discrete-time fourier series concept - unt computer …rakl/class3010/chapter7.pdf · ·...
TRANSCRIPT
1
Discrete-Time Fourier Methods
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl
1
Discrete-Time Fourier Series Concept
A signal can be represented as a linear combination of sinusoids.
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 2
The relationship between complex and real sinusoids
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 3
Discrete-Time Fourier Series Concept
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 4
Discrete-Time Fourier Series Concept
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 5
Discrete-Time Fourier Series Concept
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 6
Discrete-Time Fourier Series Concept
2
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 7
Discrete-Time Fourier Series Concept
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 8
Discrete-Time Fourier Series Concept
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 9
Discrete-Time Fourier Series Concept
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 10
Discrete-Time Fourier Series Concept
The Discrete-Time Fourier Series
The discrete-time Fourier series (DTFS) is similar to the CTFS.A periodic discrete-time signal can be expressed as
x n[ ] = cx k[ ]e j2πkn /N
k= N∑ cx k[ ] = 1
Nx n[ ]e− j2πkn /N
n=n0
n0 +N−1
∑where cx k[ ] is the harmonic function, N is any period of x n[ ] and the notation,
k= N∑ means a summation over any range of
consecutive k’s exactly N in length.
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 11
The Discrete Fourier Transform The discrete Fourier transform (DFT) is almost identical to the DTFS.A periodic discrete-time signal can be expressed as
x n[ ] = 1N
X k[ ]e j2πkn /N
k= N∑ X k[ ] = x n[ ]e− j2πkn /N
n=n0
n0 +N−1
∑where X k[ ] is the DFT harmonic function, N is any period of x n[ ] and the notation,
k= N∑ means a summation over any range of
consecutive k’s exactly N in length. The main difference between the DTFS and the DFT is the location of the 1/N term. So X k[ ] = N cx k[ ].
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 12
3
The Discrete Fourier Transform
Because the DTFS and DFT are so similar, and because the DFT isso widely used in digital signal processing (DSP), we will concentrateon the DFT realizing we can always form the DTFS from cx k[ ] = X k[ ] / N .
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 13
The Discrete Fourier Transform
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 14
DFT Example
Find the DFT harmonic function for
x n⎡⎣ ⎤⎦ = u n⎡⎣ ⎤⎦ − u n − 3⎡⎣ ⎤⎦( )∗δ5 n⎡⎣ ⎤⎦using its fundamental period as the representation time.
X k⎡⎣ ⎤⎦ = x n⎡⎣ ⎤⎦e− j2πkn/ N
n= N∑
= u n⎡⎣ ⎤⎦ − u n − 3⎡⎣ ⎤⎦( )∗δ5 n⎡⎣ ⎤⎦e− j2πkn/5
n=0
4
∑
= e− j2πkn/5
n=0
2
∑ = 1− e− j6πk /5
1− e− j2πk /5 =e− j3πk /5
e− jπk /5 × e j3πk /5 − e− j3πk /5
e jπk /5 − e− jπk /5
= e− j2πkn/5
n=0
2
∑ = e− j2πk /5sin 3πk / 5( )sin πk / 5( ) = 3e− j2πk /5 drcl k / 5,3( )
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 15
The DFT Harmonic Function
We know that x n[ ] = 1N
X k[ ]e j2πkn /N
k= N∑ so we can find x n[ ]
from its harmonic function. But how do we find the harmonicfunction from x n[ ]? We use the principle of orthogonality likewe did with the CTFS except that now the orthogonality is indiscrete time instead of continuous time.
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 16
The DFT Harmonic Function
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 17
The DFT Harmonic Function
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 18
4
The DFT Harmonic Function Below is a set of complex sinusoids for N = 8. They form aset of basis vectors. Notice that the k = 7 complex sinusoid rotates counterclockwise through 7 cycles but appears to rotate clockwise through one cycle. The k = 7 complex sinusoid is exactly the same as the k = −1 complex sinusoid. This must be true because the DFT is periodic with period N .
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 19
The DFT Harmonic Function The projection of a real vector x in the direction of anotherreal vector y is
p = xTyyTy
y
If p = 0, x and y are orthogonal. If the vectors are complex-valued
p = xHyyHy
y
where the xH is the complex-conjugate transpose of x. xTyand xHy are the dot product of x and y.
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 20
The DFT Harmonic Function
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 21
The DFT Harmonic Function
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 22
The DFT Harmonic Function
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 23
The DFT Harmonic Function
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 24
5
The DFT Harmonic Function
The most common definition of the DFT is
X k[ ] = x n[ ]e− j2πkn /N
n=0
N−1
∑ , x n[ ] = 1N
X k[ ]e j2πkn /N
k= N∑
Here the beginning point for x n[ ] is taken as n0 = 0 . This is the form of the DFT that is implemented in practically all computer languages.
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 25
Convergence of the DFT
• The DFT converges exactly with a finite number of terms. It does not have a “Gibbs phenomenon” in the same sense that the CTFS does
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 26
The Discrete Fourier Transform
X k[ ] is called the DFT harmonic function of x n[ ] and k is the harmonic number just as we have seen in the CTFS. x n[ ]and X k[ ] form a DFT pair based on N points.
x n[ ] D F TN← →⎯⎯ X k[ ]
From x n[ ] = 1N
X k[ ]e j2πkn /N
k= N∑ we see that x n[ ] is formed
by a linear combination of functions of the form e j2πkn /N eachof which has a period N . Therefore x n[ ] must also be periodicwith period (but not necessarily fundamental period) N .
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 27
DFT Properties
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 28
DFT Properties
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 29
DFT Properties
It can be shown (and is in the text) that if x n[ ] is an even function, X k[ ] is purely real and if x n[ ] is an odd functionX k[ ] is purely imaginary.
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 30
6
The Dirichlet Function
The functional form sin πNt( )N sin πt( )
appears often in discrete-time signal analysis and is given the special name Dirichlet function. That is
drcl t,N( ) = sin πNt( )N sin πt( )
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 31
DFT Pairs
e j2πn /N D F TmN← →⎯⎯ mNδmN k − m[ ]
cos 2πqn / N( ) D F TmN← →⎯⎯ mN / 2( ) δmN k − mq[ ]+δmN k + mq[ ]( )
sin 2πqn / N( ) D F TmN← →⎯⎯ jmN / 2( ) δmN k + mq[ ]−δmN k − mq[ ]( )
δN n[ ] D F TmN← →⎯⎯ mδmN k[ ]
1 D F TN← →⎯⎯ NδN k[ ]
u n − n0[ ]− u n − n1[ ]( )∗δN n[ ] D F TN← →⎯⎯ e− jπk n1+n0( )/N
e− jπk /N n1 − n0( )drcl k / N ,n1 − n0( ) tri n / Nw( )∗δN n[ ] D F T
N← →⎯⎯ Nw drcl2 k / N ,Nw( ) , Nw an integer
sinc n /w( )∗δN n[ ] D F TN← →⎯⎯ wrect wk / N( )∗δN k[ ]
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 32
The Fast Fourier Transform
One could write a MATLAB program to implement the DFT..
% (Acquire the input data in an array x with N elements.)
.
% Initialize the DFT array to a column vector of zeros.
X = zeros(N,1) ;
% Compute the Xn’s in a nested, double for loop.
for k = 0:N-1
for n = 0:N-1
X(k + 1) = X(k + 1) + x(n + 1)*exp(-j*2 *pi*n*k / N) ;
end
end
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 33
The Fast Fourier Transform
There is a function in MATLAB fft which accomplishes the same goal and is typically much faster. This table compares the speeds of the two methods. Mstands for computer multiplies and A stands for computer additions.γ N = 2γ ADFT MDFT AFFT MFFT ADFT / AFFT MDFT / MFFT
1 2 2 4 2 1 1 42 4 12 16 8 4 1.5 43 8 56 64 24 12 2.33 5.334 16 240 256 64 32 3.75 85 32 992 1024 160 80 6.2 12.86 64 4032 4096 384 192 10.5 21.37 128 16256 16384 896 448 18.1 36.68 256 65280 65536 2048 1024 31.9 649 512 261632 262144 4608 2304 56.8 113.8
10 1024 1047552 1048576 10240 5120 102.3 204.8
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 34
Generalizing the DFT for Aperiodic Signals
Pulse Train
This periodic rectangular-wave signal is analogous to the continuous-time periodic rectangular-wave signal used to illustrate the transition from the CTFS to the CTFT.
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 35
DFT of Pulse Train
As the period of the rectangular wave increases, the period of the DFT increases
Generalizing the DFT for Aperiodic Signals
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 36
7
Normalized DFT of
Pulse Train
By plotting versus k / N0 instead of k, the period of the normalized DFT stays at one.
Generalizing the DFT for Aperiodic Signals
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 37
The normalized DFT approaches this limit as the period approaches infinity.
Generalizing the DFT for Aperiodic Signals
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 38
Definition of the Discrete-Time Fourier Transform (DTFT)
Forward F Form Inverse
x n[ ] = X F( )e j2πFn dF1∫
F← →⎯ X F( ) = x n[ ]e− j2πFnn=−∞
∞
∑
Forward Ω Form Inverse
x n[ ] = 12π
X e jΩ( )e jΩn dΩ2π∫
F← →⎯ X e jΩ( ) = x n[ ]e− jΩnn=−∞
∞
∑
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 39
The function e− jΩ appears in the forward DTFT raised to the nth power. It is periodic in Ω with fundamental period 2π . n is an integer. Therefore e- jΩn is periodic with fundamental period 2π / n and 2π is also a period of e− jΩn . The forward DTFT is
X e jΩ( ) = x n[ ]e− jΩnn=−∞
∞
∑a weighted summation of functions of the form e− jΩn , all of which repeat with
every 2π change in Ω. Therefore X e jΩ( ) is always periodic in Ω with period
2π . This also implies that X F( ) is always periodic in F with period 1.
The Discrete-Time Fourier Transform
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 40
DTFT Pairs We can begin a table of DTFT pairs directly from the definition. (There is a more extensive table in the text.)
δ n[ ] F← →⎯ 1
α n u n[ ] F← →⎯e jΩ
e jΩ −α=
11−αe− jΩ
, α < 1 , −α n u −n −1[ ] F← →⎯e jΩ
e jΩ −α=
11−αe− jΩ
, α > 1
α n sin Ω0n( )u n[ ] F← →⎯e jΩα sin Ω0( )
e j2Ω − 2αe jΩ cos Ω0( ) +α 2 , α < 1 , −α n sin Ω0n( )u −n −1[ ] F← →⎯e jΩα sin Ω0( )
e j2Ω − 2αe jΩ cos Ω0( ) +α 2 , α > 1
α n cos Ω0n( )u n[ ] F← →⎯e jΩ e jΩ −α cos Ω0( )⎡⎣ ⎤⎦
e j2Ω − 2αe jΩ cos Ω0( ) +α 2 , α < 1 , −α n cos Ω0n( )u −n −1[ ] F← →⎯e jΩ e jΩ −α cos Ω0( )⎡⎣ ⎤⎦
e j2Ω − 2αe jΩ cos Ω0( ) +α 2 , α > 1
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 41
The Generalized DTFT By generalizing the CTFT to include transform that have impulseswe were able to find CTFT's of some important practical functions.The same is true of the DTFT. The DTFT of a constant
X F( ) = Ae− j2πFnn=−∞
∞
∑ = A e− j2πFnn=−∞
∞
∑does not converge. The CTFT of a constant turned out to be an impulse. Since the DTFT must be periodic that cannot the the transform of a constant in discrete time. Instead the transform mustbe a periodic impulse.
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 42
8
The Generalized DTFT
To shorten the process, find the inverse DTFT of a periodicimpulse of the form Aδ1 F( ). Using the formula
x n[ ] = Aδ1 F( )e j2πFndF1∫ = A δ F( )e j2πFndF
−1/2
1/2
∫ = A
proving that the DTFT of a constant A is Aδ1 F( ) or, in radian-
frequency form A F← →⎯ 2πAδ2π Ω( ).
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 43
The Generalized DTFT
Now consider the function X F( ) = Aδ1 F − F0( ) , −1 / 2 < F0 <1 / 2. Its inverse DTFT is
x n[ ] = Aδ1 F − F0( )e j2πFn dF1∫ = A δ F − F0( )e j2πFn dF
−1/2
1/2
∫ = Aej2πF0n
Now change x n[ ] to x n[ ] = Acos 2πF0n( ) = A / 2( ) e j2πF0n + e− j2πF0n( ).Then
Acos 2πF0n( ) F← →⎯ A / 2( ) δ1 F − F0( ) +δ1 F + F0( )⎡⎣ ⎤⎦or
Acos Ω0n( ) F← →⎯ πA δ2π Ω −Ω0( ) +δ2π Ω +Ω0( )⎡⎣ ⎤⎦
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 44
Forward DTFT Example
Find the forward DTFT of x n[ ] = u n − n0[ ]− u n − n1[ ].
u n − n0[ ]− u n − n1[ ] F← →⎯ u n − n0[ ]− u n − n1[ ]( )e− j2πFnn=−∞
∞
∑ = e− j2πFnn=n0
n1 −1
∑Let m = n − n0 . Then
u n − n0[ ]− u n − n1[ ] F← →⎯ e− j2πF m+n0( )
m=0
n1 −n0 −1
∑ = e− j2πFn0 e− j2πFmm=0
n1 −n0 −1
∑Summing this geometric series
u n − n0[ ]− u n − n1[ ] F← →⎯ e− j2πFn01− e− j2πF n1 −n0( )
1− e− j2πF
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 45
Forward DTFT Example
u n − n0[ ]− u n − n1[ ] F← →⎯ e− j2πFn01− e− j2πF n1 −n0( )
1− e− j2πF
Factor out e− jπF n1 −n0( ) from the numerator and e− jπF from the denominator
u n − n0[ ]− u n − n1[ ] F← →⎯ e− j2πFn0e− jπF n1 −n0( )
e− jπFe jπF n1 −n0( ) − e− jπF n1 −n0( )
e jπF − e− jπF
By the definition of the sine function in terms of complex exponentials
u n − n0[ ]− u n − n1[ ] F← →⎯e− jπF n0 +n1( )
e− jπFsin πF n1 − n0( )( )
sin πF( )
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 46
Forward DTFT Example
Consider the special case of n0 + n1 = 1⇒ n0 = 1− n1 (making thefunction a discrete-time rectangular pulse of width 2n0 +1 centered at n = 0.
u n − n0[ ]− u n − n1[ ] F← →⎯sin πF n1 − n0( )( )
sin πF( ) , n0 + n1 = 1
Compare this to the CTFT of a rectangular pulse of width w centered at t = 0.
rect t /w( ) F← →⎯ wsinc wf( ) = sin πwf( )π f
The DTFT is a periodically-repeated sinc function.
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 47
More DTFT Pairs
δ n[ ] F← →⎯ 1
u n[ ] F← →⎯1
1− e− j2πF+ 1 / 2( )δ1 F( ) , u n[ ] F← →⎯
11− e− jΩ
+ πδ1 Ω( )
sinc n / w( ) F← →⎯ w rect wF( )∗δ1 F( ) , sinc n / w( ) F← →⎯ w rect wΩ / 2π( )∗δ2π Ω( ) tri n / w( ) F← →⎯ wdrcl2 F,w( ) , tri n / w( ) F← →⎯ wdrcl2 Ω / 2π ,w( ) 1 F← →⎯ δ1 F( ) , 1 F← →⎯ 2πδ2π Ω( ) δN0
n[ ] F← →⎯ 1 / N0( )δ1/N0F( ) , δN0
n[ ] F← →⎯ 2π / N0( )δ2π /N0Ω( )
cos 2πF0n( ) F← →⎯ 1 / 2( ) δ1 F − F0( ) + δ1 F + F0( )⎡⎣ ⎤⎦ , cos Ω0n( ) F← →⎯ π δ2π Ω − Ω0( ) + δ2π Ω +Ω0( )⎡⎣ ⎤⎦ sin 2πF0n( ) F← →⎯ j / 2( ) δ1 F + F0( ) − δ1 F − F0( )⎡⎣ ⎤⎦ , sin Ω0n( ) F← →⎯ jπ δ2π Ω +Ω0( ) − δ2π Ω − Ω0( )⎡⎣ ⎤⎦
u n − n0[ ]− u n − n1[ ] F← →⎯e j2πF
e j2πF −1e− j2πn0F − e− j2πn1F( ) = e− jπF n0 +n1( )
e− jπFn1 − n0( )drcl F,n1 − n0( )
u n − n0[ ]− u n − n1[ ] F← →⎯e jΩ
e jΩ −1e− jn0Ω − e− jn1Ω( ) = e− jΩ n0 +n1( )/2
e− jΩ /2 n1 − n0( )drcl Ω / 2π ,n1 − n0( )
We can now extend the table of DTFT pairs.
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 48
9
DTFT Properties
α x n[ ]+ β y n[ ] F← →⎯ α X F( ) + β Y F( ) , α x n[ ]+ β y n[ ] F← →⎯ α X e jΩ( ) + β Y e jΩ( ) x n − n0[ ] F← →⎯ e− j2πFn0 X F( ) , x n − n0[ ] F← →⎯ e− jΩn0 X e jΩ( ) e j2πF0n x n[ ] F← →⎯ X F − F0( ) , e jΩ0n x n[ ] F← →⎯ X e j Ω−Ω0( )( )If y n[ ] = x n /m[ ] , n /m an integer
0 , otherwise⎧⎨⎩
⎫⎬⎭
then y n[ ] F← →⎯ X mF( ) or y n[ ] F← →⎯ X e jmΩ( )x* n[ ] F← →⎯ X* −F( ) , x* n[ ] F← →⎯ X* e− jΩ( )x n[ ]− x n −1[ ] F← →⎯ 1− e− j2πF( )X F( ) , x n[ ]− x n −1[ ] F← →⎯ 1− e− jΩ( )X e jΩ( )
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 49
DTFT Properties
x m[ ]m=−∞
n
∑ F← →⎯X F( )
1− e− j2πF+ 1
2X 0( )δ1 F( ) , x m[ ]
m=−∞
n
∑ F← →⎯X e jΩ( )1− e− jΩ
+π X e j0( )δ2π Ω( )
x −n[ ] F← →⎯ X −F( ) , x −n[ ] F← →⎯ X e− jΩ( ) x n[ ]∗y n[ ] F← →⎯ X F( )Y F( ) , x n[ ]∗y n[ ] F← →⎯ X e jΩ( )Y e jΩ( ) x n[ ]y n[ ] F← →⎯ X F( )Y F( )see note , x n[ ]y n[ ] F← →⎯ 1 / 2π( )X e jΩ( )Y e jΩ( )see note
e j2πFnn=−∞
∞
∑ = δ1 F( ) , e jΩnn=−∞
∞
∑ = 2πδ2π Ω( )
x n[ ] 2
n=−∞
∞
∑ = X F( ) 2 dF1∫ , x n[ ] 2
n=−∞
∞
∑ = 1 / 2π( ) X e jΩ( ) 2dΩ
2π∫Note: x t( )y t( ) = x τ( )y t −τ( )dτ
T∫ where T is a period of both x and y( )
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 50
DTFT Properties
Find and graph the inverse DTFT of
X F( ) = rect 50 F −1 / 4( )( ) + rect 50 F +1 / 4( )( )⎡⎣ ⎤⎦ ∗δ1 F( )Start with
sinc n /w( ) F← →⎯ wrect wF( )∗δ1 F( )In this case w = 50.
1 / 50( )sinc n / 50( ) F← →⎯ rect 50F( )∗δ1 F( )Then, using the frequency-shifting property
e j2πF0n x n[ ] F← →⎯ X F − F0( ) e jπn /2 1 / 50( )sinc n / 50( ) F← →⎯ rect 50 F −1 / 4( )( )∗δ1 F( )
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 51
DTFT Properties
Frequency shifting the other direction
e− jπn /2 1 / 50( )sinc n / 50( ) F← →⎯ rect 50 F +1 / 4( )( )∗δ1 F( )
Combining the last two results and using cos x( ) = ejx + e− jx
2 1 / 25( )sinc n / 50( )cos πn / 2( ) F← →⎯
rect 50 F −1 / 4( )( ) + rect 50 F +1 / 4( )( )⎡⎣ ⎤⎦ ∗δ1 F( )
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 52
DTFT Properties Time scaling in discrete time is quite different from timescaling in continuous-time. Let z n[ ] = x an[ ]. If a is not an integer some values of z n[ ] are undefined and a DTFT cannot be found for it. If a is an integer greater than one, some values of x n[ ] will not appear in z n[ ] because of decimation and there cannot be a unique relationship between their DTFT’s
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 53
DTFT Properties
Time scaling does not work for time compression because ofdecimation. But it does work for a special type of time expansion.
Let z n[ ] = x n /m[ ] , n /m an integer0 , otherwise
⎧⎨⎩
. Then Z F( ) = X mF( ).
So the time-scaling property of the DTFT is
z n[ ] = x n /m[ ] , n /m an integer0 , otherwise
⎧⎨⎩
, z n[ ] F← →⎯ X mF( )
or z n[ ] F← →⎯ X e jmΩ( )
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 54
10
DTFT Properties
In the time domain, the response of a system is the convolutionof the excitation with the impulse response y n[ ] = x n[ ]∗h n[ ]In the frequency domain the response of a system is the productof the excitation and the frequency response of the system
Y e jΩ( ) = X e jΩ( )H e jΩ( )
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 55
DTFT Properties
Find the signal energy of x n[ ] = 1 / 5( )sinc n /100( ). Thestraightforward way of finding signal energy is directly from
the definition Ex = x n[ ] 2
n=−∞
∞
∑ .
Ex = 1 / 5( )sinc n /100( ) 2
n=−∞
∞
∑ = 1 / 25( ) sinc2 n /100( )n=−∞
∞
∑In this case we run into difficulty because we don't know howto sum this series.
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 56
DTFT Properties
We can use Parseval's theorem to find the signal energy fromthe DTFT of the signal.
1 / 5( )sinc n /100( ) F← →⎯ 20rect 100F( )∗δ1 F( )Parseval's theorem is
x n[ ] 2
n=−∞
∞
∑ = X F( ) 2 dF1∫
For this case
Ex = 20rect 100F( )∗δ1 F( ) 2 dF1∫ = 20rect 100F( ) 2 dF
−1/2
1/2
∫
Ex = 400 dF−1/200
1/200
∫ = 4
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 57
Transform Method Comparisons
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 58
Transform Method Comparisons
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 59
Transform Method Comparisons
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 60
11
Transform Method Comparisons Using the equivalence property of the impulse and fact that both e j2πF and δ1 F( ) have a fundamental period of one,
Y F( ) = 1 / 2( ) e jπ /6 δ1 F −1 /12( )e jπ /6 − 0.9
+ e− jπ /6 δ1 F +1 /12( )e− jπ /6 − 0.9
⎡⎣⎢
⎤⎦⎥
Finding a common denominator and simplifying,
Y F( ) = 1 / 2( )δ1 F −1 /12( ) 1− 0.9e jπ /6( ) +δ1 F +1 /12( ) 1− 0.9e− jπ /6( )1.81−1.8cos π / 6( )
Y F( ) = 0.4391 δ1 F −1 /12( ) +δ1 F +1 /12( )⎡⎣ ⎤⎦ + j0.8957 δ1 F +1 /12( )−δ1 F −1 /12( )⎡⎣ ⎤⎦ y n[ ] = 0.8782cos 2πn /12( ) +1.7914sin 2πn /12( ) y n[ ] = 1.995cos 2πn /12 −1.115( )
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 61
Transform Method Comparisons The DFT can often be used to find the DTFT of a signal. The
DTFT is defined by X F( ) = x n[ ]e− j2πFnn=−∞
∞
∑ and the DFT
is defined by X k[ ] = x n[ ]e− j2πkn /N
n=0
N−1
∑ . If the signal x n[ ] is causal
and time limited, the summation in the DTFT is over a finite range of n values beginning with 0 and we can set the value of N by letting N −1 be the last value of n needed to cover that finite
range. Then X F( ) = x n[ ]e− j2πFnn=0
N−1
∑ . Now let F→ k / N yielding
X k / N( ) = x n[ ]e− j2πkn /N = X k[ ]n=0
N−1
∑
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 62
Transform Method Comparisons
The result
X k / N( ) = x n[ ]e− j2πkn /N = X k[ ]n=0
N−1
∑is the DTFT of x n[ ] at a discrete set of frequencies F = k / N orΩ = 2πk / N . If that resolution in frequency is not sufficient, Ncan be made larger by augmenting the previous set of x n[ ] valueswith zeros. That reduces the space between frequency pointsthereby increasing the resolution. This technique is called zeropadding.
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 63
Transform Method Comparisons We can also use the DFT to approximate the inverse DTFT.
The inverse DTFT is defined by x n[ ] = X F( )e j2πFn dF1∫
and the inverse DFT is defined by x n[ ] = 1N
X k[ ]e j2πkn /N
k=0
N−1
∑ .
We can approximate the inverse DTFT by
x n[ ] ≅ X k / N( )e j2πFn dFk /N
k+1( )/N
∫k=0
N−1
∑ = X k / N( ) e j2πFn dFk /N
k+1( )/N
∫k=0
N−1
∑
x n[ ] ≅ X k / N( ) ej2π k+1( )n /N − e j2πk /N
j2πnk=0
N−1
∑ = ej2πn /N −1j2πn
X k / N( )e j2πkn /N
k=0
N−1
∑
x n[ ] ≅ e jπn /N sinc n / N( ) 1N
X k / N( )e j2πkn /N
k=0
N−1
∑
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 64
Transform Method Comparisons
For n << N ,
x n[ ] ≅ 1N
X k / N( )e j2πkn /N
k=0
N−1
∑This is the inverse DFT with X k[ ] = X k / N( ).
Use this result to find the inverse DTFT of
X F( ) = rect 50 F −1 / 4( )( ) + rect 50 F +1 / 4( )( )⎡⎣ ⎤⎦ ∗δ1 F( )with the inverse DFT.
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 65
Transform Method Comparisons
N = 512 ; % Number of pts to approximate X(F)
k = [0:N -1]' ; % Harmonic numbers
% Compute samples from X(F) between 0 and 1 assuming
% periodic repetition with period 1
X = rect(50 *(k / N - 1 / 4)) + rect(50 *(k / N - 3 / 4)) ;
% Compute the approximate inverse DTFT and
% center the function on n = 0
xa = real(fftshift(ifft(X))) ;
n = [-N / 2:N / 2 - 1]' ; % Vector of discrete times for plotting
% Compute exact x[n] from exact inverse DTFT
xe = sinc(n / 50).*cos(pi*n / 2) / 25 ;
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 66
12
Transform Method Comparisons
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 67
The Four Fourier Methods
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 68
Relations Among Fourier Methods
Discrete Frequency Continuous Frequency
Continuous Time x t( )y t( ) F S← →⎯⎯ X k[ ]∗Y k[ ] x t( )y t( ) F← →⎯ X f( )∗Y f( )Discrete Time x n[ ]y n[ ] F S← →⎯⎯ Y k[ ]X k[ ] x n[ ]y n[ ] F← →⎯ X F( )Y F( )
Discrete Frequency Continuous Frequency
Continuous Time x t( )y t( ) F S← →⎯⎯ T0 X k[ ]Y k[ ] x t( )∗y t( ) F← →⎯ X f( )Y f( )Discrete Time x n[ ]y n[ ] F S← →⎯⎯ N0 Y k[ ]X k[ ] x n[ ]∗y n[ ] F← →⎯ X F( )Y F( )
Multiplication-Convolution Duality
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 69
Relations Among Fourier Methods
Discrete Frequency Continuous Frequency
Continuous Time 1T0
x t( ) 2 dtT0∫ = X k[ ] 2
k=−∞
∞
∑ x t( ) 2 dt−∞
∞
∫ = X f( ) 2 df−∞
∞
∫
Discrete Time x n[ ] 2
n= N∑ = 1
NX k[ ] 2
k= N∑ x n[ ] 2
n=−∞
∞
∑ = X F( ) 2 dF1∫
Parseval’s Theorem
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 70
Relations Among Fourier Methods
Discrete Frequency Continuous Frequency
Continuous Time x t − t0( ) F S← →⎯⎯ cx k[ ]e− j2πkt0 /N x t − t0( ) F← →⎯ X jω( )e− jωt0
Discrete Time x n − n0[ ] F S← →⎯⎯ X k[ ]e− j2πkn0 /N x n − n0[ ] F← →⎯ X e jΩ( )e− jΩn0
Discrete Frequency Continuous Frequency
Continuous Time x t( )e+ j2πkt /T F ST← →⎯⎯ cx k − k0[ ] x t( )e+ jω0t F← →⎯ X j ω −ω0( )( )
Discrete Time x n[ ]e+ j2πkn /N F SN← →⎯⎯ X k − k0[ ] x n[ ]e+ jΩ0n F← →⎯ X e j Ω−Ω0( )( )
Time and Frequency Shifting
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 71
Relations Among Fourier Methods
M. J. Roberts - All Rights Reserved. Edited by Dr. Robert Akl 72