61
 Chapter 2 Discrete-Time Signals and Systems Der-Feng Tseng Department of Electrical Engineering National Taiwan University of Science and Technology (through the courtesy of Prof. Peng-Hua Wang of National T aipei University) February 19, 2015 Der-Feng Tseng  (NTUST)  DSP Chapter 2  Fe br uary 19, 2015 1 / 61

1497-Chapter2

Embed Size (px)

DESCRIPTION

1497-Chapter2

Citation preview

  • Chapter 2Discrete-Time Signals and Systems

    Der-Feng Tseng

    Department of Electrical EngineeringNational Taiwan University of Science and Technology

    (through the courtesy of Prof. Peng-Hua Wang of National Taipei University)

    February 19, 2015

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 1 / 61

  • Outline

    1 2.1 Discrete-Time Signals: Sequences

    2 2.2 Discrete-Time Systems

    3 2.3 Linear Time-Invariant Systems

    4 2.4 Properties of Linear Time-Invariant Systems

    5 2.5 Linear Constant-Coefficient Difference Equations

    6 2.6 Frequency-Domain Representation

    7 2.7 Representation of Sequences by Fourier Transform

    8 2.8 Symmetry Properties of the Fourier Transform

    9 2.9 Fourier Transform Theorems

    10 2.10 Discrete-Time Random Random Signals

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 2 / 61

  • 2.1 Discrete-Time Signals: Sequences

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 3 / 61

  • Definition

    A sequence of number x, in which the nth number in the sequence isdenoted x[n] is written as x = {x[n]}, < n

  • Basic sequences 1/3

    The unit sample sequence is defined by

    [n] =

    {0, n 6= 0

    1, n = 0.

    Any sequence can be expressed as a sum of scaled, delayed impulse.

    The unit step sequence is defined by

    u[n] =

    {1, n 0

    0, n < 0.

    The exponential sequence is x[n] = An.

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 5 / 61

  • Basic sequences 2/3

    Any sequence can be expressed as a sum of scaled, delayed impulse

    p[n] =

    k=

    p[k][n k]

    u[n] can be expressed as a sum of delayed impulse

    u[n] =

    k=0

    [n k], or u[n] =

    nk=

    [k]

    Example 2.1 An exponential sequence x[n] that is zero for n < 0 can bewritten as x[n] = Anu[n].

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 6 / 61

  • Basic sequences 3/3

    The complex exponential sequence is x[n] = Aej(0n+).

    The sinusoidal sequence is x[n] = A cos(0n+ ) where 0 is call thefrequency and is called the phase.

    Difference between continuous and discrete complex exponentials

    fi(t) = ejit : f1(t) = f2(t) 1 = 2

    xi[n] = ejin : x1[n] = x2[n] 1 2 = 2r

    where i = 1, 2 and r is an integer.

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 7 / 61

  • Basic Operations

    Delay/shift: x[n n0] is delayed sequence of x[n] where n0 is aninteger.

    Periodic sequence: x[n] = x[n+N ] where N is the period.

    Difference between continuous and discrete complex exponentials

    f(t) = ej0t is periodic with period 2/0

    x[n] = ej0n is periodic with period N if 0N = 2k

    where k is an integer.

    Example 2.2 What is the period of x1[n] = cos(n/4), x2[n] =cos(3n/8), and x3[n] = cos(n)?

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 8 / 61

  • 2.2 Discrete-Time Systems

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 9 / 61

  • Definition

    A system is defined as a transformation or operator that maps aninput x[n] to an output y[n].

    y[n] = T{x[n]}

    Example 2.3 The ideal delay system is defined by y[n] = x[n nd] wherend is a positive integer. If nd < 0, this is a time-advancesystem.

    Example 2.4 The moving average system is defined by

    y[n] = (x[n+M1] + x[n+M1 1] + + x[n]

    + x[n 1] + + x[nM2]) /(M1 +M2 + 1)

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 10 / 61

  • Properties 1/4

    Memoryless: y[n] depends on only x[n], not x[n n0] for n0 6= 0.

    Example 2.3 y[n] = x[n nd] is not memoryless unless nd = 0.

    Example 2.4 The moving-average system is not memoryless unlessM1 = M2 = 0.

    Example 2.5 y[n] = (x[n])2 is memoryless.

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 11 / 61

  • Properties 2/4

    Additivity: T{x1[n] + x2[n]} = T{x1[n]}+ T{x2[n]}.

    Homogeneity/scaling: T{ax1[n]} = aT{x1[n]}.

    A system with additivity but not homogeneity: T {x[n]} = {x[n]}

    Linearity/superposition:T{ax1[n] + bx2[n]} = aT{x1[n]}+ bT{x2[n]}.

    Example 2.6 The accumulator is linear

    y[n] = x[n] + x[n 1] + =n

    k=

    x[k] =k=0

    x[n k]

    Example 2.7 w[n] = log10(|x[n]|) is nonlinear.

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 12 / 61

  • Properties 3/4

    Time-invariant: If y[n] = T{x[n]} then y[n n0] = T{x[n n0]}.

    Example 2.8 The accumulator is time-invariant.

    Example 2.9 The compressor y[n] = x[Mn] is not time-invariant where Mis a positive integer.

    Causality: y[n] = T{x[n]} depends on only x[n], x[n 1], . . .

    Example 2.10 Forward difference y[n] = x[n+ 1] x[n] is not causal.Backward difference y[n] = x[n] x[n 1] is causal.

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 13 / 61

  • Properties 4/4

    BIBO stability: Every bounded input sequence produce a boundedoutput sequence.

    Example 2.11 The accumulator is not stable.

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 14 / 61

  • 2.3 Linear Time-Invariant Systems

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 15 / 61

  • LTI Systems

    Linear time-invariant (LTI) system.

    If a system y[n] = T{x[n]} is LTI, then it can be characterized byh[n] = T{[n]} in the sense

    y[n] =

    k=

    x[k]h[n k] , x[n] h[n]

    The above equation is called the convolution sum. h[n] is called theimpulse response of the system.

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 16 / 61

  • Proof of Convolution

    y[n] = T{x[n]}

    = T

    {

    k=

    x[k][n k]

    }(representation)

    =

    k=

    x[k]T{[n k]} (linearity)

    =

    k=

    x[k]h[n k] (time-invariant)

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 17 / 61

  • Example 2.13

    h[n] = u[n] u[nN ] = 1, 0 n N 1.

    x[n] = anu[n]

    h[n] x[n] =

    k=

    x[k]h[n k] =

    k=

    h[k]x[n k]

    =

    N1k=0

    anku[n k]

    =

    0, n < 0,1an+1

    1a , 0 n N 1,

    anN+1(1aN

    1a

    ), n > N 1.

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 18 / 61

  • Convolution by Tabular

    Compute {x[2], x[1], x[0], x[1]} {h[1], h[0], h[1]}.

    x2 x1 x0 x1 h1 h0 h1

    h1x2 h1x1 h1x0 h1x1h0x2 h0x1 h0x0 h0x1

    + h1x2 h1x1 h1x0 h1x1y3 y2 y1 y0 y1 y2

    Prob. 2.22a Compute {0, 1} {2, 1}.

    Prob. 2.22b Compute {2,1} {1, 2, 1}.

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 19 / 61

  • 2.4 Properties of Linear Time-Invariant Systems

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 20 / 61

  • Properties of LTI Systems

    Commutative: x[n] h[n] = h[n] x[n]

    Distributive: x[n] (h1[n] + h2[n]) = x[n] h1[n] + x[n] h2[n]

    Cascade: h1[n] h2[n]

    Parallel connection: h1[n] + h2[n]

    Finite-duration impulse response (FIR) systems have only finitenonzero impulse response.

    Infinite-duration impulse response (IIR) systems have infinite nonzeroimpulse response.

    Causality: h[n] = 0 for n < 0

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 21 / 61

  • BIBO stability

    An LTI system is stable if and only if

    k=

    |h[k]| = S

  • Examples 1/2

    Example 2.3 Ideal delay system: h[n] = [n nd] (FIR, stable)

    Example 2.4 Moving average

    h[n] =1

    M1 +M2 + 1([n +M1] + [n +M1 2] +

    + [n] + [n 1] + + [n M2])

    Example 2.5 y[n] = (x[n])2, not LTI systems.

    Example 2.6 Accumulator

    h[n] = [n] + [n 1] + = u[n]

    (IIR, not stable)

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 23 / 61

  • Examples 1/2

    Example 2.7 w[n] = log10(|x[n]|) not LTI systems.

    Example 2.9 Compressor: y[n] = x[Mn] not LTI systems.

    Example 2.10 Forward difference: y[n] = [n + 1] [n].

    Example 2.10 Backward difference: y[n] = [n] [n 1]. Backwarddifference is the inverse system of accumulator.

    u[n] ([n] [n 1]) = u[n] u[n 1] = [n]

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 24 / 61

  • 2.5 Linear Constant-Coefficient DifferenceEquations

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 25 / 61

  • LCC Difference Equations

    An N th-order linear constant-coefficient (LCC) difference equations

    Nk=0

    aky[n k] =

    Mm=0

    bmx[nm]

    Example 2.14 Difference equation representation of the accumulator.

    y[n] =

    nk=

    x[k]

    Example 2.15 Difference equation representation of a moving-averagesystem.

    y[n] =1

    M2 + 1

    M2k=0

    x[n k]

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 26 / 61

  • Solve LCC DE 1/2

    For the difference equation

    Nk=0

    aky[n k] =

    Mm=0

    bmx[nm],

    the solution can be written by y[n] = yh[n] + yp[n].yh[n] is the solution to x[n] = 0, called the homogeneous solution. Theassociated homogeneous equation is

    Nk=0

    aky[n k] = 0.

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 27 / 61

  • Solve LCC DE 2/2

    Let y[n] = zn, we haveNk=0

    akzk = 0

    Denote its solution as z1, z2, . . . , zk. Then

    yh[n] =

    Nk=1

    akznk .

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 28 / 61

  • Example 2.16

    Solve y[n] = ay[n 1] + x[n], x[n] = K[n], y[1] = c

    Solution. For n 0,

    y[0] = ay[1] + x[0] = ac+K

    y[1] = ay[0] + x[1] = a(ac+K) = a2c+ aK

    For n < 0, y[n 1] = a1(y[n] x[n]), we have

    y[2] = a1(y[1] x[1]) = a1c

    y[3] = a1(y[2] x[2]) = a2c

    Therefore, y[n] = an+1c+Kanu[n].

    This system is NOT causal.

    This system is NOT linear.

    This system is NOT time-invariant.

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 29 / 61

  • Initial-Rest Conditions

    We can convert an LTI system to an LCC difference equation.

    We may not obtain an LTI system from an LCC difference equation.

    We may obtain several LTI systems from a LCC difference equation.

    Linearity, time invariance, and causality of system will depend on theauxiliary conditions.

    Initial-rest conditions. If x[n] = 0 for n < n0, then y[n] = 0 forn < n0. The system is LTI and causal for initial rest conditions.

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 30 / 61

  • 2.6 Frequency-Domain Representation

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 31 / 61

  • Eigenfunctions

    Suppose that an LTI system with impulse response h[n] have its inputx[n] = ejn, < n

  • Example 2.17

    The frequency response of the ideal delay system y[n] = x[n nd] isH(ej) = ejd . We have HR(e

    j) = cos(nd), HI(ej) = sin(nd).

    The magnitude response |H(ej)| = 1 and phase responseH(ej) = nd.

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 33 / 61

  • Linearity

    If x[n] =k

    kejkn, then y[n] =

    k

    kH(ejk)ejkn.

    Example 2.18

    x[n] = A cos(0n+ ) =A

    2ejej0n +

    A

    2ejej0n

    y[n] =A

    2ejH(ej0)ej0n +

    A

    2ejH(ej0)ej0n

    If h[n] is real, it can be shown that H(ej0) = H |ast(ej0), and

    y[n] = A|H(ej0)| cos(0n+ + H(ej0)

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 34 / 61

  • Example 2.20

    The frequency response of the moving-average system is

    H(ej) =1

    M1 +M2 + 1

    M2n=M1

    ejn =ej(M2M1)/2

    M1 +M2 + 1

    sin[(M1 +M2 + 1)/2]

    sin(/2)

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 35 / 61

  • Causal Input 1/2

    An input x[n] = ejnu[n] is applied to an LTI system and generates theoutput y[n] = yss[n] + yt[n] where yss[n] = H(e

    j)ejn is thesteady-state response and yt[n] is the transient response.

    y[n] =

    k=

    h[k]x[n k]

    =

    nk=

    h[k]ej(nk)u[n k]

    =

    k=

    h[k]ej(nk)

    yss[n]

    k=n+1

    h[k]ej(nk)

    yt[n]

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 36 / 61

  • Causal Input 2/2

    yt[n] =

    k=n+1

    h[k]ejkejn

    If h[n] = 0 except 0 n M , then yt[n] = 0 for n > M 1.

    If h[n] has infinite duration, then

    |yt[n]|

    k=n+1

    |h[k]|

    k=0

    |h[k]|

    If the system is stable, |yt[n]| is bounded.

    In fact, if the system is stable, then

    |yt[n]|

    k=n+1

    |h[k]| 0

    by Cauchy condition.

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 37 / 61

  • 2.7 Representation of Sequences by FourierTransform

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 38 / 61

  • Fourier Transform of Sequences

    X(ej) =

    n=

    x[n]ejn

    x[n] =1

    2

    X(ej)ejnd

    X(ej) = XR(ej) + jXI(e

    j)

    X(ej) = |X(ej)|ejX(ej)

    X(ej) has a period 2.ARG[X(ej)] = X(ej), X(ej) < . ARG[X(ej)] maybe not continuous.If we want a continuous phase functions for 0 < < , we use thenotation arg[X(ej)].

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 39 / 61

  • Convergence 1/2

    If x[n] is absolute summable, i.e.,

    n=

    |x[n]|

  • Convergence 2/2

    If x[n] is square summable, i.e.,

    n=

    |x[n]|2

  • Example 2.22

    Hlp(ej) =

    {1, || < c,

    0, otherwise, hlp[n] =

    sincn

    n, < n

  • Example 2.22

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 43 / 61

  • Example 2.23 & 2.24

    Example 2.23 Impulse train

    x[n] = 1, X(ej) =

    r=

    2( + 2r)

    = 2(), < .

    Note that X(ej) is periodic with period of 2.

    Example 2.24 Complex exponential sequences

    x[n] = ej0n, X(ej) =

    r=

    2( 0 + 2r)

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 44 / 61

  • Example: Fourier Transform of Unit Step Function

    x[n] = u[n], X(ej) =1

    1 ej+

    r=

    ( + 2r)

    Proof. Let x[n] = 12 + s[n] where s[n] =12 for n 0 and s[n] =

    12 for

    n < 0. Then the Fourier transform of x[n] is the sum of the Fouriertransform of a constant 12 and that of s[n]. Consider

    t[n] =

    {an, n 0

    an, n < 0., |a| < 1.

    We have s[n] = 12 lima1t[n] and S(ej) =

    1

    2lima1

    T (ej).

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 45 / 61

  • Example: Fourier Transform of Unit Step Function

    It is easy to evaluate T (ej)

    T (ej) =

    n=0

    anejn

    1n=

    anejn =1 + a2 2aej

    (1 aej)(1 aej)

    and

    lima1

    T (ej) =2

    1 ej

    Therefore,

    X(ej) =1

    1 ej+ (), < .

    Remark. The () in X(ej) represents the effect of DC component inx[n].

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 46 / 61

  • 2.8 Symmetry Properties of the Fourier Transform

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 47 / 61

  • Even-Odd Decomposition 1/2

    For a complex sequence x[n]

    conjugate-symmetric: x[n] = x[n].conjugate-antisymmetric: x[n] = x[n].

    For a real sequence x[n]

    even: x[n] = x[n].odd: x[n] = x[n].

    Any complex (real) sequence can be represented as a sum of aconjugate-symmetric (even) sequence and a conjugate-antisymmetric(odd) sequence: x[n] = xe[n] + xo[n] where

    xe[n] =1

    2(x[n] + x[n])

    xo[n] =1

    2(x[n] x[n]).

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 48 / 61

  • Even-Odd Decomposition 2/2

    Any complex (real) Fourier transform X(ej) can be represented as a sumof a conjugate-symmetric (even) function Xe(e

    j) and aconjugate-antisymmetric (odd) function Xo(e

    j)

    X(ej) = Xe(ej) +Xo(e

    j)

    where

    Xe(ej) =

    1

    2

    [X(ej) +X(ej)

    ]Xo(e

    j) =1

    2

    [X(ej)X(ej)

    ].

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 49 / 61

  • Properties

    If x[n]F X(ej), then

    1 x[n]F X(ej)

    2 x[n]F X(ej)

    3 {x[n]}F Xe(e

    j)

    4 j{x[n]}F Xo(e

    j)

    5 xe[n]F {X(ej)}

    6 xo[n]F j{X(ej)}

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 50 / 61

  • Properties

    If x[n] is real and x[n]F X(ej), then

    1 Fourier Transform is conjugate symmetric. X(ej) = X(ej)

    2 Real part is even. {X(ej)} = {X(ej)}

    3 Imaginary part is odd. {X(ej)} = {X(ej)}

    4 Magnitude is even. |X(ej)| = |X(ej)|

    5 Phase is odd. X(ej) = X(ej)

    6 xe[n]F {X(ej)}

    7 xo[n]F j{X(ej)}

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 51 / 61

  • 2.9 Fourier Transform Theorems

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 52 / 61

  • Theorems 1/2

    Linearity

    ax1[n] + bx2[n]F aX1(e

    j) + bX2(ej)

    Time shiftx[n nd]

    F ejndX(ej)

    Frequency shift

    ej0nx[n]F X(ej(0))

    Time-reversalx[n]

    F X(ej)

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 53 / 61

  • Theorems 2/2

    Differentiation in Frequency

    nx[n]F j

    dX(ej)

    d

    Parsevals Theorem

    n=

    |x[n]|2 =1

    2

    |X(ej)|2d

    Convolution Theoremx[n] h[n]

    F X(ej)H(ej)

    Modulation Theorem

    x[n]w[n]F

    1

    2

    X(ej)W (ej())d

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 54 / 61

  • Examples 2.26 2.29

    Example 2.26 Find F{anu[n 5]}.

    Example 2.27 Find inverse Fourier transform of

    X(ej) =1

    (1 aej)(1 aej)

    Example 2.28 Find inverse Fourier transform of

    H(ej) =

    {ejnd , c < || < ,

    0, || < c.

    Example 2.29 Find impulse response of the following stable, LTI system

    y[n]1

    2y[n 1] = x[n]

    1

    4x[n 1].

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 55 / 61

  • 2.10 Discrete-Time Random Random Signals

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 56 / 61

  • I-O Relationships 1/2

    Let the input sequence x[n] be a random precess with mean functionmx[n] , E{x[n]} and autocorrelation function

    xx[n, n+ k] , E{x[n]x[n + k]}.

    A wide-sense stationary (WSS) random process has meanindependent of n

    mx[n] = mx

    and has autocorrelation function depending on time difference only

    xx[n, n+ k] = xx[k].

    Suppose that a stable LTI system is characterized by

    y[n] =

    k=

    h[n k]x[k] =

    k=

    h[k]x[n k].

    When the input sequence x[n] is a WSS random precess, is theoutput y[n] also a WSS random process?

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 57 / 61

  • I-O Relationships 2/2

    The mean of y[n] is

    my[n] = mx

    k=

    h[k] = mxH(ej0) = my

    The autocorrelation function yy[m] is

    yy[n, n+m] = E{y[n]y[n +m]} =

    =

    xx[m ]chh[]

    where

    chh[] =

    k=

    h[k]h[ + k]

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 58 / 61

  • Power Density Spectrum 1/2

    Assume that mx = 0. Let xx(ej) = F{xx[m]}, yy(e

    j) =F{yy [m]}, and Chh(e

    j) = F{chh[]}. We have

    yy(ej) = Chh(e

    j)xx(ej)

    Chh(ej) = H(ej)H(ej) = |H(ej)|2

    yy(ej) = xx(e

    j)|H(ej)|2

    xx(ej) is called the power density spectrum of x[n] because

    total power = E{x2[n]} = xx[0]

    =1

    2

    xx(ej)d

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 59 / 61

  • Power Density Spectrum 2/2

    Since x[n] is real, we have

    xx(ej) is real and even because xx[m] is even and real.

    xx(ej) 0.

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 60 / 61

  • Example 2.30

    A white noise is a signal for which xx[m] = 2x[m].

    xx(ej) = 2x.

    yy(ej) = 2x|H(e

    j)|2

    Der-Feng Tseng (NTUST) DSP Chapter 2 February 19, 2015 61 / 61

    2.1 Discrete-Time Signals: Sequences2.2 Discrete-Time Systems2.3 Linear Time-Invariant Systems2.4 Properties of Linear Time-Invariant Systems2.5 Linear Constant-Coefficient Difference Equations2.6 Frequency-Domain Representation2.7 Representation of Sequences by Fourier Transform2.8 Symmetry Properties of the Fourier Transform2.9 Fourier Transform Theorems2.10 Discrete-Time Random Random Signals