Table 5.2.

CTFT pairs for elementary CT signals

Time domain CT signals (1) Constant (2) Impulse function (3) Unit step function (4) Causal decaying exponential function (5) Two-sided decaying exponential function (6) First-order time-rising causal decaying exponential function (7) N th-order time-rising causal decaying exponential function (8) Sign function (9) Complex exponential (10) Periodic cosine function (11) Periodic sine function (12) Causal cosine function (13) Causal sine function (14) Causal decaying exponential cosine function x(t) = 1 (t) u(t) eat u(t) ea|t| teat u(t) 1 2

Frequency domain

X ()e jt dt

X () =

x(t)e jt dt

Comments

2 () 1 () + 1 a + j 1 j a>0 a>0 a>0

2a a 2 + 2

1 (a + j)2 n! (a + j)n+1

t n eat u(t)

a>0

sgn(t) = ej0 t cos(0 t) sin(0 t)

1 t >0 1 t < 0

2 j 2( 0 )

cos(0 t)u(t) sin(0 t)u(t) eat cos(0 t)u(t)

[( 0 ) + ( + 0 )] [( 0 ) ( + 0 )] j j [( 0 ) + ( + 0 )] + 2 2 0 2 0 [( 0 ) ( + 0 )] + 2 2j 0 2 a + j 2 (a + j)2 + 0 0 2 (a + j)2 + 0

a>0

(15) Causal decaying eat sin(0 t)u(t) exponential sine function t 1 |t| /2 (16) Rectangular function rect = 0 |t| > /2 (17) Sinc function W sinc Wt

a>0 =0

sinc rect

2 2W
2

(18) Triangular function (19) Impulse train

|t| t |t| 1 = 0 otherwise (t kT0 )

1 || W 0 || > W >0 angular frequency 0 = 2 /T0

sinc 0

2 ( m0 )

k=

m=

(20) Gaussian function

et

/2 2

2 2 2 e /2

Table 5.4. Symmetry and transformation properties of the CTFT

Time domain Transformation properties Linearity Scaling Time shifting Frequency shifting Time differentiation Time integration

Frequency domain

1 x(t) = 2

X ()e

jt

X () =

x(t)e jt dt

Comments a1 , a2 C a , real-valued t0 , real-valued 0 , real-valued

t

a1 x1 (t) + a2 x2 (t) x(at) e j0 t x(t) dn x dt n

t

x(t t0 )

a1 X 1 () + a2 X 2 () 1 X |a| a ejt0 X () X ( 0 ) ( j)n X () X () + X (0)() j dn X ( j) dn 2 x()

n

provided dx/dt exists provided exists

x( )d

x( )d

Frequency differentiation Duality Time convolution Frequency convolution

t x(t) X (t) x1 (t) x2 (t) x1 (t) x2 (t)

provided dX/d exists X () if x(t) convolution in time domain

CTFT

X 1 ()X 2 () 1 [X 1 () X 2 ()] 2 |x(t)|2 dt =

multiplication in time domain energy in a signal

Parsevals relationship

Ex =

1 2

|X ()|2 d

Symmetry properties CTFT: X () = X ()

Hermitian property

x(t) is a real-valued function

real and imaginary components Re{X ()} = Re{X ()} Im{X ()} = Im{X ()} magnitude and phase spectra |X ()| = |X ()| <X () = <X ()

real component is even; imaginary component is odd magnitude spectrum is even; phase spectrum is odd simplied CTFT expression for even signals simplied CTFT expression for odd signals CTFT is real-valued and even

Even function

x(t) is even

X () = 2
0

x(t) cos(t)dt

Odd function

x(t) is odd

X () = j2
0

x(t) sin(t)dt

Real-valued and even function

x(t) is even and real-valued

Re{X ()} = Re{X ()} Im{X ()} = 0

CT signals x(t)

CTFT X ( j) =

Laplace transform

x(t)e

jt

dt

X (s) =

x(t)est dt

(1) Impulse function x(t) = (t) (2) Unit step function x(t) = u(t) (3) Causal gate function x(t) = u(t) u(t a) (4) Causal decaying exponential function x(t) = eat u(t) (5) Causal ramp function x(t) = tu(t) (6) Higher-order causal ramp function x(t) = t n u(t) (7) First-order time-rising causal decaying exponential function x(t) = teat u(t) (8) Higher-order time-rising causal decaying exponential function x(t) = t n eat u(t) (9) Causal cosine wave x(t) = cos(0 t)u(t) (10) Causal sine wave x(t) = sin(0 t)u(t)

1 1 j 1 j

1 ROC: entire s-plane 1 s ROC: Re{s} > 0 1 (1 eas ) s ROC: Re{s} > 0 1 a+ s ROC: Re{s} > a 1 s2 ROC: Re{s} > 0 n! s n+1 ROC: Re{s} > 0 1 (a + s)2 ROC: Re{s} > a n! (a + s)n+1 ROC: Re{s} > a s 2 0 + s 2 ROC: Re{s} > 0 0 2 0 + s 2 ROC: Re{s} > 0
2 20 + s 2 2 s 40 + s 2

() +

(1 eja ) () +

1 a + j

does not exist

does not exist

1 (a + j)2 provided a > 0. n! (a + j)n+1 provided a > 0 [( 0 ) + ( + 0 )] j + 2 0 2 [( 0 ) ( + 0 )] 2j 0 + 2 0 2 [() + ( 20 ) + ( + 20 )] 2 + 1 j + 2 j2 2 40 2

(11) Squared causal cosine wave x(t) = cos2 (0 t)u(t)

ROC: Re{s} > 0

2 20 2 s 40 + s 2

(12) Squared causal sine wave x(t) = sin2 (0 t)u(t)

[() ( 20 ) ( + 20 )] 2 + 1 j 2 j2 2 40 2 a + j 2 (a + j)2 + 0

ROC: Re{s} > 0 a+ s 2 (a + s)2 + 0

(13) Causal decaying exponential cosine function

Table 11.2. DTFTs and DTFSs for elementary DT sequences Note that the DTFS does not exist for aperiodic sequences
Sequence: x[k] DTFS: Dn = Dn = 1 does not exist does not exist 1 Dn = for all n K0 does not exist does not exist does not exist 1 K0 x[k]ejn0 k
k= K 0

DTFT: X () = X () = 2

k=

x[k]ejk

(1) x[k] = 1 (2) x[k] = [k] (3) x[k] = [k k0 ] (4) x[k] =

m=

m=

( 2m)

(k m K 0 )

X ( ) = 1 X () = ejk0 2m 2 X () = K 0 m= K0 X ( ) = X () = X () =
m=

(5) x[k] = u[k] (6) x[k] = p u[k] with | p| < 1 (7) First-order time-rising decaying exponential x[k] = (k + 1) p k u[k], with | p| < 1. (8) Complex exponential (periodic) x[k] = e jk0 K 0 = 2 p/0 (9) Complex exponential (aperiodic) x[k] = e jk0 , 2/0 = rational (10) Cosine (periodic) x[k] = cos(0 k) K 0 = 2 p/0 (11) Cosine (aperiodic) x[k] = cos(0 k), 2/0 = rational (12) Sine (periodic) x[k] = sin(0 k) K 0 = 2 p/0 (13) Sine (aperiodic) x[k] = sin(0 k),
k

( 2m ) +

1 (1 pej )2

1 1 pej

1 1 ej

Dn =

n = p r K0

0 elsewhere for < r <

X () = 2

m=

( 0 2m )

does not exist

X () = 2

m=

( 0 2m )

1 n = p r K0 Dn = 2 0 elsewhere for < r < does not exist

X () =

+ X () =

m=

( + 0 2m ) ( 0 2m )

m=

+ 1 j n = p r K0 Dn = 2 0 elsewhere for < r < does not exist X () = j

m=

( + 0 2m ) ( 0 2m )

m=

j X () = j

m=

( + 0 2m ) ( 0 2m )

m=

m=

( + 0 2m )

Table 11.2. (cont.)

Sequence: x[k] DTFS: Dn = 1 K0 x[k]ejn0 k
k= K 0

DTFT: X () =

k=

x[k]ejk

(14) Rectangular (periodic) 1 |k| N x[k] = 0 N < |k| K 0 /2 x[k] = x[k + K 0 ] (15) Rectangular (aperiodic) 1 |k| N x[k] = 0 elsewhere (16) Sinc Wk W sinc = x[k] = sin(W k) for 0 < W < k (17) Arbitrary periodic sequence with period K 0 x[k] = Dn e jn0 k
n= K 0

does not exist

2n + k = r K 0 X () = 2 (2N 1)/K 0 Dn K0 2N +1 n= Dn = 1 sin K 0 n elsewhere K 0 sin 1 n K0

sin

X () = does not exist

2N + 1 2 1 sin 2

1 || W 0 W < || X () = X ( + 2 ) X () = x[k]ejn0 k X () = 2
n=

Dn =

1 K0

k= K 0

Dn

2n K0

Appendix D, and has been used later in this chapter in solving Examples 11.15 and 11.18.

11.3 Existence of the DTFT

Definition 11.4 The DTFT X () of a DT sequence x[k] is said to exist if |X ()| < , for < < . (11.29)

The above denition for the existence of the DTFT satises our intuition that a valid function should be nite for all values of the independent variable. Substituting the value of the DTFT X () from Eq. (11.28b), Eq. (11.29) can be expressed as follows:
k=

x[k]ejk < ,

which is satised if

|x[k]| |ejk | < .

Table 11.5. Properties of the DTFS: sequences x [k], x 1 [k], and x 2 [k] are periodic with a period of K 0
Properties Time domain x[k] = x1 [k] x1 [k] Periodicity Linearity Scaling Time shifting Frequency shifting Time differencing Time summation x[k] a1 x1 [k] + a2 x2 [k] x (m) [k] with period m K 0 x[k k0 ] 2 n 0 k x[k] exp j K0 x[k] x[k 1]
k

Frequency domain Dn e jn0 k Dn =

x Dn 1 x Dn 2

Comments
0 = 2/K 0 0 = 2/K 0 0 = 2/K 0

n= K 0

1 K0

x[k]ejn0 k
k= K 0

x x a 1 Dn 1 + a 2 Dn 2 1 Dn m

Dn = Dn+K 0

a1 , a2 C m = 1, 2, 3, . . . k0 R n0 R

exp j Dnn 0

2k0 n Dn K0 2 n K0

1 exp j 1 1 exp j
x x K 0 Dn 1 Dn 2

Dn Dn summation S is nite only if D0 = 0 convolution over a period K 0 multiplication in time domain power of a periodic sequence

S=

x[m]
m=

2 n K0

Periodic convolution
n= K 0

x1 [n]x2 [n k]

Frequency convolution Parsevals relationship

x1 [k]x2 [k]
m= K 0

x2 x Dm1 Dmn

1 K0

k= K 0

|x[k]| =

n= K 0

|Dn |2

Symmetry properties
DTFS: Dn = Dn

Comments real component is even; imaginary component is odd

real and imaginary components: Re{Dn } = Re{Dn } Hermitian property x[k] is a real-valued sequence Im{Dn } = Im{Dn }

magnitude and phase spectra: |Dn | = |Dn | <Dn = <Dn

magnitude spectrum is even; phase spectrum is odd DTFS is real-valued and even DTFS is imaginary and odd

Real-valued and even function Real-valued and odd function

x[k] is an even and real-valued sequence x[k] is an odd and real-valued sequence

Re{Dn } = Re{Dn } Im{Dn } = 0 Re{Dn } = 0 Im{Dn } = Im{Dn }

Table 11.6. Properties of the discrete-time Fourier transform (DTFT)

Transformation properties Time domain x[k] = x1 [k] x2 [k] Periodicity Linearity Scaling Time shifting Frequency shifting Time differencing Time summation x[k] a1 x1 [k] + a2 x2 [k] x (m) [k] x[k k0 ] 1 2
2

Frequency domain X ()e jk d X ( ) = X1( ) X 2 () a1 X 1 () + a2 X 2 () X (m ) exp( jk0 )X () [1 exp( j)]X () 1 X ( ) + 1 exp( j) X (0)
m= k=

Comments

x[k]ejk

X () = X ( + 2 )

a1 , a2 C k0 R

m = 1, 2, 3, . . .
0 R

exp( jk 0 )x[k] x[k] x[k 1]

k

X ( 0 )

S=

x[m]
m=

provided summation S is nite

( 2 m) over period K 0 multiplication in time domain energy in a signal

Time convolution Periodic convolution Frequency convolution Parsevals relationship

x1 [k] x2 [k] x1 [k]x2 [k]

2

x1 [k] x2 [k]

X 1 ()X 2 () X 1 ()X 2 () 1 X 1 ()X 2 ( )d 2

2

1 Ex = |x[k]| = 2 k=

|X ()|2 d
2

Symmetry properties DTFT: X () = X () real and imaginary component: Im{X ()} = Im{X ()} magnitude and phase spectra: |X ()| = |X ()| <X () = <X () Real-valued and even function Real-valued and odd function x[k] is even and real-valued x[k] is odd and real-valued Re{X ()} = Re{X ()} Im{X ()} = 0 Re{X ()} = 0 Im{X ()} = Im{X ()} Re{X ()} = Re{X ()} real component is even: imaginary component is odd magnitude spectrum is even; phase spectrum is odd DTFT is real-valued and even DTFT is imaginary and odd

Hermitian property

x[k] is a real-valued function

Table 13.1. Unilateral z-transform pairs for several causal DT sequences

DT sequence 1 X (z)z k1 dz x[k] = 2 j
C

z-transform with ROC X (z) =

k=

x[k]z k

(1) Unit impulse x[k] = [k] (2) Delayed unit impulse x[k] = [k k0 ] (3) Unit step x[k] = u[k] (4) Exponential x[k] = k u[k] (5) Delayed exponential x[k] = k1 u[k 1] (6) Ramp x[k] = ku[k] (7) Time-rising exponential x[k] = k k u[k] (8) Causal cosine x[k] = cos(0 k)u[k] (9) Causal sine x[k] = sin(0 k)u[k] (10) Exponentially modulated cosine x[k] = k cos(0 k)u[k] (11) Exponentially modulated sine I x[k] = k sin(0 k)u[k] (12) Exponentially modulated sine II x[k] = r k sin(0 k + )u[k], with R.
(a)

1, ROC: entire z-plane z k0 ,ROC: entire z-plane, except z = 0 z 1 , = 1 1z z1 ROC: |z| > 1 ROC: |z| > || ROC: |z| > || ROC: |z| > 1

1 z , = 1 z 1 z z 1 1 , = 1 z 1 z

z 1 z = , (1 z 1 )2 (z 1)2

z 1 z = , ROC: |z| > || (1 z 1 )2 (z )2

z[z cos 0 ] 1 z 1 cos 0 , = 2 1 2z 1 cos 0 + z 2 z 2z cos 0 + 1 z sin 0 z 1 sin 0 = 2 , 1 2z 1 cos 0 + z 2 z 2z cos 0 + 1

ROC: |z| > 1 ROC: |z| > 1

1 z 1 cos 0 z[z cos 0 ] = 2 , ROC: |z| > || 1 2z 1 cos 0 + 2 z 2 z 2z cos 0 + 2 z 1 sin 0 z sin 0 = 2 , ROC: |z| > 1 cos + 2 z 2 1 2z z 2z cos 0 + 2 0 ROC : |z| ||(a)

z(Az + B) A + Bz 1 = 2 , 1 + 2 z 1 + 2 z 2 z + 2 z + 2 , and = tan1

Where r =

A2 2 + B 2 2AB , 0 = cos1 2 2

A 2 2 . B A

DTFT is obtained by computing the z-transform along the unit circle in the complex z-plane. Table 13.1 lists the z-transforms for several commonly used sequences. Comparing Table 13.1 with Table 11.2, we observe that when the sequence is causal and its DTFT exists, the DTFT can be obtained from the z-transform by substituting z = ej . Since the substitution z = ej can only be made if the ROC contains the unit circle, an alternative condition for the existence of the DTFT is the inclusion of the unit circle within the ROC of the z-transform. If the ROC of a z-transform does not include the unit circle, we cannot substitute z = ej

Table 13.2. Properties of the z-transform for transform pairs x[k]X(z), ROC: R x ; x[k]u[k]X (c) (z), ROC: z z R x ; x 1 [k]X 1 (z), ROC: R1 ; x 2 [k]X 2 (z), ROC: R 2
Properties Linearity Time scaling Time shifting (non-causal) Time shifting (causal) Time domain a1 x1 [k] + a2 x2 [k] x [k] for m = 1, 2, 3, . . . x[k m] x[k m]u[k m] x[k + m]u[k] x[k m]u[k] Frequency shifting Time differencing ej0 k x[k] x[k] x[k 1]
k (m)

z-domain a1 X 1 (z) + a2 X 2 (z) X (z ) z m X (z) z m X (c) (z)

m1 m

ROC at least R1 R2 (Rx )1/m

z m X (c) (z) z m

x[k]z k
k=0 m

z m X (c) (z) + z m X (ej0 z) (1 z )X (z) z X (z) z1 z dX (z) dz

1

x[k]z k
k=1

Rx , except for the possible deletion or addition of z = 0 or z= Rx Rx , except for the possible deletion of the origin Rx ( |z| > 1) Rx at least R1 R2 provided x[k] = 0 for k < 0 provided x[] exists

Time accumulation z-domain differentiation Time convolution Initial-value theorem Final-value theorem
(a)

y[k] = kx[k]

x[m] (a)
m=0

x1 [k] x2 [k]

X 1 (z)X 2 (z) x[0] = lim X (z)

z

x[] = lim x[k]= lim(z 1)X (z)

k z1

Provided that the sequence y[k] has a nite value for all k.

Proof Using Eq. (13.7), the z-transform of a1 x1 [k] + a2 x2 [k] is calculated as follows: Z {a1 x1 [k] + a2 x2 [k]} =
k=0

{a1 x1 [k] + a2 x2 [k]} z k

= a1

k=0

x1 [k]z k + a2
X 1 (z)

k=0

x2 [k]z k ,
X 2 (z)

which proves the algebraic expression, Eq. (13.16). To determine the ROC of the linear combination, we note that the z-transform X 1 (z) is nite within the specied ROC, R . Similarly, X (z) is nite within its ROC, R . Therefore, the