Chap3-Discrete Fourier Transform

Embed Size (px)

Citation preview

  • 8/9/2019 Chap3-Discrete Fourier Transform

    1/83

    Copyright 2005. Shi Ping CUC

    Chapter 3

    Discrete Fourier Transform

    Review

    Features in common

    We need a numerically computable transform, that is

    Discrete Fourier Transform (DFT)

    The DTFT provides the frequency-domain ( )

    representation for absolutely summable sequences.

    The z-transform provides a generalized frequency-

    domain ( ) representation for arbitrary sequences.

    [

    z

    Defined for infinite-length sequences.

    Functions of continuous variable ( or ).

    They are not numerically computable transform.[ z

  • 8/9/2019 Chap3-Discrete Fourier Transform

    2/83

    Copyright 2005. Shi Ping CUC

    Chapter 3

    Discrete Fourier Transform

    Content

    The Family of Fourier Transform

    The Discrete Fourier Series (DFS)

    The Discrete Fourier Transform (DFT)

    The Properties of DFT

    The Sampling Theorem in Frequency Domain

    Approximating to FT (FS) with DFT (DFS)

    Summary

  • 8/9/2019 Chap3-Discrete Fourier Transform

    3/83

    Copyright 2005. Shi Ping CUC

    The Family of Fourier Transform

    Introduction

    Fourier analysis is named afterJean Baptiste Joseph

    Fourier(1768-1830), a French mathematician andphysicist.

    A signal can be eithercontinuous ordiscrete, and it can

    be eitherperiodicoraperiodic. The combination ofthese two features generates the four categories of

    Fourier Transform.

  • 8/9/2019 Chap3-Discrete Fourier Transform

    4/83

    Copyright 2005. Shi Ping CUC

    The Family of Fourier Transform

    Aperiodic-ContinuousFourier Transform

    g

    g

    ;

    g

    g;

    ;;!

    !;

    dejXtx

    dtetxjX

    tj

    tj

    )(2

    1)(

    )()(

    T

  • 8/9/2019 Chap3-Discrete Fourier Transform

    5/83

    Copyright 2005. Shi Ping CUC

    The Family of Fourier Transform

    Periodic-ContinuousFourier Series

    g

    g!

    ;

    ;

    ;!

    !;

    k

    tjk

    T

    T

    tjk

    ejkXtx

    dtetxTjkX

    0

    0

    0

    0

    )()(

    )(

    1

    )(

    0

    2

    20

    0

    0

    0

    22

    TF

    TT !!;

  • 8/9/2019 Chap3-Discrete Fourier Transform

    6/83

    g

    g!

    !

    !

    T

    T

    [[

    [[

    [

    T

    deeXnx

    enxeX

    njj

    n

    njj

    )(

    2

    1)(

    )()(

    Copyright 2005. Shi Ping CUC

    The Family of Fourier Transform

    Aperiodic-DiscreteDTFT

  • 8/9/2019 Chap3-Discrete Fourier Transform

    7/83

    Copyright 2005. Shi Ping CUC

    The Family of Fourier Transform

    Periodic-DiscreteDFS (DFT)

    !

    !

    !

    !

    1

    0

    2

    1

    0

    2

    )(1

    )(

    )()(

    N

    k

    nkN

    j

    N

    n

    nkN

    j

    ekXN

    nx

    enxkX

    T

    T

  • 8/9/2019 Chap3-Discrete Fourier Transform

    8/83

    Copyright 2005. Shi Ping CUC

    The Family of Fourier Transform

    Summary

    Time function Frequency function

    Continuous and Aperiodic Aperiodic and Continuous

    Continuous and Periodic( ) Aperiodic and Discrete( )

    Discrete ( ) and Aperiodic Periodic( ) and Continuous

    Discrete ( ) and Periodic ( )

    Periodic( )

    and Discrete( )

    0T

    T

    T0

    T

    0

    0

    2

    T

    T!;

    Ts

    T2

    !;

    Ts

    T2!

    0

    0

    2

    T

    T!;

    return

  • 8/9/2019 Chap3-Discrete Fourier Transform

    9/83

    Copyright 2005. Shi Ping CUC

    The Discrete Fourier Series (DFS)

    Definition

    Periodic time functions can be synthesized as a linear

    combination of complex exponentials whose frequencies

    are multiples (or harmonics) of the fundamental frequency

    Periodic continuous-time function )()( rTtxtx !

    Periodic discrete-time function )()( rNnxnx !

    g

    g!

    !

    k

    ktT

    j

    ekXtx

    T2

    )()(

    !

    !1

    0

    2

    )(1

    )(N

    k

    knN

    j

    ekX

    N

    nx

    T

    fundamental

    frequencyt

    Tj

    e

    T2

    nN

    j

    e

    T2fundamentalfrequency

  • 8/9/2019 Chap3-Discrete Fourier Transform

    10/83

    Copyright 2005. Shi Ping CUC

    The Discrete Fourier Series (DFS)

    !

    !

    !

    ! elsewhere,0

    ,1

    1

    1112

    2

    1

    0

    2mNr

    e

    e

    Ne

    N rN

    j

    rNN

    jN

    n

    rnN

    j

    T

    TT

    )(1

    )(

    )(1

    )(

    1

    0

    )(1

    0

    1

    0

    1

    0

    1

    0

    rXeN

    kX

    eekXN

    enx

    N

    n

    nrkN

    jN

    k

    N

    n

    rnN

    jN

    k

    knN

    jN

    n

    rnN

    j

    !

    -

    !

    -

    !

    !

    !

    !

    !

    !

  • 8/9/2019 Chap3-Discrete Fourier Transform

    11/83

    Copyright 2005. Shi Ping CUC

    The Discrete Fourier Series (DFS)

    1

    0

    2

    )()(N

    n

    knN

    j

    enxkX

    T

    )()(

    )()(

    1

    0

    1

    0

    )(

    kXenx

    enxmNkX

    N

    n

    knN

    j

    N

    n

    nmNkN

    j

    !!

    !

    !

    !

    Because:

    The is a periodic sequence with fundamental

    period equal to N

    )(kX

  • 8/9/2019 Chap3-Discrete Fourier Transform

    12/83

    Copyright 2005. Shi Ping CUC

    The Discrete Fourier Series (DFS)

    !

    !

    !!

    !!

    !

    1

    0

    1

    0

    ~1]~[IDFS~

    ~]~[DFS~

    Let

    N

    k

    kN

    N

    n

    nk

    N

    j

    N

    WkN

    knx

    Wnxnxk

    eW NT

  • 8/9/2019 Chap3-Discrete Fourier Transform

    13/83

    Copyright 2005. Shi Ping CUC

    The Discrete Fourier Series (DFS)

    Relation to the z-transform

    ee

    !elsewhere,0

    10, Nnnxnx

    kN

    jez

    zXkX T2|)()(~!!

    The DFS represents N evenly spaced samples ofthe z-transform around the unit circle.

    )(~

    kX

    )(zX

    !

    !

    !!1

    0

    1

    0

    ))(()(~

    ,)()(2

    N

    n

    nkjN

    n

    n NenxkXznxzXT

  • 8/9/2019 Chap3-Discrete Fourier Transform

    14/83

    Copyright 2005. Shi Ping CUC

    The Discrete Fourier Series (DFS)

    Relation to the DTFT

    k

    j

    N

    n

    nkN

    jN

    n

    njj

    NeXkX

    enxkXenxeX

    T[

    [

    T

    [[

    2|)()(

    ~

    )()(~

    )()(1

    0

    21

    0

    !

    !

    !

    !

    !!

    ee

    !elsewhere,0

    10),(~

    )(Nnnx

    nx

    The DFS is obtained by evenly sampling the DTFT at

    intervals. It is called frequency resolution and represents the

    sampling intervalin the frequency domain.

    NT2

  • 8/9/2019 Chap3-Discrete Fourier Transform

    15/83

    Copyright 2005. Shi Ping CUC

    The Discrete Fourier Series (DFS)

    jIm[z]

    Re[z]

    0!k

    k

    j

    N

    ekX T[[

    2)()(~

    !!

    N

    T2

    N=8frequency resolution

  • 8/9/2019 Chap3-Discrete Fourier Transform

    16/83

    Copyright 2005. Shi Ping CUC

    The properties of DFS

    The Discrete Fourier Series (DFS)

    Linearity

    )(

    ~

    )(

    ~

    )](~

    )(~

    [DFS 11 kXbkXanxbnxa

    Shift of a sequence

    )(~

    )(~

    )](~[DFS kXekXWmnxmk

    Nj

    mk

    N

    T

    !!

    Modulation

    )(~

    )](~[DFS lkXnxWlnN

    !

  • 8/9/2019 Chap3-Discrete Fourier Transform

    17/83

    Copyright 2005. Shi Ping CUC

    The Discrete Fourier Series (DFS)

    Periodic convolution

    !

    !

    !

    !!

    !

    1

    012

    1

    0

    21

    21

    )(~)(~

    )(~)(~)](~

    [IDFS)(~then

    )(~

    )(~

    )(~if

    N

    m

    N

    m

    mnxmx

    mnxmxkYny

    kXkXkY

  • 8/9/2019 Chap3-Discrete Fourier Transform

    18/83

    Copyright 2005. Shi Ping CUC

    !

    !

    !

    !

    !

    !

    !

    !!

    -

    !

    -

    !

    !!

    1

    0

    12

    1

    0

    21

    1

    0

    1

    0

    )(21

    1

    0

    2

    1

    0

    1

    1

    0

    2121

    )(~)(~)(~)(~

    )(~1)(~

    )(~

    )(~1

    )(~

    )(~1

    )](~

    )(~

    [IDFS)(~

    N

    m

    N

    m

    N

    m

    N

    k

    kmnN

    N

    k

    nk

    N

    N

    m

    mk

    N

    N

    k

    nk

    N

    mnxmxmnxmx

    WkXN

    mx

    WkXWmxN

    WkXkX

    N

    kXkXny

    return

  • 8/9/2019 Chap3-Discrete Fourier Transform

    19/83

    Copyright 2005. Shi Ping CUC

    Introduction

    The Discrete Fourier Transform (DFT)

    The DFS provided us a mechanism fornumerically

    computingthe discrete-time Fourier transform. But most of the

    signals in practice are not periodic. They are likely to be of

    finite length.

    Theoretically, we can take care of this problem by defining a

    periodicsignal whoseprimaryshape is that of the finite length

    signal and then using the DFS on this periodic signal.

    Practically, we define a new transform called the Discrete

    Fourier Transform, which is the primary period of the DFS.

    This DFT is the ultimate numerically computable Fourier

    transform for arbitrary finite length sequences.

  • 8/9/2019 Chap3-Discrete Fourier Transform

    20/83

    Copyright 2005. Shi Ping CUC

    Finite-length sequence & periodic sequence

    The Discrete Fourier Transform (DFT)

    )(nx Finite-length sequence that has N samples

    )(~ nx periodic sequence with the period of N

    )()(~)(

    ,0

    10),(~)(

    nRnxnxelsewhere

    Nnnxnx

    N!

    ee!

    ))(()(~

    )()(~

    N

    r

    nxnx

    rNnxnx

    !

    ! g

    g!

    Window operation

    Periodic extension

  • 8/9/2019 Chap3-Discrete Fourier Transform

    21/83

    Copyright 2005. Shi Ping CUC

    The definition of DFT

    The Discrete Fourier Transform (DFT)

    10,)(1

    )]([ID T)(

    10,)()]([D T)(

    1

    0

    1

    0

    ee

    ee

    NnWkXN

    kXnx

    NkWnxnxkX

    N

    n

    nk

    N

    N

    n

    nk

    N

    )()(~)()(1

    )(

    )()(~)()()(

    1

    0

    1

    0

    nRnxnRWkXN

    nx

    kRkXkRWnxkX

    N

    N

    n

    N

    nk

    N

    N

    N

    n

    NnkN

    !!

    !!

    !

    !

    return

  • 8/9/2019 Chap3-Discrete Fourier Transform

    22/83

    Copyright 2005. Shi Ping CUC

    The Properties of DFT

    Linearity

    )()()]()([DFT 11 kbkanbxnax

    N3-point DFT, N3=max(N1,N2)

    Circular shift of a sequence

    )()]())(([ kXWRmkm

    !

    )())(()]([ kRlkXWl

    !

    Circular shift in the frequency domain

  • 8/9/2019 Chap3-Discrete Fourier Transform

    23/83

    Copyright 2005. Shi Ping CUC

    The Properties of DFT

    The sum of a sequence

    !!

    !!

    !!1

    0

    0

    1

    00

    )()()(N

    n

    N

    n

    n

    NnxWnxX

    The first sample of sequence

    !

    !1

    0

    )(1

    )0(N

    k

    kXN

    x

    )())(()]([

    )()]([

    kRkNNxnXDFT

    kXnxDFT

    NN!

    !

  • 8/9/2019 Chap3-Discrete Fourier Transform

    24/83

    Copyright 2005. Shi Ping CUC

    The Properties of DFT

    Circular convolution

    )()()())(()(

    )())(()()()(

    12

    1

    0

    12

    1

    0

    2121

    nxnxnRnxmx

    nRmnxmxnxnx

    N

    N

    m

    N

    N

    N

    m

    N

    !

    -

    !

    -

    !

    !

    !

    N

    N

    )()()]()([DFT2121

    kXkXnxnx !N

    )()(1

    )]()([DFT 2121 kXkXN

    nxnx ! N

    Multiplication

  • 8/9/2019 Chap3-Discrete Fourier Transform

    25/83

    Copyright 2005. Shi Ping CUC

    The Properties of DFT

    Circular correlation

    g

    g!

    g

    g! !! nnxy nymnxmnynxmr )(*)()(*)()(

    Linear correlation

    Circular correlation

    )()(*))((

    )())((*)()(

    1

    0

    1

    0

    mRymx

    mRmyxmr

    N

    N

    N

    N

    N

    Nxy

    !

    !

    !

    !

  • 8/9/2019 Chap3-Discrete Fourier Transform

    26/83

    Copyright 2005. Shi Ping CUC

    The Properties of DFT

    )()(*))((

    )())((*)(

    )]([IDFT)(

    )()()(

    1

    0

    1

    0

    *

    mRnymnx

    mRmnynx

    kRmrthen

    kYkXkRif

    N

    N

    n

    N

    N

    N

    n

    N

    xyxy

    xy

    !

    !

    !

    !

    !

    !

  • 8/9/2019 Chap3-Discrete Fourier Transform

    27/83

    Copyright 2005. Shi Ping CUC

    The Properties of DFT

    Parsevals theorem

    !

    !

    !1

    0

    *1

    0

    * )()(1

    )()(N

    k

    N

    n

    kYkXN

    nynx

    1

    0

    21

    0

    2

    1

    0

    *

    1

    0

    *

    )(1

    )(

    )()(1

    )()(then

    )()(let

    NN

    n

    NN

    n

    Nnx

    N

    nxnx

    nynx

  • 8/9/2019 Chap3-Discrete Fourier Transform

    28/83

    Copyright 2005. Shi Ping CUC

    The Properties of DFT

    Conjugate symmetry properties of DFT

    and)(nxep )(nxop

    Let be a N-point sequence)(nx Nnxnx ))(()(~ !

    ]))(())(([2

    1

    )](

    ~

    )(

    ~

    [2

    1

    )(

    ~

    ]))(())(([2

    1)](~)(~[

    2

    1)(~

    NNo

    NNe

    nNx

    nx

    nx

    nx

    nx

    nNxnxnxnxnx

    !!

    !!

    It can be proved that

    )(

    ~

    )(

    ~

    )(~)(~

    *

    *

    nx

    nx

    nxnx

    oo

    ee

  • 8/9/2019 Chap3-Discrete Fourier Transform

    29/83

    Copyright 2005. Shi Ping CUC

    The Properties of DFT

    ? A

    ? A )())(())((2

    1

    )()(~)(

    )())(())((2

    1

    )()(~)(

    nRnNn

    nRnn

    nRnNn

    nRnn

    NNN

    Noop

    NNN

    Np

    !

    !

    !

    !

    Circular

    conjugate

    symmetric

    component

    Circular

    conjugate

    antisymmetriccomponent

  • 8/9/2019 Chap3-Discrete Fourier Transform

    30/83

    Copyright 2005. Shi Ping CUC

    The Properties of DFT

    )()()( nxnxnxopep

    !

    )())(()(

    )())(()(*

    *

    nRnNxnx

    nRnNxnx

    NNopop

    NNepep

    !

    !

  • 8/9/2019 Chap3-Discrete Fourier Transform

    31/83

    Copyright 2005. Shi Ping CUC

    The Properties of DFT

    )()()(kXkXkX

    opp

    !

    )())(()(

    )())(()(

    *

    *

    kRkNXkX

    kRkNXkX

    NNopop

    NNepep

    and)(kXep )(kXop

  • 8/9/2019 Chap3-Discrete Fourier Transform

    32/83

    Copyright 2005. Shi Ping CUC

    The Properties of DFT

    )]())((Im[)](Im[

    )]())((Re[)](Re[

    kRkNXkX

    kRkNXkX

    NNepep

    NNepep

    !

    !

    )]())((Im[)](Im[

    )]())((Re[)](Re[

    kRkNXkX

    kRkNXkX

    NNopop

    NNopop

    !

    !

  • 8/9/2019 Chap3-Discrete Fourier Transform

    33/83

    Copyright 2005. Shi Ping CUC

    The Properties of DFT

    )())(()(then

    )())(()(if

    kRkNXkX

    nRnNxnx

    NN

    NN

    !

    !

    Circular even sequences

    Circular odd sequences

    )())(()(then)())(()(ifkRkNXkX

    nRnNxnx

    NN

    NN

    !

    !

  • 8/9/2019 Chap3-Discrete Fourier Transform

    34/83

    Copyright 2005. Shi Ping CUC

    The Properties of DFT

    )()())((

    )())(()]([DFT**

    **

    kNXkRkNX

    kRkXnx

    NN

    NN

    !!

    !

    Conjugate sequences

    )()]())(([D

    )]())(([D

    **

    *

    knRnNx

    nRnx

    NN

    NN

    !!

  • 8/9/2019 Chap3-Discrete Fourier Transform

    35/83

    Copyright 2005. Shi Ping CUC

    The Properties of DFT

    _ a

    ? A )())(())((21

    )()](Re[DFT

    * kRkNXkX

    kXnx

    NNN

    ep

    !

    !

    Complex-value sequences

    _ a? A )())(())((

    2

    1)()](Im[DFT

    * kRkNXkX

    kXnxj

    NNN

    op

    !

    !

  • 8/9/2019 Chap3-Discrete Fourier Transform

    36/83

    Copyright 2005. Shi Ping CUC

    The Properties of DFT

    !

    !

    )(]))(())(([2

    1

    DFT

    )](Re[)]([DFT

    *

    nnn

    kXnp

    !

    !

    )(]))(())(([2

    1DFT

    )](Im[)]([DFT

    *nRnNxnx

    kXjnx

    NNN

    op

  • 8/9/2019 Chap3-Discrete Fourier Transform

    37/83

    Copyright 2005. Shi Ping CUC

    The Properties of DFT

    )())(()(then

    sequencevalue-realis)(i*

    kRkNXkX

    nx

    NN!

    Real-value sequences

    Imaginary-value sequences

    )())(()(tpartimagi aryaso ly)(if

    *kRkNXkX

    nx

    NN!

  • 8/9/2019 Chap3-Discrete Fourier Transform

    38/83

    Copyright 2005. Shi Ping CUC

    The Properties of DFT

    Summary

    )()()(

    )](Im[)](Re[)(

    kXkXkX

    nxjnxnx

    opep !

    !

    DDD

    )](Im[)](Re[)(

    )()()(

    kXjkXkX

    nxnxnx opep

    !

    !

    DDD

    example

  • 8/9/2019 Chap3-Discrete Fourier Transform

    39/83

    Copyright 2005. Shi Ping CUC

    The Properties of DFT

    Linear convolution & circular convolution

    !

    g

    g!

    !!

    !

    1

    0

    2121

    211

    )()()()(

    )()()(N

    mm

    l

    mnxmxmnxmx

    nxnxny

    Linear convolution

    )(1 nx

    )(nx

    N1 point sequence, 0n N1-1

    N2 point sequence, 0n N2-1

    )(nyl L point sequence, L= N1+N2-1

  • 8/9/2019 Chap3-Discrete Fourier Transform

    40/83

    Copyright 2005. Shi Ping CUC

    The Properties of DFT

    Circular convolution

    ee

    ee!

    1,0

    10),()(

    1

    11

    1LnN

    Nnnn

    We have to make both and L-point

    sequences by padding an appropriate number of zeros

    in order to make L point circular convolution.

    )(nx )(2 nx

    ee

    ee!

    ,

    ),()(

    LnN

    Nnnxnx

  • 8/9/2019 Chap3-Discrete Fourier Transform

    41/83

    Copyright 2005. Shi Ping CUC

    The Properties of DFT

    )()(

    )()()(

    )()()(

    )())(()()()()(

    2

    1

    0

    1

    1

    0

    21

    1

    0

    2121

    nRrLny

    nRmrLnxmx

    nRmrLnxmx

    nRmnxmxnxnxny

    L

    r

    l

    L

    r

    L

    m

    L

    L

    m r

    L

    L

    m

    Lc

    -

    !

    -

    !

    - !

    -

    !!

    g

    g!

    g

    g!

    !

    !

    g

    g!

    !

    L

  • 8/9/2019 Chap3-Discrete Fourier Transform

    42/83

    Copyright 2005. Shi Ping CUC

    The Properties of DFT

    )()()( nRrnynyr

    lc

    -

    !

    g

    g!

    )()()()(isthat

    )()(then1if

    2121

    21

    nxnxnxnx

    nynyNNL

    lc

    !

    !u

    L

    return

  • 8/9/2019 Chap3-Discrete Fourier Transform

    43/83

    Copyright 2005. Shi Ping CUC

    The Sampling Theorem in Frequency Domain

    Sampling in frequency domain

    g

    g!!

    !!

    m

    km

    NWzWmxzXkX k

    N

    )(|)()(~

    g

    g!

    g

    g!

    !

    !

    g

    g!

    !

    !

    -

    !

    -

    !

    !!

    rm

    N

    mk

    N

    N

    k

    kn

    N

    m

    km

    N

    N

    k

    kn

    NN

    rNnxWN

    mx

    WWmxN

    WkXN

    kXnx

    )()(

    )(

    )()]([IDFS)(

    )(

  • 8/9/2019 Chap3-Discrete Fourier Transform

    44/83

    Copyright 2005. Shi Ping CUC

    The Sampling Theorem in Frequency Domain

    g

    g!

    !r

    NrNnxnx )()(~

    Frequency Sampling TheoremForM point finite duration sequence, if the frequency

    sampling number N satisfy:

    MN uthen

    )()()(~)( nxnRnxnxNNN

    !!

  • 8/9/2019 Chap3-Discrete Fourier Transform

    45/83

    Copyright 2005. Shi Ping CUC

    The Sampling Theorem in Frequency Domain

    Interpolation formula of )(zX

    !

    !

    !

    !

    !

    !

    !

    !

    !

    !

    !

    -

    !

    -

    !

    -

    !!

    1

    01

    1

    01

    1

    0

    1

    0

    11

    0

    1

    0

    1

    0

    1

    0

    1

    0

    1)(1

    11)(1

    )(1

    )(1

    )(1

    )()(

    N

    kk

    N

    NN

    kk

    N

    NNk

    N

    N

    k

    N

    n

    nk

    N

    N

    k

    N

    n

    nnk

    N

    N

    n

    n

    N

    k

    nk

    N

    N

    n

    n

    zWkX

    N

    zzWzWkX

    N

    zWkXN

    zWkXN

    zWkXN

    znxzX

  • 8/9/2019 Chap3-Discrete Fourier Transform

    46/83

    Copyright 2005. Shi Ping CUC

    The Sampling Theorem in Frequency Domain

    1

    1

    0

    1

    01

    1

    11)(

    )()(1

    )(1)(

    !

    !

    !*

    *!

    !

    zW

    z

    Nz

    zkXzW

    kX

    N

    zzX

    k

    N

    N

    k

    N

    k

    k

    N

    kk

    N

    N

    Interpolation

    function

  • 8/9/2019 Chap3-Discrete Fourier Transform

    47/83

    Copyright 2005. Shi Ping CUC

    The Sampling Theorem in Frequency Domain

    !

    !

    *!*!1

    0

    1

    0

    )2

    ()()()()(N

    k

    N

    k

    j

    K

    jwk

    NkXekXeX

    T

    [[

    2

    1

    2

    2

    sin

    sin)(

    !*

    NjN

    eN

    [

    [

    [

    [

    Interpolation

    function

    Interpolation formula of )( [jeX

    return

  • 8/9/2019 Chap3-Discrete Fourier Transform

    48/83

    Copyright 2005. Shi Ping CUC

    Approximating to FT (FS) with DFT (DFS)

    Approximating to FT of continuous-time aperiodic

    signal with DFT

    g

    g;

    g

    g

    ;

    ;;!

    !;

    dejXtx

    dtetxjX

    tj

    tj

    )(21)(

    )()(

    T

    CTFT

  • 8/9/2019 Chap3-Discrete Fourier Transform

    49/83

    Copyright 2005. Shi Ping CUC

    g

    g!

    g

    gppp

    n

    TdtTdtnTt ,,

    Sampling in time domain

    g

    g!

    ;g

    g

    ; }!;n

    nTjtjenTxTdtetxjX )()()(

    ;;

    g

    g

    ;

    ;;}

    ;;!S

    dejXntx

    dejXtx

    nTj

    tj

    0)(

    2

    1)(

    )(2

    1

    )(

    T

    T

    Tfss

    T

    T

    22 !!;

    Approximating to FT (FS) with DFT (DFS)

  • 8/9/2019 Chap3-Discrete Fourier Transform

    50/83

    Copyright 2005. Shi Ping CUC

    Truncation in time domain

    )1~0(:,,)~0(:00 ! NNTTTt

    !

    ;};1

    0

    )()(N

    n

    nTjenTxTjX

    S

    dejXnTxnTj

    0)(

    21)(T

    Approximating to FT (FS) with DFT (DFS)

  • 8/9/2019 Chap3-Discrete Fourier Transform

    51/83

    Copyright 2005. Shi Ping CUC

    Sampling in frequency domain

    NT

    TT

    TFNT

    fN

    FT

    ddk

    s

    N

    n

    s

    TT

    TT

    22

    22,1

    ,,

    0

    0

    0

    00

    0

    0

    1

    0

    00

    00

    !!;

    !!;!!!

    ;p;;p;;p;

    !

    ;

    )]([T

    )()()(1

    0

    21

    0

    0

    0

    nxT

    enxTenTxTjkXN

    n

    nkN

    jN

    n

    nTjk

    !

    !};

    !

    !

    ;T

    Approximating to FT (FS) with DFT (DFS)

  • 8/9/2019 Chap3-Discrete Fourier Transform

    52/83

    Copyright 2005. Shi Ping CUC

    )]([I T1

    )]([I T

    )(1

    )(1

    )(2

    )(

    00

    1

    0

    2

    0

    1

    0

    2

    00

    1

    0

    0

    0 0

    ;!;!

    ;!

    ;!

    ;;

    }

    !

    !

    !

    ;

    jkXT

    jkXf

    ejkXN

    f

    ejkXN

    NF

    ejkXnTx

    s

    N

    k

    nkN

    j

    s

    N

    k

    nkN

    j

    N

    k

    nTjk

    T

    T

    T

    demo

    Approximating to FT (FS) with DFT (DFS)

  • 8/9/2019 Chap3-Discrete Fourier Transform

    53/83

    Copyright 2005. Shi Ping CUC

    Approximating to FS of continuous-time periodic

    signal with DFS

    g

    g!

    ;

    ;

    ;!

    !;

    k

    tjk

    Ttjk

    ejkXtx

    dtetxT

    jkX

    0

    00

    )()(

    )(1

    )(

    0

    00

    0

    0

    00

    22

    TF

    TT !!;

    Approximating to FT (FS) with DFT (DFS)

  • 8/9/2019 Chap3-Discrete Fourier Transform

    54/83

    Copyright 2005. Shi Ping CUC

    !

    !1

    00

    0

    0

    ,,N

    n

    T

    TdNTTTdnTt

    Sampling in time domain

    )]([1

    )(1

    )()(1

    0

    21

    00

    00

    nxN

    enxN

    enTxT

    TjkX

    N

    n

    nkN

    jN

    n

    nTjk

    !

    !};

    !

    !

    ;T

    Approximating to FT (FS) with DFT (DFS)

  • 8/9/2019 Chap3-Discrete Fourier Transform

    55/83

    Copyright 2005. Shi Ping CUC

    Truncating in frequency domain

    ),(:let,, !@! NkNFfNTT s3

    )]([IDFS)(1

    )()()(

    )()(

    0

    1

    0

    2

    0

    1

    0

    2

    0

    1

    0

    0

    0

    0

    0

    ;!;!

    ;!;}

    ;!

    !

    !

    !

    ;

    g

    g!

    ;

    jkXNejkXN

    N

    ejkXejkXnTx

    ejkXtx

    N

    k

    nkN

    j

    N

    k

    nkN

    jN

    k

    nTjk

    k

    tjk

    T

    T

    Approximating to FT (FS) with DFT (DFS)

  • 8/9/2019 Chap3-Discrete Fourier Transform

    56/83

    Some problems

    Aliasing

    Otherwise, the aliasing will occur in frequency domainhs

    hs

    ffff

    2,2 !"Sampling in time domain:

    Sampling in frequency domain:

    0

    0

    1

    FT !

    Period in time domain0T Frequency resolution0F

    NT

    T

    F

    fs!!

    0

    0

    and

    is contradictoryhf 0F

    Approximating to FT (FS) with DFT (DFS)

  • 8/9/2019 Chap3-Discrete Fourier Transform

    57/83

    Spectrum leakage

    sequencelength-finite),()()(

    sequencelength-infinite),(

    12

    1

    nRnxnx

    nx

    N!

    )()()(j

    R

    jjeWeXeX !

    Spectrum extension (leakage)

    Spectrum aliasing

    Approximating to FT (FS) with DFT (DFS)

    demo

  • 8/9/2019 Chap3-Discrete Fourier Transform

    58/83

    Copyright 2005. Shi Ping CUC

    Fence effect

    N

    fF

    f

    F

    fN

    s

    ss

    !!;

    !! 000

    0 ,22 TT

    [

    Frequency resolution

    00

    11

    TT

    fF

    s!!!

    demo

    Approximating to FT (FS) with DFT (DFS)

  • 8/9/2019 Chap3-Discrete Fourier Transform

    59/83

    Copyright 2005. Shi Ping CUC

    Comments

    return

    demo

    Zero-padding is an operation in which more zeros are

    appended to the original sequence. It can provides closely

    spaced samples of the DFT of the original sequence.

    The zero-padding gives us a high-density spectrum and

    provides a better displayed version for plotting. But it does

    not give us a high-resolution spectrum because no new

    information is added.

    To get a high-resolution spectrum, one has to obtainmore data from the experiment or observation.

    example

    Approximating to FT (FS) with DFT (DFS)

  • 8/9/2019 Chap3-Discrete Fourier Transform

    60/83

    Copyright 2005. Shi Ping CUC

    Summary

    return

    The frequency representations of x(n)

    )(nx

    )(kX

    )(zX

    )([jeX

    Time

    sequence

    z-transform of x(n)

    Complex

    frequency domain

    DTFT of x(n)

    Frequency

    domainDFT of x(n)

    Discrete frequencydomain

    ZT

    DTFTDFT

    kN

    T2!

    [jz! kjz

    2

    !

    interpolation

    interpolation

  • 8/9/2019 Chap3-Discrete Fourier Transform

    61/83

    Copyright 2005. Shi Ping CUC

    Illustration of the four Fourier transforms

    Discrete Fourier SeriesSignals that are discrete and

    periodic

    DTFTSignals that are discrete andaperiodic

    Fourier SeriesSignals that are continuousand periodic

    FourierTransformSignals that are continuousand aperiodic

    ~

  • 8/9/2019 Chap3-Discrete Fourier Transform

    62/83

    Copyright 2005. Shi Ping CUC

    )(~ ny

    n0 1 2 3 4 5 6

    0!n m

    )(~1 mx

    0

    m

    )(~2 mnx

    0

    1!n2!n

    3!n4!n5!n

    6!n

    return

    ? A1

  • 8/9/2019 Chap3-Discrete Fourier Transform

    63/83

    Copyright 2005. Shi Ping CUCreturn

    ? A)())(()(2

    1)( nRnNxnxnx NNep !

    n

    )(nx

    0

    n

    N

    nNx

    ))((

    0 55

    n

    )(nxep

    05

  • 8/9/2019 Chap3-Discrete Fourier Transform

    64/83

    Copyright 2005. Shi Ping CUCreturn

    n

    )(nxep

    0 5

    )())(()( * nRnNxnx NNepep !

    n

    epn))((

    0 5

    )(nRN

  • 8/9/2019 Chap3-Discrete Fourier Transform

    65/83

    Copyright 2005. Shi Ping CUC

    0 1 2 3 4 5 6 7 8 9 10

    0

    5

    10

    O r ig i n a l s e q u e n c e

    n

    x

    (n

    )

    0 1 2 3 4 5 6 7 8 9 10

    0

    5

    10

    C i r c u l ar c o n j u g a t e s y m m e t r ic c o m p o n e n t

    n

    xep(n

    )

    0 1 2 3 4 5 6 7 8 9 10 -4

    -2

    0

    2

    4C i r c u l ar c o n j u g a t e an t is y m m e t r ic c o m p o n e n t

    n

    xo

    p

    (n

    )

    return

    )()8.(1 11 nRn

    v

  • 8/9/2019 Chap3-Discrete Fourier Transform

    66/83

    Copyright 2005. Shi Ping CUC

    0 1 2 3 4 5 6 7 8 9 1 0

    0

    5

    1 0

    Ci

    l

    v

    0 1 2 3 4 5 6 7 8 9 1 0

    0

    2 0

    4 0

    T h e D FT

    f

    k

    0 1 2 3 4 5 6 7 8 9 1 0

    0

    2 0

    4 0

    k

    return

    )(kX

    )())(( nkNX NN

  • 8/9/2019 Chap3-Discrete Fourier Transform

    67/83

    Copyright 2005. Shi Ping CUC

    0 1 2 3 4 5 6 7 8 9 1 0 -4

    -2

    0

    2

    4Ci

    l

    d d

    e q u e n c e

    n )

    n

    0 1 2 3 4 5 6 7 8 9 1 0

    -1 0

    0

    1 0

    T h e i !

    in a r y "

    a rt

    f D FT [x (n ) ]

    k

    0 1 2 3 4 5 6 7 8 9 1 0

    -1 0

    0

    1 0

    k

    return

    )(kXd

    )())(( nRkNXNN

    d

  • 8/9/2019 Chap3-Discrete Fourier Transform

    68/83

    Copyright 2005. Shi Ping CUCreturn

    )())(()(*

    kkNXkXNN

    !

    ? Anumberrealais)0(

    )0()())(()0( *0

    *

    X

    XkRkNXXk

    NN

    @

    !!!

    ? A

    numberrealais)2

    (

    )

    2

    ()())(()

    2

    (

    evenisif

    *

    2

    *

    NX

    NXRNX

    NX

    N

    NNN

    @

    !!!

  • 8/9/2019 Chap3-Discrete Fourier Transform

    69/83

    Copyright 2005. Shi Ping CUCreturn

    )())(()(*

    kRkNXkXNN

    !

    numb rimaginaryanis)0(

    )0()())(()0( *0

    *

    X

    XRNXXNN

    @

    !!!

    ? A

    numberimginaryanis)2

    (

    )

    2

    ()())(()

    2

    (

    evenisif

    *

    2

    *

    NX

    NXkRkNX

    NX

    N

    Nk

    NN

    @

  • 8/9/2019 Chap3-Discrete Fourier Transform

    70/83

    Copyright 2005. Shi Ping CUCreturn

    )(,)( 21 nxnx N-point real-value sequences

    )]([DFT)()],([DFT)(2211

    nxnx !!

    )()()]([)]([

    )]()([)]([)()()()(

    kjkj

    jykYjy

    !!

    !!!

    _ a )())(()(2

    1)()](Re[DFT)(1 kRkNYkYkYykX NNep !!!

    _ a ? A )())(()(2

    1)(

    1)](Im[DFT)(2 kRkNYkY

    jkY

    jnykX NNop !!!

  • 8/9/2019 Chap3-Discrete Fourier Transform

    71/83

    Copyright 2005. Shi Ping CUC

    0 1 2 3 4 5 6 7 8 9

    0

    2

    4

    68

    1 0

    1 2

    L in e a r c o n vo lut io n

    n 0 1 2 3 4 5 6 7 8 9

    0

    2

    4

    68

    1 0

    1 2

    Circ ula rc o n vo lut io n N = 6

    n

    0 1 2 3 4 5 6 7 8 9

    0

    2

    46

    8

    1 0

    1 2

    Circ ula rc o n vo lut io n N = 7

    n 0 1 2 3 4 5 6 7 8 9

    0

    2

    46

    8

    1 0

    1 2

    Circ ula rc o n vo lut io n N = 5

    n

    return

    ],2,3,2,1[)(],2,2,1[)( 21 !! nn

    )([*

  • 8/9/2019 Chap3-Discrete Fourier Transform

    72/83

    Copyright 2005. Shi Ping CUC

    -1 -0 .8 -0 .6 -0 .4 -0 .2 0 0 .2 0 .4 0 .6 0 .8 10

    0 .2

    0 .4

    0 .6

    0 .8

    1

    M a g n i tude R e s p o n s e , N = 8

    fr e q u e n c y in p iu n i ts

    -1 -0 .8 -0 .6 -0 .4 -0 .2 0 0 .2 0 .4 0 .6 0 .8 1-1

    -0.5

    0

    0 .5

    1P h a s e R e s p o n s e

    fr e q u e n c y in p iu n i ts

    p

    i

    )([*

    return

    N

    T2N

    T4

  • 8/9/2019 Chap3-Discrete Fourier Transform

    73/83

    Copyright 2005. Shi Ping CUC

    0 1 2 3 4 5 6 7 0

    1

    2

    3

    4

    5

    6

    X (k),N = 8

    k

    0 0 .1 0 .2 0 .3 0 .4 0 .5 0 .6 0 .7 0 .8 0 .9 1

    0

    2

    4

    fr e q u e n c y in p iu n i ts

    return

    )()0( [*X )()(N

    X T[*

    )4

    ()2(N

    T[ *

    )6()3(N

    X T[ *

    ttx )80(0)( v )( ;jX

  • 8/9/2019 Chap3-Discrete Fourier Transform

    74/83

    Copyright 2005. Shi Ping CUC

    0 5 1 0 1 5 2 0 2 5 0

    2

    4

    6

    8

    1 0

    t

    -1 -0 .5 0 0 .5 10

    1 0

    2 0

    3 0

    4 0

    5 0

    ra d

    0 5 1 0 1 5 2 0 2 5 0

    2

    4

    6

    8

    1 0

    n

    -2 -1 0 1 20

    1 0

    2 0

    3 0

    4 0

    5 0

    p i

    atx )8.0(0)( v! )( ;jXa

    FT

    DTFT

    )(nx )([jeX

    )()( R )()([[ jj

    eReX

  • 8/9/2019 Chap3-Discrete Fourier Transform

    75/83

    Copyright 2005. Shi Ping CUC

    -1 0 -5 0 5 1 0 0

    2

    4

    6

    8

    1 0

    n

    -2 -1 0 1 20

    1 0

    2 0

    3 0

    4 0

    5 0

    p i

    -1 0 0 1 0 0

    2

    4

    6

    8

    1 0

    n

    -1 0 0 1 0 0

    1 0

    2 0

    3 0

    4 0

    5 0

    k

    )()( nRnx )()(jj

    eReX

    )(~ nxN )(~

    kN

    DTFT

    DF

    )( )(kX

  • 8/9/2019 Chap3-Discrete Fourier Transform

    76/83

    Copyright 2005. Shi Ping CUC

    -1 0 0 1 0 0

    2

    4

    6

    8

    1 0

    n

    -1 0 0 1 0 0

    1 0

    2 0

    3 0

    4 0

    5 0

    k

    return

    )(nxN

    )(kXN

    DFT

    )(1 nx )([j

    eX

  • 8/9/2019 Chap3-Discrete Fourier Transform

    77/83

    Copyright 2005. Shi Ping CUCreturn

    [0

    )([j

    eR

    [0

    )(2[j

    X

    n

    )(nR

    0

    n

    )(nx

    0

    n

    )(1

    0[0

    )(eX

  • 8/9/2019 Chap3-Discrete Fourier Transform

    78/83

    Copyright 2005. Shi Ping CUC

    0 0 .2 0 .4 0 .6 0 .8 1 1 .2 1 .4 1 .6 1 .8 20

    2

    4

    0 0 .2 0 .4 0 .6 0 .8 1 1 .2 1 .4 1 .6 1 .8 20

    2

    4

    0 0 .2 0 .4 0 .6 0 .8 1 1 .2 1 .4 1 .6 1 .8 20

    2

    4

    p i

    p i

    p i

    DTFTDFT

    DTFT

    DFT

    DTFT

    DFT

    return

    ],,,[)( !nx

    ],,,,1,1,1,1[)( !nx

    ],,,,,,,,,,,,1,1,1,1[)( !nx

    s ig n a l (n) 0 < n < 1 9

    )50cos()480cos()( TT

  • 8/9/2019 Chap3-Discrete Fourier Transform

    79/83

    Copyright 2005. Shi Ping CUC

    0 2 4 6 8 1 0 1 2 1 4 1 6 1 8 2 0 -2

    -1

    0

    1

    2s ig n a lx(n),0

  • 8/9/2019 Chap3-Discrete Fourier Transform

    80/83

    Copyright 2005. Shi Ping CUC

    0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 0 1 0 0 -2

    -1

    0

    1

    2s ig n a lx(n),0

  • 8/9/2019 Chap3-Discrete Fourier Transform

    81/83

    Copyright 2005. Shi Ping CUC

    0 1 0 2 0 3 0 4 0 5 0 6 0 7 0 8 0 9 0 1 0 0 -2

    -1

    0

    1

    2s ig n a lx(n),0

  • 8/9/2019 Chap3-Discrete Fourier Transform

    82/83

    Copyright 2005. Shi Ping CUC

    0 5 0 1 0 0 1 5 0 2 0 0 2 5 0 3 0 0 3 5 0 4 0 0 -2

    -1

    0

    1

    2s ig n a lx(n),0

  • 8/9/2019 Chap3-Discrete Fourier Transform

    83/83

    Suppose kHz4Hz,100 ee hfF

    Determine N,,0

    TT

    Solutions

    FT .!u!

    msThs

    125.01042

    1

    2

    113

    !vv

    !e!

    102422

    80010125.0

    1.0

    10

    3

    0

    !!!

    !vu!

    mN

    T

    TN