Fast Fourier Transform (FFT) (Theory and Implementation)

Chapter 19
Fast Fourier Transform (FFT)

(Theory and Implementation)
Learning Objectives
Chapter 19, Slide 2
DFT algorithm.
Conversion of DFT to FFT algorithm.
Implementation of the FFT algorithm.
Dr. Naim Dahnoun, Bristol University, (c) Texas Instruments 2004
DFT Algorithm
The Fourier transform of an analogue

signal x(t) is given by:
X
jt
x
t
e
dt
The Discrete Fourier Transform (DFT) of

a discrete-time signal x(nT) is given by:
N 1
X k xn e
2
nk
N
n 0
Chapter 19, Slide 3
Where:
k 0,1, N 1
xnT xn
DFT Algorithm
If we let:
2
N
X k
then:
WN
N 1
xnW
nk
N
n 0
Sampled signal
2
Amplitude
1
0
-1
-2
20
40
60
80
Sample
Frequency Domain
100
120
Magnitude
0.8
0.6
0.4
0.2
0
Chapter 19, Slide 4
0.1
0.2
0.3
Normalised Frequency
0.4
0.5
DFT Algorithm
X k
N 1
x[n] = input
xnWNnk
n 0
X[k] = frequency bins

W
= twiddle factors
X(0)
= x[0]WN0 + x[1]WN0*1 ++ x[N-1]WN0*(N-1)
X(1)
= x[0]WN0 + x[1]WN1*1 ++ x[N-1]WN1*(N-1)

:
X(k)
= x[0]WN0 + x[1]WNk*1 ++ x[N-1]WNk*(N-1)

:
X(N-1) = x[0]WN0 + x[1]WN (N-1)*1 ++ x[N-1]WN (N-1)(N-1)
Note: For N samples of x we have N frequencies

representing the signal.
Chapter 19, Slide 5
Performance of the DFT Algorithm
The DFT requires N2 (NxN) complex

multiplications:
The DFT also requires (N-1)*N complex

additions:
Chapter 19, Slide 6
Each X(k) requires N complex

multiplications.
Therefore to evaluate all the values of the
DFT ( X(0) to X(N-1) ) N2 multiplications are
required.
Each X(k) requires N-1 additions.

Therefore to evaluate all the values of the
DFT (N-1)*N additions are required.
120
120
Number of Additions
Number of Multiplications
Performance of the DFT Algorithm
100
80
60
40
20
100
80
60
40
20
0
0
0
Number of Samples
Chapter 19, Slide 7
10
10
Number of Samples
Can the number of computations required

be reduced?
DFT FFT
Chapter 19, Slide 8
A large amount of work has been devoted

to reducing the computation time of a
DFT.
This has led to efficient algorithms which
are known as the Fast Fourier Transform
(FFT) algorithms.
DFT FFT
X k
N 1
xnWNnk ; 0 k N 1
n 0
[1]
x[n] = x[0], x[1], , x[N-1]
Lets divide the sequence x[n] into even

and odd sequences:
Chapter 19, Slide 9
x[2n]
= x[0], x[2], , x[N-2]
x[2n+1] = x[1], x[3], , x[N-1]
DFT FFT
Equation 1 can be rewritten as:

X k
N
1
2
x2nW
2 nk
N
n 0
N
1
2
x2n 1W
2 n 1k
N
[2]
n 0
Since:
WN2 nk e
2
2 nk
N
2
nk
N 2
WN2 n 1k WNk W Nnk
W Nnk
Then:
X k
N
1
2
n 0
x2nW Nnk WNk

2
n 0
Y k WNk Z k
Chapter 19, Slide 10
N
1
2
x2n 1W Nnk
2
DFT FFT
The result is that an N-point DFT can be

divided into two N/2 point DFTs:
X k
N 1
xnW
nk
N
; 0 k N 1
N-point DFT
n 0
Where Y(k) and Z(k) are the two N/2

point DFTs operating on even and odd
samples respectively:
X k
N
1
2
n 0
x1 nW Nnk WNk
2
Y k WNk Z k
N
1
2
n 0
x2 nW Nnk
2
Two N/2point DFTs
DFT FFT
Periodicity and symmetry of W can be

exploited to simplify the DFT further:
X k
N
1
2
x1 nW Nnk W Nk
n 0
N
1
2
x2 nW Nnk
n 0
[3]
X k
2
Or:
And:
WN
N
2
N
n k
x1 n W N 2
n 0
2
WN
2
N
1
2
N
2
2
k
N
2 N
N 2
2
k
N 2
e
j
N
1
2
N
k
WN 2
x2
n 0
2
k
N
2 N
N 2 2
e
2
k
N 2
N
n k
n W N 2
2
2
k
N
W Nk
WNk
: Symmetry
: Periodicity
DFT FFT
Symmetry and periodicity:
W86
W8 5
W87
W84
W8
W80 = W88
3
W8
1=
W8
WNk+N/2 = -WNk
WN/2k+N/2 = WN/2k
W8k+4 = -W8k
W8k+8 = W8k
W82
DFT FFT
Finally by exploiting the symmetry and

periodicity, Equation 3 can be written as:
N
1
2
N
1
2
X k x1 nW Nnk WNk x2 nW Nnk

2 n 0
n0
2
2
[4]
Y k WNk Z k
DFT FFT
N
k 0, 1
2
N
k
X k Y k WN Z k ; k 0, 1
2
X k Y k WNk Z k ;
Y(k) and WNk Z(k) only need to be

calculated once and used for both
equations.
Note: the calculation is reduced from 0 to
N-1 to 0 to (N/2 - 1).
DFT FFT
N
k 0, 1
2
N
k
X k Y k WN Z k ; k 0, 1
2
X k Y k WNk Z k ;
Y(k) and Z(k) can also be divided into N/4

point DFTs using the same process shown
above:
Y k U k W Nk V k
2
Y k U k W Nk V k
4
Z k Pk W Nk Qk
2
Z k Pk W Nk Qk
4
The process continues until we reach 2

point DFTs.
DFT FFT
x[0]
x[2]
x[4]
N/2 point
DFT
x[N-1]
X[0] = y[0]+W0z[0]
X[1] = y[1]+W1z[1]
y[N-2]
x[N-2]
x[1]
x[3]
x[5]
y[0]
y[1]
y[2]
N/2 point
DFT
z[0]
z[1]
z[2]
W0
W1
-1
-1
X[N/2] = y[0]-W0z[0]
X[N/2+1] = y[1]-W1z[1]
z[N/2-1]
Illustration of the first decimation in time

FFT.
FFT Implementation
Calculation of the output of a butterfly:

Y(k) =
WNkZ(k) = (Lr+jLi)(Wr+jWi)
Key:
U=Ur+jUi = X(k)
Ur+jUi
-1
L=Lr+jLi = X(k+N/2)
U = Upper
r = real
L = Lower
i = imaginary
Different methods are available for

calculating the outputs U and L.
The best method is the one with the least
number of multiplications and additions.
FFT Implementation

Ur+jUi
U=Ur+jUi
(Lr+jLi)(Wr+jWi)
L=Lr+jLi
-1
Lr jLi Wr jWi LrWr jLrWi jLiWr LiWi

U LrWr LiWi j LrWi LiWr U r jU i
LrWr LiWi U r j LrWi LiWr U i
L U r jU i LrWr LiWi j LrWi LiWr

U r LrWr LiWi j U i LrWi LiWr
FFT Implementation

U=Ur+jUi = (Lr Wr - Li Wi + Ur)+j(Lr Wi + Li Wr + Ui)
Ur+jUi
(Lr+jLi)(Wr+jWi)
-1
L=Lr+jLi = (Ur - Lr Wr + Li Wi)+j(Ui - Lr Wi - Li Wr)
To further minimise the number of

operations (* and +), the following are
calculated only once:
temp1 = LrWr
temp2 = LiWi
temp1_2 = temp1 - temp2
temp3 = LrWi
temp4 = LiWr
temp3_4 = temp3 + temp4
Ur = temp1 - temp2 + Ur = temp1_2 + Ur

Ui = temp3 + temp4 + Ui = temp3_4 + Ui
Lr = Ur - temp1 + temp2 = Ur - temp1_2
Li = Ui - temp3 - temp4
= Ui - temp3_4
FFT Implementation (Butterfly Calculation)
Converting the butterfly calculation into

C code:
temp1 = (y[lower].real * WR);
temp2 = (y[lower].imag * WI);
temp3 = (y[lower].real * WI);
temp4 = (y[lower].imag * WR);
temp1_2 = temp1 - temp2;
temp3_4 = temp 3 + temp4;
y[upper].real = temp1_2 + y[upper].real;
y[upper].imag = temp3_4 + y[upper].imag;
y[lower].imag = y[upper].imag - temp3_4;

y[lower].real = y[upper].real - temp1_2;
FFT Implementation
To efficiently implement the FFT

algorithm a few observations are made:
Each stage has the same number of

butterflies (number of butterflies = N/2, N is
number of points).
The number of DFT groups per stage is equal
to (N/2stage).
The difference between the upper and lower
leg is equal to 2stage-1.
The number of butterflies in the group is
equal to 2stage-1.
FFT
FFT Implementation
Example: 8 point FFT
W0
W0
W0
W0
-1
-1
W0
-1
W2
-1
-1
W0
-1
W0
-1
W1
-1
W0
-1
W2
-1
W2
-1
W3
-1
Decimation in time FFT:
Number of stages = log2N

Number of blocks/stage = N/2stage
Number of butterflies/block = 2stage-1
FFT Implementation
(1) Number of stages:
W0
W0
W0
W0
-1
-1
W0
-1
W2
-1
-1
W0
-1
W0
-1
W1
-1
W0
-1
W2
-1
W2
-1
W3
-1

FFT Implementation
Stage 1
W0
W0
W0
W0

Nstages = 1
-1
-1
W0
-1
W2
-1
-1
W0
-1
W0
-1
W1
-1
W0
-1
W2
-1
W2
-1
W3
-1

FFT Implementation
Stage 1
W0
W0
W0
W0
Stage 2

Nstages = 2
-1
-1
W0
-1
W2
-1
-1
W0
-1
W0
-1
W1
-1
W0
-1
W2
-1
W2
-1
W3
-1

FFT Implementation
Stage 1
W0
W0
W0
W0
Stage 2
Stage 3

Nstages = 3
-1
-1
W0
-1
W2
-1
-1
W0
-1
W0
-1
W1
-1
W0
-1
W2
-1
W2
-1
W3
-1

FFT Implementation
Stage 1
W0
W0
W0
W0
Stage 2
Stage 3

Nstages = 3
(2) Blocks/stage:
Stage 1:
-1
-1
W0
-1
W2
-1
-1
W0
-1
W0
-1
W1
-1
W0
-1
W2
-1
W2
-1
W3
-1

FFT Implementation
Stage 1
Stage 2
Stage 3

Nstages = 3
(2) Blocks/stage:
Stage 1: Nblocks = 1
Block 1
W0
W0
W0
W0
-1
-1
W0
-1
W2
-1
-1
W0
-1
W0
-1
W1
-1
W0
-1
W2
-1
W2
-1
W3
-1

FFT Implementation
Stage 1
Stage 2
Stage 3

Nstages = 3
(2) Blocks/stage:
Block 1
W0
-1
Block 2
W0
W0
W0
-1
W0
-1
W2
-1
-1
W0
-1
W0
-1
W1
-1
W0
-1
W2
-1
W2
-1
W3
-1

FFT Implementation
Stage 1
Stage 2
Stage 3

Nstages = 3
(2) Blocks/stage:
Block 1
W0
-1
Block 2
W0
-1
W0
-1
W2
-1
Block 3
W0
W0
-1
W0
-1
W0
-1
W1
-1
W0
-1
W2
-1
W2
-1
W3
-1

FFT Implementation
Stage 1
Stage 2
Stage 3

Nstages = 3
(2) Blocks/stage:
Block 1
W0
-1
Block 2
W0
-1
W0
-1
W2
-1
Block 3
W0
-1
W0
Block 4
W0
-1
W0
-1
W1
-1
W0
-1
W2
-1
W2
-1
W3
-1

FFT Implementation
Stage 1
Block 1
W0
W0
W0
W0
Stage 2
Stage 3

Nstages = 3
(2) Blocks/stage:
-1
-1
W0
-1
W2
-1
-1
W0
-1
W0
-1
W1
-1
W0
-1
W2
-1
W2
-1
W3
-1

FFT Implementation
Stage 1
Block 1
W0
W0
Block 2
W0
W0
Stage 2
Stage 3

Nstages = 3
(2) Blocks/stage:
-1
-1
W0
-1
W2
-1
-1
W0
-1
W0
-1
W1
-1
W0
-1
W2
-1
W2
-1
W3
-1

FFT Implementation
Stage 1
W0
Block 1
W0
W0
W0
Stage 2
Stage 3

Nstages = 3
(2) Blocks/stage:
-1
-1
W0
-1
W2
-1
-1
W0
-1
W0
-1
W1
-1
W0
-1
W2
-1
W2
-1
W3
-1

FFT Implementation
Stage 1
W0
W0
W0
W0
Stage 2
Stage 3

Nstages = 3
(2) Blocks/stage:
(3) Bflies/block:
Stage 1:
-1
-1
W0
-1
W2
-1
-1
W0
-1
W0
-1
W1
-1
W0
-1
W2
-1
W2
-1
W3
-1

FFT Implementation
Stage 1
W0
W0
W0
W0
Stage 2
Stage 3

Nstages = 3
(2) Blocks/stage:
(3) Bflies/block:
Stage 1: Nbtf = 1
-1
-1
W0
-1
W2
-1
-1
W0
-1
W0
-1
W1
-1
W0
-1
W2
-1
W2
-1
W3
-1

FFT Implementation
Stage 1
W0
W0
W0
W0
Stage 2
Stage 3
-1
-1
W0
-1
W2
-1
-1
W0
-1
W0
-1
W1
-1
W0
-1
W2
-1
W2
-1
W3
-1

Nstages = 3
(2) Blocks/stage:
(3) Bflies/block:
Stage 1: Nbtf = 1
Stage 2: Nbtf = 1

FFT Implementation
Stage 1
W0
W0
W0
W0
Stage 2
Stage 3
-1
-1
W0
-1
W2
-1
-1
W0
-1
W0
-1
W1
-1
W0
-1
W2
-1
W2
-1
W3
-1

Nstages = 3
(2) Blocks/stage:
(3) Bflies/block:
Stage 1: Nbtf = 1
Stage 2: Nbtf = 2

FFT Implementation
Stage 1
Stage 2
Stage 3

W
-1
Nstages = 3
W
-1
(2) Blocks/stage:
W
-1
W
-1
W
-1
W
-1
W
-1
W
W
W
-1
-1
(3) Bflies/block:
W
-1
W
-1
W
-1
Stage 1: Nbtf = 1
Stage 2: Nbtf = 2
Stage 3: Nbtf = 1
0

FFT Implementation
Stage 1
Stage 2
Stage 3

W
-1
Nstages = 3
W
-1
(2) Blocks/stage:
W
-1
W
-1
W
-1
W
-1
W
-1
W
W
W
-1
-1
(3) Bflies/block:
W
-1
W
-1
W
-1
Stage 1: Nbtf = 1
Stage 2: Nbtf = 2
Stage 3: Nbtf = 2
0

FFT Implementation
Stage 1
Stage 2
Stage 3

W
-1
Nstages = 3
W
-1
(2) Blocks/stage:
W
-1
W
-1
W
-1
W
-1
W
-1
W
W
W
-1
-1
(3) Bflies/block:
W
-1
W
-1
W
-1
Stage 1: Nbtf = 1
Stage 2: Nbtf = 2
Stage 3: Nbtf = 3
0

FFT Implementation
Stage 1
Stage 2
Stage 3

W
-1
Nstages = 3
W
-1
(2) Blocks/stage:
W
-1
W
-1
W
-1
W
-1
W
-1
W
W
W
-1
-1
(3) Bflies/block:
W
-1
W
-1
W
-1
Stage 1: Nbtf = 1
Stage 2: Nbtf = 2
Stage 3: Nbtf = 4
0

FFT Implementation
Stage 1
W0
-1
W0
-1
W0
Stage 3
-1
W0
Start Index
Input Index
Twiddle Factor Index
Stage 2
-1
0
1
W0
-1
W2
-1
W0
W0
-1
W1
-1
W0
-1
W2
-1
W2
-1
W3
-1
0
2
0
4
N/2 = 4
FFT Implementation
Stage 1
W0
-1
W0
-1
W0
Stage 3
-1
W0
Start Index
Input Index
Stage 2
-1
0
1
N/2 = 4
W0
-1
W2
-1
W0
W0
-1
W1
-1
W0
-1
W2
-1
W2
-1
W3
-1
0
2
0
4
4 /2 = 2
FFT Implementation
Stage 1
W0
-1
W0
-1
W0
Stage 3
-1
W0
Start Index
Input Index
Stage 2
-1
0
1
N/2 = 4
W0
-1
W2
-1
W0
W0
-1
W1
-1
W0
-1
W2
-1
W2
-1
W3
-1
0
2
4 /2 = 2
0
4
2 /2 = 1
FFT Implementation
Stage 1
W0
Stage 3
-1
W0
-1
W0
-1
W0
Start Index
Input Index
Indicies Used
Stage 2
-1
0
1
N/2 = 4
W0
W0
-1
W2
-1
W0
W0
-1
W1
-1
W0
-1
W2
-1
W2
-1
W3
-1
0
2
0
4
4 /2 = 2
W0
2 /2 = 1
W2
W1
W0
W2
W3
FFT Implementation
The most important aspect of converting
the FFT diagram to C code is to
calculate the upper and lower indices of
each butterfly:
GS = N/4;
step = 1;
NB = N/2;
/* Group step initial value */

/* Initial value */
/* NB is a constant */
for (k=N; k>1; k>>1)

/*
{
for (j=0; j<N; j+=GS)
/*
{
for (n=j; n<(j+GS-1); n+=step) /*
{
upperindex = n;
lowerindex = n+step;
}
}
/* Change the GS and step for the next
Repeat this loop for each stage */

Repeat this loop for each block */
Repeat this loop for each butterfly */
stage */
GS = GS << 1;
step = step << 1;
}
FFT Implementation
How to declare and access twiddle factors:

(1) How to declare the twiddle factors:
struct {
short real; // 32767 * cos (2*pi*n) -> Q15 format
short imag; // 32767 * sin (2*pi*n) -> Q15 format
} w[] = { 32767,0,
32767,-201,
...
};
(2) How to access the twiddle factors:

short temp_real, temp_imag;
temp_real = w[i].real;
temp_imag = w[i].imag;
FFT Implementation
Links:
Project location:
\CD ROM 2004\DSP Code 6711 Update\Chapter 19 - Fast Fourier

Transform\FFT DIF _in C_Fixed point\fft_bios.pjt
\CD ROM 2004\DSP Code for DSK6416\Chapter 19 - Fast Fourier
Transform\FFT DIF _in C_Fixed point\fft_project.pjt
\CD ROM 2004\DSP Code for DSK6713\Chapter 19 - Fast Fourier
Transform\FFT DIF _in C_Fixed point\fft_project.pjt
Laboratory sheet: FFTLabSheet.pdf
FFT test on the DSK6713
FFT test on the DSK6416
Chapter 19
Fast Fourier Transform (FFT)
(Theory and Implementation)
- End -

Fast Fourier Transform (FFT) (Theory and Implementation)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Fast Fourier Transform (FFT) (Theory and Implementation)

Uploaded by

Copyright:

Available Formats

Chapter 19

Fast Fourier Transform (FFT)

Chapter 19, Slide 2

Dr. Naim Dahnoun, Bristol University, (c) Texas Instruments 2004

The Fourier transform of an analogue

The Discrete Fourier Transform (DFT) of

Chapter 19, Slide 3

Chapter 19, Slide 4

Dr. Naim Dahnoun, Bristol University, (c) Texas Instruments 2004

X[k] = frequency bins

= x[0]WN0 + x[1]WN0*1 ++ x[N-1]WN0*(N-1)

= x[0]WN0 + x[1]WN1*1 ++ x[N-1]WN1*(N-1)

= x[0]WN0 + x[1]WNk*1 ++ x[N-1]WNk*(N-1)

X(N-1) = x[0]WN0 + x[1]WN (N-1)*1 ++ x[N-1]WN (N-1)(N-1)

Note: For N samples of x we have N frequencies

Dr. Naim Dahnoun, Bristol University, (c) Texas Instruments 2004

Performance of the DFT Algorithm

The DFT requires N2 (NxN) complex

The DFT also requires (N-1)*N complex

Chapter 19, Slide 6

Each X(k) requires N complex

Each X(k) requires N-1 additions.

Performance of the DFT Algorithm

Chapter 19, Slide 7

Can the number of computations required

Chapter 19, Slide 8

A large amount of work has been devoted

Dr. Naim Dahnoun, Bristol University, (c) Texas Instruments 2004

x[n] = x[0], x[1], , x[N-1]

Lets divide the sequence x[n] into even

Chapter 19, Slide 9

Dr. Naim Dahnoun, Bristol University, (c) Texas Instruments 2004

Equation 1 can be rewritten as:

WN2 n 1k WNk W Nnk

x2nW Nnk WNk

Dr. Naim Dahnoun, Bristol University, (c) Texas Instruments 2004

The result is that an N-point DFT can be

Where Y(k) and Z(k) are the two N/2

Two N/2point DFTs

Dr. Naim Dahnoun, Bristol University, (c) Texas Instruments 2004

Periodicity and symmetry of W can be

Chapter 19, Slide 12

Dr. Naim Dahnoun, Bristol University, (c) Texas Instruments 2004

Chapter 19, Slide 13

Dr. Naim Dahnoun, Bristol University, (c) Texas Instruments 2004

Finally by exploiting the symmetry and

X k x1 nW Nnk WNk x2 nW Nnk

Chapter 19, Slide 14

Dr. Naim Dahnoun, Bristol University, (c) Texas Instruments 2004

Y(k) and WNk Z(k) only need to be

Chapter 19, Slide 15

Dr. Naim Dahnoun, Bristol University, (c) Texas Instruments 2004

Y(k) and Z(k) can also be divided into N/4

The process continues until we reach 2

Chapter 19, Slide 16

Dr. Naim Dahnoun, Bristol University, (c) Texas Instruments 2004

Illustration of the first decimation in time

Chapter 19, Slide 17

Dr. Naim Dahnoun, Bristol University, (c) Texas Instruments 2004

Calculation of the output of a butterfly:

Different methods are available for

Chapter 19, Slide 18

Dr. Naim Dahnoun, Bristol University, (c) Texas Instruments 2004

= x[0]WN0 + x[1]WN01 ++ x[N-1]WN0(N-1)

= x[0]WN0 + x[1]WN11 ++ x[N-1]WN1(N-1)

= x[0]WN0 + x[1]WNk1 ++ x[N-1]WNk(N-1)