier Transform

Effect of sampling rate and duration:

Example 7-2

Reference: Chapter 7 (Introduction to Digital Signal Processing, Theory and


Applications Using Matlab by Kathleen Ossman)
© A K M Baki
Discrete Fourier Transform

Effect of sampling rate and duration:


𝑥 𝑡 = sin(2 × 𝜋 × 2500 × 𝑡)

Example 7-2

Sufficient sampling
rate and longer
duration

Reference: Chapter 7 (Introduction to Digital Signal Processing, Theory and


Applications Using Matlab by Kathleen Ossman)
© A K M Baki
Discrete Fourier Transform

Effect of sampling rate and duration:


𝑥 𝑡 = sin(2 × 𝜋 × 2500 × 𝑡)

Example 7-2

Sufficient sampling
rate and longer
duration

Reference: Chapter 7 (Introduction to Digital Signal Processing, Theory and


Applications Using Matlab by Kathleen Ossman)
© A K M Baki
Discrete Fourier Transform

Effect of sampling rate and duration:


𝑥 𝑡 = sin(2 × 𝜋 × 2500 × 𝑡)

Example 7-2

Sampling rate too


low: aliasing effect

Reference: Chapter 7 (Introduction to Digital Signal Processing, Theory and


Applications Using Matlab by Kathleen Ossman)
© A K M Baki
Discrete Fourier Transform

Effect of sampling rate and duration:


𝑥 𝑡 = sin(2 × 𝜋 × 2500 × 𝑡)

Example 7-2

No bin at 2.5 KHz


So spectrum
spreads into
closest bin
(leakage)
Reference: Chapter 7 (Introduction to Digital Signal Processing, Theory and
Applications Using Matlab by Kathleen Ossman)
© A K M Baki
Discrete Fourier Transform

Effect of sampling rate and duration:


𝑥 𝑡 = sin(2 × 𝜋 × 2500 × 𝑡)

Example 7-2

No bin at 2.5 KHz


Longer duration
improves
resolution and
reduces leakage
Reference: Chapter 7 (Introduction to Digital Signal Processing, Theory and
Applications Using Matlab by Kathleen Ossman)
© A K M Baki
Discrete Fourier Transform

Effect of sampling rate and duration:

WINDOWING CAN MINIMIZE


SPECTRAL LEAKAGE

© A K M Baki
Discrete Fourier Transform

Frequency spectrum of a rectangular pulse:


Frequency spectrum of a rectangular pulse is a weighted sinc
function. 1
𝑥 𝑡 = 𝑟𝑒𝑐𝑡 𝑡 − ⇔ 𝑋(𝑓) = 𝑒 −𝑗𝜋𝑓 𝑠𝑖𝑛𝑐(𝑓)
2

Example 7-2

© A K M Baki
Discrete Fourier Transform

MATLAB code:

Example 7-2

© A K M Baki
Fast Fourier Transform (FFT)

Expression for DFT:


𝑁−1 𝑁−1
−𝑗2𝜋𝑛𝑘 𝑁 −𝑗2𝜋 𝑁 𝑛𝑘
𝑋 𝑘 = 𝑥(𝑛)𝑒 = 𝑥(𝑛) 𝑒
n=0 n=0

−𝑗2𝜋 𝑁
𝑊𝑁 = 𝑒 → 𝑇𝑤𝑖𝑑𝑑𝑙𝑒 𝐹𝑎𝑐𝑡𝑜𝑟
𝑊𝑁 = cos 2 𝜋 𝑁 − 𝑗 sin 2 𝜋 𝑁

© A K M Baki
Fast Fourier Transform (FFT)

Expression for DFT:


𝑁−1
𝑛𝑘
𝑋 𝑘 = 𝑥(𝑛) 𝑊𝑁
n=0

−𝑗2𝜋 𝑁
𝑊𝑁 = 𝑒 → 𝑇𝑤𝑖𝑑𝑑𝑙𝑒 𝐹𝑎𝑐𝑡𝑜𝑟
𝑊𝑁 = cos 2 𝜋 𝑁 − 𝑗 sin 2 𝜋 𝑁

© A K M Baki
Fast Fourier Transform (FFT)

Expression of DFT in terms of Twiddle


Factor:

Requires:
N samples of input signal, then
N×N complex multiplications
© A K M Baki
Fast Fourier Transform (FFT)

Twiddle Factor can be computed ahead of


time and stored in the DSP-memory:

Steps to be followed:
1. Take N samples of analog signal.
2. Multiply each sample by twiddle factor raised to some
power.
3. Sum up the results.

© A K M Baki
Fast Fourier Transform (FFT)

Twiddle Factor and it’s properties:

𝑊𝑁 = cos 2 𝜋 𝑁 − 𝑗 sin 2 𝜋 𝑁

Reference: Chapter 7 (Introduction to Digital Signal Processing, Theory and


Applications Using Matlab by Kathleen Ossman)
© A K M Baki
Fast Fourier Transform (FFT)

Twiddle Factor and it’s properties:


Key symmetric properties-

𝑊𝑁 = cos 2 𝜋 𝑁 − 𝑗 sin 2 𝜋 𝑁

Reference: Chapter 7 (Introduction to Digital Signal Processing, Theory and


Applications Using Matlab by Kathleen Ossman)
© A K M Baki
Fast Fourier Transform (FFT)

Twiddle Factor and it’s properties:


Key symmetric properties-

Reference: Chapter 7 (Introduction to Digital Signal Processing, Theory and


Applications Using Matlab by Kathleen Ossman)
© A K M Baki
Fast Fourier Transform (FFT)

How FFT works-


o The FFT decomposes an N point time domain signal into N time
domain signals each composed of a single point.

o The second step is to calculate the N frequency spectra


corresponding to these N time domain signals.

o Lastly, the N spectra are synthesized into a single frequency


spectrum.

Reference: Chapter 7 (Introduction to Digital Signal Processing, Theory and


Applications Using Matlab by Kathleen Ossman)
© A K M Baki
Recommended Readings

Required
 S. W. Smith, The Scientist and Engineer’s Guide to Digital Signal Processing, 2nd
ed., San Diego, Calif.: California Technical Pub.,1999.
 B. P. Lathi and R. Green, Essentials of Digital Signal Processing, New York N.Y.
Cambridge University Press, 2014
 K. A. K. Ossman, Introduction to Digital Signal Processing Theory and Applications
Using MATLAB.

Recommended
 J.G Proakis and D.G Manolakis, Digital Signal Processing, Principles, Algorithm,
and Applications, 3rd ed., India: Prentice-Hall, 2000
 Jose Maria Giron Sierra, Signal Processing with Matlab Examples, Volume 1,
Springer-2018

You might also like