Signals & systems Ch.3 Fourier Transform of Signals and LTI System 4/13/2015 Signals and systems in the Frequency domain Fourier transform Time [sec] 4/13/2015 Frequency KyungHee University [sec-1, Hz] 2 3.1 Introduction • Orthogonal vector => orthonomal vector vo (1/ 2,1/ 2 ) v1 (1/ 2,1/ 2 ) vo v1 0 vo v0 1 v1 v1 1 vi v j [i j] for i, j 0,1 Any vector in the 2- dimensional space can be represented by weighted sum of 2 orthonomal vectors Fourier h (h0 , h1 ) H (H0 , H1 ) h H 0vo H1v1 • What is meaning of magnitude of H? Fourier Transform(FT) Inverse FT 4/13/2015 h H 0vo H1v1 H 0 h vo H1 h v1 KyungHee University 3 3.1 Introduction cont’ • CDMA? 1 1 1 1 v (1 , 1 , 1 , 1) v (1 , 1 , 1 , 1) v (1 , 1 , 1 , 1) vo ( , , , ) 1 2 3 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 Orthogonal? vo v1 0 vo v0 1 v1 v1 1 vi v j [i j] for i, j 0,1, 2,3 Any vector in the 4- dimensional space can be represented by weighted sum of 4 orthonomal vectors Orthonormal function? vk (t ) A cos( k0t ) where 0 vk (t ) e jk0t where 0 2 T 2 T T vi (t )v j (t )dt [i j ] vm (t )vk* (t ) 1 * v ( t ) v (t )dt [m k ] m k T T 1 1 j ( m k ) 0t * v ( t ) v ( t ) dt e dt m k T T T T 4/13/2015 KyungHee University 4 3.1 Introduction cont’ 1 1 j ( m k ) 0t 1 * v ( t ) v ( t ) dt e dt 1 dt 1 m k T T T T T T T j 2 (Tm k ) t 1 j ( m k ) 0t 1 e dt e 0 T T j (m k )2 0 i) If m k , ii) If m k , • Fourier Series (FS) Any periodic function(signal) with period T can be represented by weighted sum of orthonormal functions. f (t ) f (t T ) F v (t ) k k k where Fk T f (t )vk (t )dt What is meaning of magnitude of Fk? Think about equalizer in audio system. 4/13/2015 KyungHee University 5 3.2 Complex Sinusoids and Frequency Response of LTI Systems cf) impulse response h[k ]x[n k ] y[n] h[k ]e k y[n] e jn h[k ]e j ( n k ) . k jk j H ( e )e jn j H (e ) , y(t ) h( )e j ( t ) d e jt h( )e j h[k ]e jk . k k How about x[n] z n for complex z? (3.1) d H ( j)e How about x(t ) e st for complex s? jt , H ( j ) h( )e j d . (3.3) y(t ) H ( j) e j (t arg{H ( j )}) . Magnitude to kill or not? 4/13/2015 Phase delay KyungHee University 6 3.3 Fourier Representations for Four Classes of signals • 3.3.1 Periodic Signals: Fourier Series Representations “Any periodic function can be represented by weighted sum of basic periodic function.” Fourier said (periodic) - (discrete) (discrete) - (periodic) Time Property Continuous (t) Discrete [n] DTFS xˆ[n] A[k ]e jk0n , k CTFS xˆ (t ) A[k ]e jk0t , (3.5) Periodic? k Periodic Fourier Series (FS) Discrete-Time Fourier Series (DTFS) 2 N 2 0 T 0 Non-periodic Fourier Transform (CTFT) Discrete-Time Fourier Transform (DTFT) (3.4) e j ( N k )0n e jN 0n e jk0n e j 2 ne jk0n e jk0n . yes! With the period of N 4/13/2015 KyungHee University 7 Fourier transform Time domain N 1 Periodic discrete 2 2 0 0 T N Periodic continuous x[n] X [k ]e jk0 n frequency domain 1 N 1 X [k ] x[n]e jk0n N n 0 ,. k 0 x(t ) X [k ]e jk0t X [k ] k 1 j jn x[n] X ( e ) e d a-periodic discrete 2 2 1 X ( j )e jt d a-Periodic continuous x(t ) 2 z-transform x[n] 1 X ( z) z n 1 2 j 1 st X ( s ) e ds Laplace transform x(t ) 2 j 4/13/2015 KyungHee University dz j 1 T X (e ) T x(t )e jk0t dt x[n]e jn n X ( j ) x(t )e jt dt X ( z) x[n]z n n X (s) x(t )e st dt 8 3.3 Fourier Representations for Four Classes of signals cont’ • 3.3.2 Non-periodic Signals: Fourier-Transform Representations (aperiodic) - (continuous) (continuous) - (aperiodic) Inverse continuous time Fourier Transform (CTFT) (aperiodic) (discrete) X ( j )e jt d. - (continuous) - (periodic) Inverse discrete time Fourier Transform (DTFT) 4/13/2015 1 xˆ (t ) 2 KyungHee University 1 xˆ[n] 2 X (e j )e jn d. 9 3.4 Discrete-Time Periodic Signals: The Discrete-Time Fourier Series DTFS ;0 x[n] X [k ] (periodic) (discrete) - (discrete) - (periodic) N 1 Inverse DFT x[n] X [k ]e jk0n ,. DFT 1 N 1 X [k ] x[n]e jk0n N n 0 k 0 (3.10) Example 3.2 Determining DTFS Coefficients Figure 3.5 Time-domain signal for Example 3.2. Find the frequency-domain representation of the signal depicted in Fig. 3.5 4/13/2015 KyungHee University 10 3.4 Discrete-Time Periodic Signals: The Discrete-Time Fourier Series cont’ 1 2 1 X [k ] x[n]e jk 2n / 5 x[2]e jk 4 / 5 x[1]e jk 2 / 5 x[0]e j 0 x[1]e jk 2 / 5 x[2]e jk 4 / 5 5 n2 5 Just inner product to orthonormal vectors in the 5 dimensional space!! X[k] 1 * 1 x c k ( x[2], x[1], x[0], x[1], x[2]) (e jk 4 / 5 , e jk 2 / 5 , e j 0 , e jk 2 / 5 , e jk 4 / 5 ) 5 5 c2 (e j 8 / 5 , e j 4 / 5 , e j 0 , e j 4 / 5 , e j 8 / 5 ) c3 (e j12 / 5 , e j 6 / 5 , e j 0 , e j 6 / 5 , e j12 / 5 ) 1 * 1 j 8 / 5 j 4 / 5 j 0 j 4 / 5 j 8 / 5 c 2 c 3 (e ,e ,e ,e ,e ) (e j12 / 5 , e j 6 / 5 , e j 0 , e j 6 / 5 , e j12 / 5 ) 5 5 1 * ci c j [i j ] 5 4/13/2015 KyungHee University 11 3.4 Discrete-Time Periodic Signals: The Discrete-Time Fourier Series cont’ 1 1 1 1 X [k ] 1 e jk 2 / 5 e jk 2 / 5 1 j sin(k 2 / 5). 5 2 2 5 1 sin( 4 / 5) j 0.232 e j 0.531 5 5 1 X [ 0] 0.2e j 0 DC!! 5 X [2] X [1] sin( 2 / 5) 1 j 0.276 e j 0.760 5 5 X [1] X [2] 1 sin( 2 / 5) j 0.276 e j 0.760 5 5 1 sin(4 / 5) j 0.232e j 0.531. 5 5 Figure 3.6 Magnitude and phase of the DTFS coefficients for the signal in Fig. 3.5. Recall “weighted sum of orthonormal vectors in the 5 dimensional space” 4/13/2015 KyungHee University 12 3.4 Discrete-Time Periodic Signals: The Discrete-Time Fourier Series cont’ Example 3.3 Computation of DTFS Coefficients by Inspection v a 0 c0 a1c1 orthonormal c0 and c1 DFT (v0 , v1 ) (a0 , a1 ) Determine the DTFS coefficients of , using the method of inspection. x[n] e j n 3 e 2 j n 3 j n j n 1 1 e j e 3 e j e 3 2 2 Figure 3.8 Magnitude and phase of DTFS coefficients for Example 3.3. 4/13/2015 KyungHee University 13 3.4 Discrete-Time Periodic Signals: The Discrete-Time Fourier Series cont’ Example 3.4 DTFS Representation of an Impulse Train Find the DTFS Coefficients of the N-periodic impulse train x[n] [n lN ], l Figure 3.9 A discrete-time impulse train with period N. 1 N 1 X [k ] [n]e jkn2 / N N n 0 1 . N How about in other period? 1 X [k ] N 3N / 2 [n N ]e n N / 2 jkn2 / N 1 .` N Draw X[k] in the k-axis. (impulse train) - (impulse train) 4/13/2015 KyungHee University 14 3.4 Discrete-Time Periodic Signals: The Discrete-Time Fourier Series cont’ Example 3.5 The Inverse DTFS of periodic X[k] Use Eq. (3.10) to determine the time-domain signal x[n] from the DTFS coefficients depicted in Fig. 3.10 Figure 3.10 Magnitude and phase of DTFS coefficients for Example 3.5. x[n] 4 X [k ]e jk 2n / 9 k 4 e j 2 / 3e j 6n / 9 2e j / 3e j 4n / 9 1 2e j / 3e j 4n / 9 e j 2 / 3e j 6n / 9 2 cos(6n / 9 2 / 3) 4 cos(4n / 9 / 3) 1. 4/13/2015 KyungHee University 15 3.4 Discrete-Time Periodic Signals: The Discrete-Time Fourier Series cont’ Example 3.6 DTFS Representation of a Square Wave Find the DTFS coefficients for the N-periodic square wave -M n M . M n N M 1, x[n] 0, Figure 3.11 Square wave for Example 3.6. 1 X [k ] N N M 1 x[n]e jk0 n n M 1 N N M 1 e n M jk0 n 1 N 1 2 M jk0 ( mM ) 1 jk0M 2 M jk0m X [k ] e e e . N m 0 N m 0 1 X [k ] N 2M 1 1 , N m0 2M e jk0 M X [k ] N 4/13/2015 1 e jk0 ( 2 M 1) , jk 0 1 e k 0, N ,2 N , k 0, N ,2 N , , M e jk0 n . n M (3.15) 등비수열의 합 DC=mean=평균 1 e jk0 M X [k ] N 1 1 e jk0 ( 2 M 1) jk 0 1 e 1 e jk0 / 2 e jk0 ( 2 M 1) / 2 e jk0 ( 2 M 1) / 2 jk0 / 2 jk 0 / 2 jk 0 / 2 N e e e KyungHee University 16 3.4 Discrete-Time Periodic Signals: The Discrete-Time Fourier Series cont’ 1 sin(k 0 (2M 1) / 2) , N sin(k 0 / 2) k 0, N ,2 N ,. 1 sin(k (2M 1) / N ) , sin(k / N ) X [k ] N (2M 1) / N , k 0, N ,2 N , . k 0, N ,2 N , X [k ] DC, also 1 sin(k (2M 1) / N ) 2M 1 . k 0, N , 2 N , N sin( k / N ) N Then X [k ] 1 sin(k (2M 1) / N ) . for all k. N sin(k / N ) lim M=0 (impulse train) 2M+1 = N? X [k ] 1 sin(k ) ? N sin(k / N ) (square) - (sinc) 4/13/2015 KyungHee University 17 3.4 Discrete-Time Periodic Signals: The Discrete-Time Fourier Series cont’ Example 3.7 Building a Square Wave form DTFS Coefficients 1 sin(k (2M 1) / N ) . N sin(k / N ) X [k ] x[n] X [k ]e k jk0 n symmetric X [0] 2 X [k ] cos(k 0 n) k 1 Evaluate one period of the Jth term in Eq. (3.18) and the 2J+1 term approximation xˆ J [n] for J=1, 3, 5, 23, and 25. J xˆ J [n] B[k ] cos(k 0 n), (3.18) k 0 4/13/2015 KyungHee University 18 3.4 Discrete-Time Periodic Signals: The Discrete-Time Fourier Series cont’ Example 3.8 Numerical Analysis of the ECG Evaluate the DTFS representations of the two electrocardiogram (ECG) waveforms depicted in Figs. 3.15(a) and (b). (a) Normal heartbeat. (b) Ventricular tachycardia. (c) Magnitude spectrum for the normal heartbeat. (d) Magnitude spectrum for ventricular tachycardia. Figure 3.15 Electrocardiograms for two different heartbeats and the first 60 coefficients of their magnitude spectra. 4/13/2015 KyungHee University 19 3.5 Continuous-Time Periodic Signals: The Fourier Series Any periodic function can be represented by weighted sum of basic periodic functions. (periodic) (discrete) (continuous) (aperiodic) X [k ]e jk0t Inverse FT x(t ) FT 1 T X [k ] x(t )e jk0t dt T 0 4/13/2015 k ,. where (3.19) Recall “orthonormal”!! (3.20) KyungHee University 20 3.5 Continuous-Time Periodic Signals: The Fourier Series cont’ Example 3.9 Direct Calculation of FS Coefficients Determine the FS coefficients for the signal x(t ) depicted in Fig. 3. 16. Solution : 1 2 2t jkt 1 2 ( 2 jk )t 1 X [k ] e e dt e dt e ( 2 jk )t 2 0 2 0 2(2 jk ) 2 0 1 e4 4 jk 2 X[0]=? FT Figure 3.16 Time-domain signal for Example 3.9. 4/13/2015 Figure 3.17 Magnitude and phase for Ex. 3.9. KyungHee University 21 3.5 Continuous-Time Periodic Signals: The Fourier Series cont’ Example 3.10 FS Coefficients for an Impulse Train. Determine the FS coefficients x(t ) for the signal defined by (t 4l ) t Solution : X [k ] 1 1 2 1 6 jk ( / 2 ) t jk ( / 2 ) t x ( t ) e dt ( t ) e dt (t 4)e jk ( / 2)t dt T 2 2 4 4 4 (impulse train) (impulse train) Example 3.11 Calculation of FS Coefficients by Inspection Determine the FS representation of the signal x(t ) 3 cos(t / 2 / 4) using the method of inspection. Solution : x(t ) 3 e 4/13/2015 X [k ]e x(t ) j (t / 2 / 4 ) 3 j / 4 , 2 e 3 X [ k ] e j / 4 , 2 0, e 2 jkt / 2 (3.21) k j (t / 2 / 4 ) 3 3 e j / 4e jt / 2 e j / 4e jt / 2 2 2 k 1 k 1 (3.22) otherwise Figure 3.18 Magnitude and phase for Ex. 3.11. KyungHee University 22 3.5 Continuous-Time Periodic Signals: The Fourier Series cont’ Example 3.12 Inverse FS : Find the time-domain signal x(t) corresponding to the k FS coefficients . Assume that the fundamental periodTis 2. X [k ] (1/ 2) e jk / 20 Solution : 0 x(t ) (1/ 2) e k k 0 jk / 20 e jkt (1/ 2) k e jk / 20 e jkt k 1 k 0 l 1 등비수열의 합 (1/ 2) k e jk / 20 e jkt (1/ 2) l e jl / 20 e jlt x(t ) 1 1 1 j (t / 20 ) j (t / 20 ) 1 (1 / 2)e 1 (1 / 2)e x(t ) 3 5 4 cos(t / 20) Figure 3.20 FS coefficients for Problem 3.9. Figure 3.21 4/13/2015 Square wave for Example 3.13. KyungHee University 23 3.5 Continuous-Time Periodic Signals: The Fourier Series cont’ Example 3.13 FS for a Square Wave Determine the FS representation of the square wave depicted in Fig. 3.21. X [k ] 1 T /2 x(t )e jkwot dt T 2 T 1 To jkwot 1 jkwot To e dt e , To T To Tjkwo 2 sin(kwoTo ) 2 e jkwoTo e jkwoTo ( ) , Tkwo 2j Tkwo 2T 1 To For k=0, X [0] dt o T To T X [k ] 2 sin(kwoTo ) 2T0 2T sin c(k 0 ) Tkw0 T T k 0 k 0 lim k 0 2 sin(kwoTo ) 2To Tkw0 T where sin c(u ) sin(u ) u Figure 3.22 The FS coefficients, X[k], –50 k 50, for three square waves. In Fig. 3.21 Ts/T = (a) 1/4 . (b) 1/16. (c) 1/64. 4/13/2015 KyungHee University 24 3.5 Continuous-Time Periodic Signals: The Fourier Series cont’ To/T=1/2? (DC) (impulse) To 0? (impulse train) (impulse train) T ? (aperiodic) (continuous) (square) (sinc) (sinc) (square) Figure 3.23 4/13/2015 Sinc function sinc(u) = sin(u)/(u) KyungHee University 25 3.5 Continuous-Time Periodic Signals: The Fourier Series cont’ Example 3.14 Square-Wave Partial-Sum Approximation Let the partial-sum approximation to the FS in Eq.(3.29), be given by J ^ x j (t ) B[k ] cos(kwo t ) k 0 This approximation involves the exponential FS Coefficients with indices Consider a square wave T 1 and T0 / T 1 / 4. Depict one period of the J k J. J th term in ^ this sum, and find x j (t ) for J 1,3,7,9, and 99. Solution : 4/13/2015 k 0 1 / 2, B[k ] (2 /(k ))(1) ( k 1) / 2 , k odd 0, k even KyungHee University 26 3.6 Discrete-Time Non-periodic Signals : The Discrete-Time Fourier Transform (discrete) (periodic) 1 x[n] 2 j X (e ) x[n]e X (e jn )e jn d (3.31) (3.32) jn n DTFT x[n] X (e j ) x[n] n 4/13/2015 x[n] 2 n KyungHee University 27 3.6 Discrete-Time Non-periodic Signals : The Discrete-Time Fourier Transform cont’ Example 3.17 DTFT of an Exponential Sequence Find the DTFT of the sequence x[n] n u[n] Solution : j X (e ) n n u[n]e jn e n jn n 0 (e j ) n n 0 1 , j 1 e 1 1 1 1 X (e j ) ((1 cos) 2 2 sin 2 )1 / 2 ( 2 1 2 cos)1 / 2 1 cos j sin sin arg{ X (e j )} arctan( ) 1 cos = 0.5 = 0.9 X (e j ) x[n] = nu[n]. magnitude phase 4/13/2015 = 0.5 KyungHee University = 0.9 28 3.6 Discrete-Time Non-periodic Signals : The Discrete-Time Fourier Transform cont’ Example 3.18 DTFT of a Rectangular Pulse Let 1, x[n] 0, n M n M Find the DTFT of x[n]. Solution : (square) (sinc) Figure 3.30 Example 3.18. (a) Rectangular pulse in the time domain. (b) DTFT in the frequency domain. 4/13/2015 KyungHee University 29 3.6 Discrete-Time Non-periodic Signals : The Discrete-Time Fourier Transform cont’ 2M X (e ) e j j ( m M ) e jm m0 2M e m 9 jm 1 e j( 2 M 1) , e 1 e j 2M 1, j X (e ) e jM jm 0, 2 , 4 , 0, 2 , 4 , e j( 2 M 1) / 2 (e j( 2 M 1) / 2 e j( 2 M 1) / 2 ) e j / 2 (e j / 2 e j / 2) ) e j( 2 M 1) / 2 e j( 2 M 1) / 2 e j / 2 e j / 2 X (e j ) sin((2M 1) / 2) sin( / 2) sin((2M 1) / 2) 2M 1 0, 2 , 4 , , sin( / 2) lim X (e j ) 4/13/2015 sin((2M 1) / 2) sin( / 2) KyungHee University 30 3.6 Discrete-Time Non-periodic Signals : The Discrete-Time Fourier Transform cont’ Example 3.19 Inverse DTFT of a Rectangular Spectrum Find the inverse DTFT of Solution : x[n] (sinc) (square) 1 2 W e jn d W 1 W sin(Wn) n 0 n x[0] lim 1, W X (e j ) 0, W 1 jn W 1 e sin(Wn ), W n 2nj x[n] 1 sin(Wn) n x[n] n0 W si nc (Wn / ) Figure 3.31 (a) Rectangular pulse in the frequency domain. (b) Inverse DTFT in the time domain. 4/13/2015 KyungHee University 31 3.6 Discrete-Time Non-periodic Signals : The Discrete-Time Fourier Transform cont’ Example 3.20 DTFT of the Unit Impulse Find the DTFT of x[n] [n] Solution : (impulse) - (DC) j X (e ) [n]e jn 1 n DTFT [n] 1 Example 3.21 Find the inverse DTFT of a Unit Impulse Spectrum. Solution : (impulse train) (impulse train) 1 x[n] 2 ()e jn d j X (e ) ( k 2 ) k 1 DTFT (), 2 4/13/2015 KyungHee University 32 3.6 Discrete-Time Non-periodic Signals : The Discrete-Time Fourier Transform cont’ Example 3.22 Two different moving-average systems y1 [n] 1 ( x[n] x[n 1]) 2 y 2 [ n] 1 ( x[n] x[n 1]) 2 h1 [n] 1 1 [n] [n 1] 2 2 h2 [n] 1 1 [n] [n 1] 2 2 solution : Figure 3.36 H 1 ( e j ) e j 2 e j 2 e 2 4/13/2015 1 1 j e 2 2 j 2 e j 2 H 2 ( e j ) j e cos( ) e 2 2 j 2 H 1 (e j ) cos( ) 2 1 1 j e 2 2 e 2 j 2 je j 2 sin( ) 2 arg{ H 1 (e j )} KyungHee University 2 H 2 (e j ) sin( ) 2 /2 , 2 arg{H 2 (e j )} / 2, 2 0 0 33 3.6 Discrete-Time Non-periodic Signals : The Discrete-Time Fourier Transform cont’ Example 3.23 Multipath Channel : Frequency Response y[n] x[n] ax[n 1] Solution : H (e j ) 1 ae j 1 a e j (arg{a}) 1 a cos( arg{a}) j a sin( arg{a}) H (e j ) ((1 a cos( arg{ a})) 2 a sin 2 ( arg{ a})) 1/ 2 (1 a 2 a cos( arg{ a})) 1/ 2 2 2 H inv (e j ) H inv (e j ) 1 , 1 ae j 1 1 a e H inv (e j ) j ( arg{a}) a 1 1 1 a cos( arg{}) j a sin( arg{}) 1 (1 a 2 a cos( arg{a}))1 / 2 2 1 H (e j ) | H (e j ) | (a) a = 0.5ej2/3. (b) a = 0.9ej2/3. | H inv (e j ) | (a) a = 0.5ej2/3. (b) a = 0.9ej2/3. 4/13/2015 KyungHee University 34 3.7 Continuous-Time Non-periodic Signals : The Fourier Transform (continuous aperiodic) (continuous aperiodic) Inverse CTFT CTFT x(t ) 1 2 (3.35) X ( jw)e jwt dw X ( jw) x(t )e jwt dt (3.26) x(t ) X ( jw) CTFT Condition for existence of Fourier transform: 4/13/2015 x(t ) dt 2 KyungHee University 35 3.7 Continuous-Time Non-periodic Signals : The Fourier Transform cont’ Example 3.24 FT of a Real Decaying Exponential at Find the FT of x(t ) e u(t ) Solution : 0 eat dt , a 0 Therefore, FT not exists. a 0, X ( jw) e at u (t )e jwt dt e ( a jw) t dt 1 e ( a jw) t a jw 0 0 1 a jw x(t ) e at u(t ) X ( jw) 1 a 2 w2 arg{X ( jw)} arctan(w / a) LPF or HPF? Cut-off from 3dB point? 4/13/2015 KyungHee University 36 3.7 Continuous-Time Non-periodic Signals : The Fourier Transform cont’ Example 3.25 FT of a Rectangular Pulse 1, x(t ) 0, T0 t T0 Find the FT of x(t). t T0 Solution : (square) (sinc) X ( jw) x(t )e jwt For w 0, T0 dt e T0 T0 jwt 1 jwt 2 dt e sin(wT0 ), jw w T0 w0 2 sin( wT0 ) 2T0 w 0 w lim X ( jw) 2 sin( wT0 ) X ( jw) 2T0 s inc (wT0 / ) w X ( jw) 2 Example 3.25. (a) Rectangular pulse. (b) FT. 4/13/2015 sin(wT0 ) w 0, sin(wT0 ) / w 0 arg{X ( jw)} , sin(wT0 ) / w 0 KyungHee University 37 3.7 Continuous-Time Non-periodic Signals : The Fourier Transform cont’ Example 3.26 Inverse FT of an Ideal Low Pass Filter!! Fine the inverse FT of the rectangular spectrum depicted in Fig.3.42(a) and given by 1, X ( jw) 0, W w W w W Solution : (sinc) -- (square) 1 x(t ) 2 1 jwt e dw e W 2 jt W W jwt W 1 sin(Wt ), t t0 1 sin(Wt ) W / , t 0 t 1 W Wt x(t ) sin(Wt ) or x(t ) si nc( ) t when 4/13/2015 t 0, lim KyungHee University 38 3.7 Continuous-Time Non-periodic Signals : The Fourier Transform cont’ Example 3.27 FT of the Unit Impulse x(t ) (t ) Solution : (impulse) - (DC) X ( jw) (t )e jwt dt 1 FT (t ) 1 Example 3.28 Inverse FT of an Impulse Spectrum Find the inverse FT of X ( jw) 2 (w) Solution : (DC) (impulse) 1 x(t ) 2 2 ( w)e jwt dw 1 FT 1 2 (w) 4/13/2015 KyungHee University 39 3.7 Continuous-Time Non-periodic Signals : The Fourier Transform cont’ Example 3.29 Digital Communication Signals Ar , Rectangular (wideband) xr (t ) 0, t T0 / 2 t T0 / 2 Separation between KBS and SBS. Narrow band ( Ac / 2)(1 cos(2t /T 0)), xc (t ) 0, t T0 / 2 t T0 / 2 Figure 3.44 Pulse shapes used in BPSK communications. (a) Rectangular pulse. 4/13/2015 KyungHee University (b) Raised cosine pulse. 40 3.7 Continuous-Time Non-periodic Signals : The Fourier Transform cont’ Solution : Pr Pc 1 T0 Ac2 4T0 Figure 3.45 BPSK (a) rectangular pulse shapes (b) raised-cosine pulse shapes. T0 / 2 3 Ac2 8 T0 / 2 T0 / 2 T0 / 2 Ar2 dt Ar2 ( Ac2 / 4)(1 cos(2t / T0 ))2 dt T0 / 2 1 T0 T0 / 2 [1 2 cos(2t / T0 1 / 2 1 / 2 cos(4t / T0 )]dt 8 Ar Ac 3 the same power constraints sin(wT0 / 2 ) sin(fT0 ) 1 8 T0 / 2 X r' ( jf ) X c ( jw) (1 cos(2t / T0 ))e jwt dt w f 2 3 T0 / 2 2 T0 / 2 jwt 1 2 T0 / 2 j ( w2 / T0 )t 1 2 T0 / 2 j ( w 2 / T0 )t X c ( jw) e dt e dt e dt 3 T0 / 2 2 3 T0 / 2 2 3 T0 / 2 T /2 sin(T0 / 2) jt e dt 2 T / 2 X r ( jw) 2 0 0 X c ( jw) 2 X c' ( jf ) 4/13/2015 2 sin(wT0 / 2) 2 sin((w 2T0 )T0 / 2) 2 sin((w 2 / T0 )T0 / 2) 3 w 3 w 2 / T0 3 w 2 / T0 2 sin(fT0 ) 2 sin( ( f 1 / T0 )T0 ) 2 sin( ( f 1 / T0 )T0 ) 0.5 0.5 3 t 3 ( f 1 / T0 ) 3 ( f 1 / T0 ) KyungHee University 41 3.7 Continuous-Time Non-periodic Signals : The Fourier Transform cont’ rectangular pulse. One sinc Raised cosine pulse 3 sinc’s The narrower main lobe, the narrower bandwidth. But, the more error rate as shown in the time domain Figure 3.47 sum of three frequency-shifted sinc functions. 4/13/2015 KyungHee University 42 Fourier transform Time domain Periodic discrete 2 2 0 0 T N Periodic continuous x(t ) X [k ]e jk0t frequency domain X [k ] k 1 x[n] X (e j )e jn d a-periodic discrete 2 2 1 X ( j )e jt d a-Periodic continuous x(t ) 2 4/13/2015 KyungHee University j 1 T X (e ) T x(t )e jk0t dt x[n]e jn n X ( j ) x(t )e jt dt 43 9.1 Linearity and Symmetry Properties of FT z (t ) ax(t ) by(t ) FT Z ( jw) aX ( jw) bY ( jw) ; w0 z (t ) ax(t ) by(t ) FS Z [k ] aX[k ] bY[k ] z[n] ax[n] by[n] DTFT Z (e j ) aX (e j ) bY (e j ) DTFS ; 0 z[n] ax[n] by[n] Z [k ] aX[k ] bY[k ] z (t ) 2 1 x(t ) y (t ) 3 2 x(t ) y(t ) FS ; 2 x(t ) X [k ] (1 /(k )) sin(k / 4) FS ; 2 y(t ) Y [k ] (1 /(k )) sin(k / 2) 3 1 FS ; 2 z (t ) Z [k ] sin(k / 4) sin(k / 2) 2k 2k 4/13/2015 KyungHee University 44 9.1 Linearity and Symmetry Properties of FT cont’ • 3.9.1 Symmetry Properties : Real and Imaginary Signals X ( jw) x(t )e jwt dt (3.37) x (t )e jwt dt (real x(t)=x*(t)) (conjugate symmetric) X ( jw) x(t )e j ( w)t dt 4/13/2015 X ( jw) X ( jw) KyungHee University (3.38) 45 9.1 Linearity and Symmetry Properties of FT cont’ • 3.9.2 SYSMMEYRY PROPERTIES : EVEN/ODD SIGNALS (even) (real) (odd) (pure imaginary) xe (t ) Re{X ( j )} xo (t ) j Im{X ( j )} For even x(t), 4/13/2015 X * jw x t e jwt dt x e jw d X jw. KyungHee University real 46 3.10 Convolution Property • 3.10.1 CONVOLUTION OF NON-PERIODIC SIGNALS (convolution) (multiplication) yt ht * xt h xt d But xt 1 2 1 yt h 2 X jwe jwt dw. given xt 1 2 X jwe jwt dw. X jwe jwt e jw dw d change the order of integration 1 2 h e jw d X jwe jwt dw. jwt yt 2 H jwX jwe dw, 1 FT yt ht * xt Y jw X jwH jw. 4/13/2015 3.40 KyungHee University 47 3.10 Convolution Property cont’ Example 3.31 Convolution problem in the frequency domain xt 1 t sint be the input to a system with impulse response ht 1 t sin2t . Find the output yt xt * ht . Solution: 1, w FT xt X jw 0, w 1, w 2 FT ht H jw . 0, w 2 FT yt xt * ht Y jw X jwH jw, 1, w Y jw , 0 , w 4/13/2015 yt 1 t sint . KyungHee University 48 3.10 Convolution Property cont’ Example 3.32 Find inverse FT’S by the convolution property Use the convolution property to find xt , where 4 xt X jw 2 sin 2 w. w FT 2 Z jw sin w. w 1, t 1 FT z t Z jw, 0, t 1 Ex 3.32 (p. 261). (a) Rectangular z(t). (b) z (t ) * z (t ) xt 4/13/2015 KyungHee University 49 3.10 Convolution Property cont’ • 3.10.2 FILTERING Continuous time Discrete time LPF HPF BPF Figure 3.53 (p. 263) Frequency dependent gain (power spectrum) 20 log H jw or 20 log H e j . Y jw H jw X jw , 2 4/13/2015 2 2 kill or not (magnitude) KyungHee University 50 3.10 Convolution Property cont’ Example 3.34 Identifying h(t) from x(t) and y(t) t The output of an LTI system in response to an input xt e2tut is yt e ut . Find frequency response and the impulse response of this system. Solution: X jw H jw 4/13/2015 1 jw 2 Y jw 1 . jw 1 jw 2 jw 1 1 1 1 . jw 1 jw 1 jw 1 jw 1 Y jw But H jw X jw ht t e t ut . KyungHee University 51 3.10 Convolution Property cont’ EXAMPLE 3.35 Equalization of multipath channel X jw H inv jwY jw or X e j H inv e j Y e j , Consider again the problem addressed in Example 2.13. In this problem, a distorted received signal y[n] is expressed in terms of a transmitted signal x[n] as yn xn axn 1, a 1. hinv n* hn n. Then H inv e j 1, hn a, 0, n0 n 1 . otherwise H inv e j H e j 1, 1 . H e j DTFT hn H e j 1 ae j . H inv e j 1 . 1 ae j hinv n a un. n 4/13/2015 KyungHee University 52 3.10 Convolution Property cont’ • 3.10.3 Convolution of periodic signals : Cyclic convolution Convolution in just one period Or better to derive in the frequency domain yt xt zt x zt d T 0 2 FS ; T yt xt zt Y k T X k Z k . 3.44 EXAMPLE 3.36 Convolution of 2 periodic signals Evaluate the periodic convolution of the sinusoidal signal zt 2 cos2t sin4t 1, 1 2 j , Z k 1 2 j , 0, X k k 1 k 2 . k 2 otherwize 2 sin k 2 . k 2 Figure 3.56 (p. 268) Square wave for Example 3.36. 4/13/2015 KyungHee University 53 3.10 Convolution Property cont’ 1 , Y k X k Z k 0, k 1 , otherwise yt 2 cos2t . Xˆ j k X k Z k , 1, J k J 0, otherwise Z k N 1 z t sin t 2 J 1 . 3.45 sin t 2 N yn xn zn xk zn k Y k N X k Z k DTFS ; k 0 Figure 3.57 4/13/2015 (p. 270) z(t) in Eq. (3.45) when J = 10. KyungHee University 54 3.11 Differentiation and Integration Properties • 3.11.1 DIFFERENTIATION IN TIME 1 xt 2 X jwe jwt dw. d 1 xt dt 2 FT X jw jwe jwt dw jwX jw. EXAMPLE 3.37 The differentiation property implies that d at jw FT e u t . dt a jw d at e u t ae at u t e at t ae at u t . dt d at a jw FT e ut 1 dt a jw a jw 4/13/2015 KyungHee University 55 3.11 Differentiation and Integration Properties cont’ Example 3.38 Resonance in MEMS accelerometer The MEMS accelerometer introduced in Section 1.10 is described by the differential equation d2 w d yt n yt wn2 yt xt . 2 dt Q dt H jw 1 . wn 2 2 jw jw wn Q Fig 3.58 (p. 273) 20log10 | H jw | n = 10,000 rad/s Q = 2/5, Q = 1, and Q = 200. Resonance in 10,000 rad/s 4/13/2015 KyungHee University 56 3.11 Differentiation and Integration Properties cont’ Example 3.39 Use the differentiation property to find the FS of the triangular wave depicted in Fig. 3.59(a) 0, FS ;w0 z t Z k 4 sin k 2 k k 0 , . Y [k ] ??? k0 Signals for (a) Triangular wave y(t). (b) T 2 , k FS ; w0 y t Y k 2T sin 2 2 2 jk dy (t ) z (t ) dt k 0 . , k 0 where j 2k j, T If we differentiate z(t) once more??? 4/13/2015 KyungHee University 57 3.11 Differentiation and Integration Properties cont’ • 3.11.2 DIFFERENTIATION IN FREQUENCY X jw xt e jwt dt, Differentiate w.r.t. ω, d X jw jtxt e jwt dt , dw d FT Then, jtxt X jw. dw Example 3.40 FT of a Gaussian pulse Use the differentiation-in-time and differentiation-in-frequency properties for the FT of the Gaussian pulse, defined by gt 1 2 et 2 and depicted in Fig. 3.60. d g t t dt 2 e t 2 2 2 tg t . 3.48 d FT g t tg (t ) jwG jw, dt 1 d FT G jw tg t j dw FT and jtg t dw G jw wG jw. d Then G jw ce w 2 . (But, c=?) G j 0 1 2 Figure 3.60 (p. 275) Gaussian pulse g(t). 4/13/2015 1 2 et 2 2 d G jw dw 2 e t 2 dt 1. 2 FT e w 2 . KyungHee University 2 58 3.11 Differentiation and Integration Properties cont’ • 3.11.3 Integration yt x d ; t t FT x d Y jw 1 X jw. jw 3.52 1 X jw X j 0 w. jw 3.53 FT Ex) ut d . Prove ut U jw jw w. 1 t u t 1 1 sgn t . 2 2 1, sgn t 0, 1, 3.54 Note u(t ) e at u(t ) where a=0 t0 FT 2 w. We know 1 t 0. t0 d sgn t 2 t . jwSgn jw 2. dt Fig. a step fn. as the sum of a constant and a signum fn. 4/13/2015 KyungHee University 2 , 0 S ( j ) j since linear 0, 0 FT u t 1 j 59 3.11 Differentiation and Integration Properties cont’ Common Differentiation and Integration Properties. d FT xt jX j dt d FS ;0 xt jk0 X k dt d FT jtxt X j d d DTFT jnxn X e j d t 1 FT x d j X j X j 0 4/13/2015 KyungHee University 60 3.12 Time-and Frequency-Shift Properties • 3.12.1 Time-Shift Property jt Z j zt e xt t0 e jt dt x e j t0 d e jt0 x e j d e jt0 X j arg[e j t0 X j ] t0 arg[X ( j)] Table 3.7 Time-Shift Properties of Fourier Representations FT xt t0 e jt0 X j FS ;0 xt t0 e jk0t0 X k DTFT xn n0 e jn0 X e j DTFS ;0 xn n0 e jk0n0 X k 4/13/2015 KyungHee University 61 3.12 Time-and Frequency-Shift Properties cont’ Example) Figure 3.62 X j 4/13/2015 2 z t xt T1 Z j e jT X j 1 sin T0 Z j e jT 1 KyungHee University 2 sin T 62 3.12 Time-and Frequency-Shift Properties cont’ • 3.12.2 Frequency-Shift Property 1 j t z t Z j e d 2 1 X j e jt d 2 Recall z t 1 2 e jt X j e j t d 1 2 X j e jt d e jt xt FT xt t0 e jt0 X j Table 3.8 Frequency-Shift Properties FT e jt xt X j FS ;0 e jk00t xt X k k0 DTFT e jn xn x e j DTFS ;0 e jk00n xn X k k0 4/13/2015 KyungHee University 63 3.12 Time-and Frequency-Shift Properties cont’ Example 3.42 FT by Using the Frequency-Shift Property e j10t , t z t t 0, Solution: We may express zt as the product of a complex sinusoid e j10t and a rectangular pulse 1, t xt 0, t FT xt X j 2 zt e j10t x(t ) sin FT X j 10 z t e j10t x(t ) 4/13/2015 2 sin 10 10 KyungHee University 64 3.12 Time-and Frequency-Shift Properties cont’ Example 3.43 Using Multiple Properties to Find an FT xt d 3t e u t e t u t 2 dt Sol) Let wt e3t ut and vt et ut 2 Then we may write xt d wt vt dt By the convolution and differentiation properties X j jW j V j The transform pair FT e at u t vt e e 1 a j W j 1 3 j e j 2 ut 2 V j e 1 j 2 t 2 2 je j 2 X j e 1 j 3 j 2 4/13/2015 KyungHee University 65 3.13 Inverse FT by Using PartialFraction Expansions • 3.13.1 Inverse FT by using FT e at u t 1 a j bM j b1 j b0 B j X j j N aN 1 j N 1 a1 j a0 A j ~ M N B j k X j f k j A j k 0 M Let v j, then v N aN 1v N 1 a1v a0 0 N roots, d k M X j for k 1,, N b j k 0 N k N k j d partial fraction X j k 1 k Cx j d k k 1 FT e dt u t 1 j d for d 0 N N xt Ck e ut X j dk t k 1 4/13/2015 FT k 1 Ck j d k KyungHee University 66 3.13 Inverse FT by Using PartialFraction Expansions cont’ • 3.13.2 Inverse Discrete-Time Fourier Transform M e jM 1e j 0 X e N e jN N 1e j( N 1) 1e j 1 j Ne jN N 1e j N 1 1e j N 1 1 d k e j k 1 v N 1v N 1 2v N 2 N 1v N 0 1 dC e X e j N k k 1 j where k N DTFT d k n un 1 1 d k e j Then xn Ck d k un n k 1 4/13/2015 KyungHee University 67 3.13 Inverse FT by Using PartialFraction Expansions cont’ Example 3.45 Inversion by Partial-Fraction Expansion X e j 5 e j 5 6 1 1 1 e j e j 2 6 6 Solution: 5 e j 5 C1 C2 6 1 1 1 1 1 e j e j 2 1 e j 1 e j 6 6 2 3 Using the method of residues described in Appendix B, We obtain And 5 e j 5 1 6 C1 1 e j 2 1 1 e j 1 e j 2 6 6 5 e j 5 1 6 C2 1 e j 1 3 1 e j 1 e j 2 6 6 5 e j 5 6 4 1 j 1 e 3 e j 2 e j 2 5 e j 5 6 1 1 j 1 e j 3 e 3 e j 3 Hence, xn 41/ 2 un 1/ 3 un n 4/13/2015 n KyungHee University 68 3.14 Multiplication (modulation) Property Given xt 1 2 yt x(t ) z (t ) X jv e jvt dv 1 2 2 and z t 1 2 Z j e jt d X jv Z j e j v t ddv Change of variable to obtain yt 1 2 1 2 X jv Z j v dve jt d FT y t xt z t Y j 1 X j Z j 2 (3.56) Where X j Z j X jv Z j v dv DTFT yn xnzn Y e j 1 X e j Z e j 2 (3.57) denotes periodic convolution. Here, X e j and Z e j are 2 -periodic. X e j Z e j X e j Z e j d 4/13/2015 KyungHee University 69 3.14 Multiplication (modulation) Property cont’ Example) Windowing in the time domain FT y t x(t ) w(t ) Y j W j 4/13/2015 2 1 X j W j 2 sin T0 Figure 3.66 The effect of Truncating the impulse response of a discrete-time system. (a) Frequency response of ideal system. (b) F for near zero. (c) F for slightly greater than / 2 (d) Frequency response of system with truncated impulse response. KyungHee University 70 3.14 Multiplication (modulation) Property cont’ Example 3.46 Truncating the sinc function Sol) hn 1, n M 1 n sin truncated by wn n 2 0, otherwise hn, n M ht n otherwise 0, H t e j hn 1 2 H e W e j j d 1, / 2 1 n sin H e j n 2 0, / 2 W e j H t e j sin 2M 1 / 2 sin / 2 1 2 /2 /2 F H e W e j 4/13/2015 DTFT ht n Ht e j F d j j , W e 0, /2 /2 KyungHee University 71 3.14 Multiplication (modulation) Property cont’ Example 3.47 Radar Range Measurement: RF Pulse Train k 500 12 j , Solution) sin(2 500t ) S k 12 j , k 500 Pk e jkT00 X k sin k0T0 k 0, otherwise x(t ) p(t ) sin(2 500t ) X [k ] P[k ] * S[k ] 1 j k 500 T00 sin k 5000T0 1 j k 500 T00 sin k 5000T0 e e k 500 k 500 2j 2j DTFT ; 2 / N yn xnzn Y k X k Z k (3.59) N 1 X k Z k X mZ k m m0 4/13/2015 KyungHee University 72 3.14 Multiplication (modulation) Property cont’ Figure 3.69 FS magnitude spectrum of FR pulse train for 0 k 1000 The result is depicted as a continuous curve, due to the difficulty of displaying 1000 steps Table 3.9 Multiplication Properties FT xt z t 1 X j Z j 2 FS ;0 xt z t X k Z k 1 DTFT xnzn X e j Z e j 2 DTFS ;0 xnzn X k Z k 4/13/2015 KyungHee University 73 3.15 Scaling Properties Z j zt e jt xate jt dt 1 / a x e j / a d , a 0 Z j 1 / a x e j / a d , a 0 Z j 1 / a x e j / a d FT 1/ a X j / a zt xat 4/13/2015 (3.60) KyungHee University 74 3.15 Scaling Properties cont’ Example 3.48 SCALING A RECTANGULAR PULSE Let the rectangular pulse 1, t 1 x(t ) 0 , t 1 t 1, t 2 y (t ) x( ) 2 0, t 2 Solution : X ( jw) 2 sin( w). y(t ) x(t / 2). a 1 / 2 w Y ( jw) 2 X (i 2w) 2 2( ) sin(2w) 2w 2 sin(2w). w 4/13/2015 KyungHee University 75 3.15 Scaling Properties cont’ d e j 2w Example 3.49 Multiple FT Properties for x(t) when X ( jw) j { }. dw 1 j (w / 3) Solution) s(t ) e t u (t ) S ( jw) FT X ( jw) j 1 1 jw d {e j 2 w S ( jw / 3)}. dw we define Y ( jw) S ( jw / 3) y(t ) 3s(3t ) 3e3t u(3t ) 3e 3t u(t ). Now we define W ( jw) e j 2wY ( jw) w(t ) y(t 2) 3e3(t 2)u(t 2). Finally, since X ( jw) j d W ( jw),. dw x(t ) tw(t ) 3te 3(t 2)u(t 2). 4/13/2015 KyungHee University 76 3.16 Parseval’s Relationships 언제 꽝? | X ( jw) |2 | x(t ) |2 Wx Table 3.10 Parseval Relationships for the Four Fourier Representations Represent ation | x(t ) |2 dt, Note that | x(t ) |2 x(t ) x * (t ). 1 x (t ) 2 * Wx x(t )[ X * ( jw)e jwt dw. 1 2 X * ( jw)e jwt dw]dt. 1 * Wx X ( jw) X ( jw) dw 2 So concludethat 1 2 2 | x ( t ) | dt | X ( jw ) | dw. 2 4/13/2015 바이올린 주자가쓴 에너지? Parseval Relation 1 | X ( jw) | dw 2 FS 1 T 2 2 | x ( t ) | dt | X [ k ] | T O K DTFT DTFS KyungHee University | x[n] |2 n 1 N N 1 1 2 FT 2 | X (e j ) | 2 d N 1 | x[n] | | X [k ] | 2 K 0 | X ( jw) | 2 dw 2 k 0 77 3.16 Parseval’s Relationships cont’ Example 3.50 Calculate the energy in a signal sin(Wn) Let x[n] n E∞= | x[n]| 2 n sin 2 (Wn ) . 2 2 n n Use the Parseval’s theorem Solution) E 1 2 1 E 2 W | X (e j ) |2 d. 1d W / . W 1, | | W x[n] X (e ) { 0, W | | DTFT 4/13/2015 j KyungHee University 78 3.17 Time –Bandwidth Product 1, | t | To FT x(t ) { X ( jw) 2 sin(wTo ) / w. 0, | t | To Figure 3.72 (p. 305) Rectangular pulse illustrating the inverse relationship between the time and frequency extent of a signal. Td [ t 2 | x(t ) |2 dt 2 | x(t ) | dt ]1/ 2 Bw [ w2 | X ( jw) |2 dw 1 / 2 ] . 2 | X ( jw) | dw Td Bw 1 / 2. 4/13/2015 KyungHee University 79 3.17 Time –Bandwidth Product cont’ Example 3.51 Bounding the Bandwidth of a Rectangular Pulse 1, | t | T Let x(t ) 0 , | t | T o Use the uncertainty principle to place a lower bound on the effective bandwidth of x(t). Solution: To t 2 dt T Td Too Todt 1/ 2 (1/(2To ))(1/ 3)t 3 |To To 1/ 2 To / 3. Bw 3 /(2To ). 4/13/2015 KyungHee University 80 3.18.1 The Duality Property of the FT 1 x(t ) 2 X ( jw)e jwt dw. X ( jw) x(t )e jwt dt . Fig. 3.73 Duality of rectangular pulses and sinc functions 1 z ( )e jv dv. 2 if we choose v t and w, then Eq. (3.66) im plies that 1 y (t ) z ( w)e jwt dw. 2 FT y (t ) z ( w). y (v ) 1 z (t )e jwt dt, 2 FT z (t ) 2y ( w), y ( w) 4/13/2015 FT f (t ) F ( jw), FT F ( jt ) 2f ( w). KyungHee University 81 3.18.1 The Duality Property of the FT cont’ Example 3.52 By using Duality, find the FT of x(t ) 1 1 jt Solution) FT FT f (t ) F ( jw) F ( jt ) 2f (w). FT f (t ) et u (t ) F ( jw) 4/13/2015 1 1 jw F ( jt ) 1 1 jt KyungHee University 82 3.18.2 & 3.18.3 • 3.18.2 The Duality Property of the DTFS N 1 x[n] X [k ]e jk o n K 0 1 X [ k ] . N DTFS ; 2 / N x[n] X [k ] N 1 x[n]e jkon n 0 DTFS ; 2 / N X [k ] 1 x[k ] N • 3.18.3 The Duality Property of the DTFT and FS z (t ) Z [k ]e jkw0t Z [k ] 1 2 k j X (e ) z (t )e jkt dt x[n]e jn . n Table 3.11 Duality Properties of Fourier Representations FT f (t ) F ( jw) FT DTFS DTFS ; 2 / N x[n] X [k ] DTFS x[n] X (e j ) DTFS x[n] 4/13/2015 1 2 X (e j )e jn . FT F ( jt ) 2f (w) DTFS ; 2 / N X [n] (1/ N ) x[k ] FS ;1 X (e jt ) x[k ] DTFT FS ;1 x[n] X (e j ) X (e jt ) x[k ]. KyungHee University 83 3.18.3 The Duality Property of the DTFT and FS cont’ Example 3.53 FS-DTFT Duality Use the duality property and the results of Example 3.39 to determine the inverse DTFT of the triangular spectrum X (e j ) depicted in Fig 3.75(a). Solution: Z [k ] e 4/13/2015 jw / 2 Y [k ] , n0 k 1 4 j sin(k / 2) , k 2 n0 KyungHee University 84