Diskrete Fourier-Transformation (DFT)

Fast Fourier Transformation (FFT)

Windowing
Conclusion

Fast Fourier Transformation

Gerald Stanje Andreas Buchinger

Department of Networked and Embedded Systems

University of Klagenfurt

14.01.2010 / Lecture "Digital Signal Processors"

Gerald Stanje, Andreas Buchinger FFT

 Diskrete Fourier-Transformation (DFT)
Fast Fourier Transformation (FFT)
Windowing
Conclusion

Content I
1 Diskrete Fourier-Transformation (DFT)
Introduction
Sampling Theorem
Properties
2 Fast Fourier Transformation (FFT)
Derivation
Algorithmn
Graphical Representation
Butterfly
Complexity DFT vs. FFT
Applications
3 Windowing
Introduction
Leakage Effect
Gerald Stanje, Andreas Buchinger FFT
 Diskrete Fourier-Transformation (DFT)
Fast Fourier Transformation (FFT)
Windowing
Conclusion

Content II
Window Types

4 Conclusion

Gerald Stanje, Andreas Buchinger FFT

 Diskrete Fourier-Transformation (DFT)
Introduction
Fast Fourier Transformation (FFT)
Sampling Theorem
Windowing
Properties
Conclusion

Introduction (1)

1 Fourier Transformation

is non-periodic, infinite, continuous

2 Sampling

3 Windowing

Gerald Stanje, Andreas Buchinger FFT

 Diskrete Fourier-Transformation (DFT)
Introduction
Fast Fourier Transformation (FFT)
Sampling Theorem
Windowing
Properties
Conclusion

Introduction (2)

4 Evalutation of Xs ω(f) at N equidistant frequency values

5 Discrete Fourier Transformation

is periodic, finite, discrete

6 Inverse Discrete Fourier Transformation

Gerald Stanje, Andreas Buchinger FFT

 Diskrete Fourier-Transformation (DFT)
Introduction
Fast Fourier Transformation (FFT)
Sampling Theorem
Windowing
Properties
Conclusion

Introduction (3)

Gerald Stanje, Andreas Buchinger FFT

 Diskrete Fourier-Transformation (DFT)
Introduction
Fast Fourier Transformation (FFT)
Sampling Theorem
Windowing
Properties
Conclusion

Sampling Theorem
Sampling frequency
fs > 2fmax ... Nyquist Sampling Theorem

Analog Signal

The samples when fs > FM

The samples when fs = 2FM

The samples when fs < 2FM

Avoid Aliasing
Increase the sampling rate
Introduce an anti-aliasing filter
Gerald Stanje, Andreas Buchinger FFT
 Diskrete Fourier-Transformation (DFT)
Introduction
Fast Fourier Transformation (FFT)
Sampling Theorem
Windowing
Properties
Conclusion

Properties (1)
Linearity
α1 x1(n) + α2 x2(n) c sα X 1(k ) + α X 2(k )
1 2

Gerald Stanje, Andreas Buchinger FFT

 Diskrete Fourier-Transformation (DFT)
Introduction
Fast Fourier Transformation (FFT)
Sampling Theorem
Windowing
Properties
Conclusion

Properties (2)

The shift theorem

f (t) c sF (ω)

f (t − τ ) c se−iωτ F (ω)

Gerald Stanje, Andreas Buchinger FFT

 Diskrete Fourier-Transformation (DFT)
Introduction
Fast Fourier Transformation (FFT)
Sampling Theorem
Windowing
Properties
Conclusion

Properties (3)

Symmetry
X [N −k ] = X ∗ [k ] k = 0, 1, ..., N −1
where the symbol ∗ indicates com-
plex conjugate

Gerald Stanje, Andreas Buchinger FFT

 Diskrete Fourier-Transformation (DFT)
Introduction
Fast Fourier Transformation (FFT)
Sampling Theorem
Windowing
Properties
Conclusion

Properties (4)
Convolution

f (t) ∗ g(t) c sF (ω) · G(ω)

Gerald Stanje, Andreas Buchinger FFT

 Diskrete Fourier-Transformation (DFT)
Derivation
Fast Fourier Transformation (FFT)
Algorithmn
Windowing
Graphical Representation
Conclusion

Derivation (1)

Initial point is the DFT

N−1
X
X(k) = x(n)WN nk k=0,1,...,N-1
n=0
k ... discrete frequency variable
n ... discrete time variable
N ... number of samples, number of frequency values
X[k] ... discrete fourier transform at the point f = k ∗ fs /N
x[n]... sample of x(t) at the point t = n ∗ Ts = n/fs
WN ...Twiddlefactor WN = e−j2π/N

Gerald Stanje, Andreas Buchinger FFT

 Diskrete Fourier-Transformation (DFT)
Derivation
Fast Fourier Transformation (FFT)
Algorithmn
Windowing
Graphical Representation
Conclusion

Derivation (2)
Twiddlefactor

(WN )nN = e−j2πn = 1

WN is periodic in n and
k with the period N

Compexity of the DFT

for N frequency points: N 2 multiplications, N(N-1) additions
N should be a power of 2
N = 2q (q ∈ N)
Gerald Stanje, Andreas Buchinger FFT
 Diskrete Fourier-Transformation (DFT)
Derivation
Fast Fourier Transformation (FFT)
Algorithmn
Windowing
Graphical Representation
Conclusion

Algorithmn
Divide the series x[n] into two series
x1 = x[2n]
x2 = x[2n + 1]
We can formulate the DFT sum as
N/2−1 N/2−1
X X
2nk
X[k] = x(2n)WN + x(2n + 1)WN (2n+1)/k
n=0 n=0
N/2−1 N/2−1
X X
X[k] = x(n)WN 2nk + WN k x(n)WN (2nk )
n=0 n=0
simplify: WN 2nk = e(−j2π/N)2nk = e−j2πnk /(N/2) = WN/2 nk
N/2−1 N/2−1
X X
⇒ X [k ] = x1(n)WN/2 nk + WN k x2(n)WN/2 (nk )
n=0 n=0
X[k] = X1[k] + WN k X2[k]
Gerald Stanje, Andreas Buchinger FFT
 Diskrete Fourier-Transformation (DFT)
Derivation
Fast Fourier Transformation (FFT)
Algorithmn
Windowing
Graphical Representation
Conclusion

Graphical Representation
The N point DFT will be splitted into two DFTs with N/2
points
Example 8 point DFT splitted into tow DFTs with 4 points:

Gerald Stanje, Andreas Buchinger FFT

 Diskrete Fourier-Transformation (DFT)
Fast Fourier Transformation (FFT)
Windowing
Conclusion

Literature I

Brad G Osgood,
“The Fourier Transform and its Applications,”
http:
//see.stanford.edu/see/lecturelist.aspx?
coll=84d174c2-d74f-493d-92ae-c3f45c0ee091;
accessed December 23, 2009.
Grünbacher Herbert,
“Die Diskrete Fouriertransformation (DFT),”
http://ti.tuwien.ac.at/rts/teaching/
courses/sigproz/files/DFT_Text.pdf; accessed
December 23, 2009.

Gerald Stanje, Andreas Buchinger FFT

 Diskrete Fourier-Transformation (DFT)
Fast Fourier Transformation (FFT)
Windowing
Conclusion

Literature II

Leonhard Reindl,
“Zeitdiskrete Signale im Frequenzbereich,”
http://electures.informatik.uni-freiburg.
de/portal/download/96/5730/06_signal.pdf;
accessed December 23, 2009.
Merten Joost,
“Diskrete Fourier-Transformation,”
http://www.uni-koblenz.de/~physik/
informatik/DSV/DFT.pdf; accessed December 23,
2009.

Gerald Stanje, Andreas Buchinger FFT

 Diskrete Fourier-Transformation (DFT)
Fast Fourier Transformation (FFT)
Windowing
Conclusion

Literature III

Andreas M. Chwatal,
“Die Schnelle Fouriertransformation,”
http://www.ads.tuwien.ac.at/teaching/ss08/
FortgeschritteneAD/folien/fft_handout.pdf;
accessed December 23, 2009.
M. Wollenhaupt,
“Einführung in die Fouriertransformation,”
http://www.physik.uni-kl.de/fileadmin/
beigang/Vorlesungen/WS_08_09/
Fouriertransformation.pdf; accessed December
23, 2009.

Gerald Stanje, Andreas Buchinger FFT

 Diskrete Fourier-Transformation (DFT)
Fast Fourier Transformation (FFT)
Windowing
Conclusion

Literature IV

James R. Graham,
“Fourier Transforms & the Frequency Domain,”
http://astro.berkeley.edu/~jrg/ngst/fft/
fourier.html; accessed December 23, 2009.
Michael Weiser,
“Fast Fourier Transformation,”
http://www.scribd.com/doc/6817958/
IEFast-Fourier-Transformation; accessed
December 23, 2009.

Gerald Stanje, Andreas Buchinger FFT

 Diskrete Fourier-Transformation (DFT)
Fast Fourier Transformation (FFT)
Windowing
Conclusion

Literature V

Jan Scholz,
“Diskrete FOURIER-Transformation,”
http://www.uni-giessen.de/~gd1186/F-Prak/
node8.html; accessed December 23, 2009.

Gerald Stanje, Andreas Buchinger FFT

 Diskrete Fourier-Transformation (DFT)
Fast Fourier Transformation (FFT)
Windowing
Conclusion

Thank You

Thank you for your attention!

This presentation was created with the LATEX Beamer class:

http://latex-beamer.sourceforge.net

Gerald Stanje, Andreas Buchinger FFT