Orthonormal Basis Functions A set of signals {0 (t ),, n (t )} are called orthogonal on an interval (a, b) if any two signals m (t ) and k (t ) in the set satisfy 0, a m (t ) k (t )dt , b * mk mk If the magnitude of each signal i (t ) is set to one, it is called normalized. A set of normalized orthogonal functions can form an orthonormal basis set。 The complex exponential signals {e j 2nf t , n 0, 1, 2, 3, } are orthogonal on any interval over a period T0 1/ f0 . 教育部網路通訊人才培育先導型計畫 0 Principles of Communications 2.3 1 Generalized Transformation of Signal A signal x(t) on any interval over a period T0, i.e., (t0 , t0+T0 ), can be expressed as following in terms of {0 (t ),, n (t ),} x(t ) cnn (t ), t0 t t0 T0 n 0 1 cn T0 t0 T0 t0 x(t ) n* (t )dt Parseval theorem:(Will be discussed later) 1 P T0 T0 x(t ) dt |cn |2 , 2 教育部網路通訊人才培育先導型計畫 n 0 T0 : Integrate over a period. Principles of Communications 2.3 2 Fourier theory Jean Baptiste Joseph Fourier (1768-1830) proposed that a periodic signal can be represented by a summation of a (possibly infinite) number of sinusoids each with a particular amplitude and phase. Assume that x(t) is a periodic signal with fundamental period T0, then x(t) can be represented as x(t ) a0 [an cosn0t bn sin n0t ], n 1 or x(t ) a0 [an cos 2nf0t bn sin 2nf0t ], n 1 0 2f 0 2 T0 The series is called Trigonometric Fourier series because it is represented in terms of sinusoids. 教育部網路通訊人才培育先導型計畫 Signals and Systems 4.3 3 Coefficients of trigonometric Fourier series (1/7) Given a periodic signal with fundamental period T0 that can be represented by the trigonometric Fourier series x(t ) a0 [an cos n0t bn sin n0t ] n 1 The problem is how to find the coefficients a0, an and bn. We begin with a0 Integrating the series term by term over one period T0, we obtain T0 x(t )dt T0 0 0 a0 dt a1 cos0tdt a2 cos 20tdt T0 T0 b1 sin 0tdt b2 sin 20tdt T0 T0 0 0 Note:Integrating the sinusoidal over one or integer number of periods is 0. 教育部網路通訊人才培育先導型計畫 Signals and Systems 4.3 4 Coefficients of trigonometric Fourier series (2/7) T0 x(t )dt a0T0 1 a0 T0 T0 x(t )dt a0 is the average value of the waveform (signal). Find coefficients an Multiplying both sides of the series equation by cos k0t and integrating over one period T0, we obtain a0 cos k0tdt an cos n0t cos k0tdt T0 x(t ) cosk0tdt T0 T0 n 1 0 bn sin n0t cos k0tdt T0 n 1 教育部網路通訊人才培育先導型計畫 Signals and Systems 4.3 5 Coefficients of trigonometric Fourier series (3/7) x(t ) cos k0tdt [an cos n0t cos k0tdt bn sin n0t cos k0t dt] T0 T0 T0 n 1 I1 0 T0 n k I3 0 nk 2 The terms I1 and I3 will be discussed in details later T0 x(t ) cos k0tdt ak T0 2 2 2 ak x(t ) cos k0tdt an T0 T0 T0 教育部網路通訊人才培育先導型計畫 T0 x(t ) cos n0tdt Signals and Systems 4.3 6 Coefficients of trigonometric Fourier series (4/7) Find coefficients bn Similarly we can obtain x(t ) sin k0tdt a0 sin k0tdt an cos n0t sin k0tdt T0 T0 T0 n 1 0 bn sin n0t sin k0tdt T0 n 1 x(t ) sin k0tdt [an cosn0t sin k0tdt bn sin n0t sin k0t dt] T0 T0 T0 n 1 I1 0 x(t ) sin k0tdt bk T0 2 bk T0 T0 T0 2 x(t ) sin k0tdt 教育部網路通訊人才培育先導型計畫 0 nk I 2 T0 nk 2 2 bn T0 I2 will be discussed in details later T0 x(t ) sin n0tdt Signals and Systems 4.3 7 Coefficients of trigonometric Fourier series (5/7) In the procedure of finding coefficients an and bn , we used the properties of integrals involving products of sines and cosines. Consider I1 cosn0t sin k0tdt T0 0 0 1 n k , I [ sin( n k )0tdt sin( n k )0tdt ] 0 If 1 T0 2 T0 1 If n k , I1 cos k0t sin k0tdt sin 2k0tdt 0 T0 2 T0 I1 cosn0t sin k0tdt 0 T0 教育部網路通訊人才培育先導型計畫 Signals and Systems 4.3 8 Coefficients of trigonometric Fourier series (6/7) Consider I 2 sin n0t sin k0tdt T0 If n k, 0 0 1 I 2 [ cos( n k )0tdt cos( n k )0tdt ] 0 T0 2 T0 If n k, I 2 sin k0t sin k0tdt sin 2 k0tdt T0 T0 0 1 cos 2k0t T0 1 1 dt dt cos 2k0tdt T0 T0 2 T0 2 2 2 0, n k I 2 T0 2 , nk 教育部網路通訊人才培育先導型計畫 Signals and Systems 4.3 9 Coefficients of trigonometric Fourier series (7/7 ) Consider I 3 cos n0t cosk0tdt T0 If n k, If n k, 0 0 1 I 3 [ cos( n k )0tdt cos( n k )0tdt ] 0 T0 2 T0 1 cos 2k0t I 3 cos k0tdt dt T0 T0 2 0 T0 1 1 dt cos 2k0tdt T0 2 T0 2 2 2 0, n k I 3 T0 2 , nk 教育部網路通訊人才培育先導型計畫 Signals and Systems 4.3 10 Trigonometric Fourier series Given a periodic signal with period T0 that can be represented by the trigonometric Fourier series, x(t ) a0 [an cos n0t bn sin n0t ] n 1 where 1 a0 T0 x(t )dt T0 the average value of the signal 2 an T0 x(t ) cos n0tdt 2 bn T0 x(t ) sin n0tdt 教育部網路通訊人才培育先導型計畫 T0 T0 n0 n0 Signals and Systems 4.3 11 Harmonic form Fourier series By using the Trigonometric equality an cosn0t bn sin n0t cn cos(n0t n ) where we have cn an bn ; 2 2 n tan1 ( bn ) an x(t ) a0 (an cos n0t bn sin n0t ) n 1 a0 cn cos(n0t n ) n 1 Let c0 = a0 and we have the harmonic form Fourier series of x(t) x(t ) c0 cn cos(n0t n ) n 1 教育部網路通訊人才培育先導型計畫 Signals and Systems 4.3 12 Limits of Fourier series at the discontinuities The Fourier series of waveform converges to the mean of the right- and lefthand limits at the discontinue point. x(t ) a0 (an cos n0t1 bn sin n0t1 ) n 1 x(t ) limits of Fourier series b x(t1 ) a or b a Discontinue at t = t1+ nT0 t1 T0 t1 t1 2T0 t1 3T0 t Limits of Fourier series ab a0 (an cos n0t1 bn sin n0t1 ) 2 n 1 教育部網路通訊人才培育先導型計畫 Signals and Systems 4.3 13 Dirichlet conditions for Fourier series A periodic signal x(t) can be represented as a Fourier series only if it satisfy the following Dirichlet conditions: x(t) is absolutely integrable over any period, that is t1 T0 t1 x(t ) dt , for any t1 x(t) has a finite number of maxima and minima within any finite interval of t. x(t) has a finite number of discontinuities within any finite interval of t, and each of these discontinuities is finite. 教育部網路通訊人才培育先導型計畫 Signals and Systems 4.3 14 Example 4-9 (1/6) Please find the trigonometric Fourier series for the periodic rectangular pulse train signal with period 2. x(t) 1 3 2 教育部網路通訊人才培育先導型計畫 /2 /2 2 3 Signals and Systems 4.3 t 15 Example 4-9 (2/6) 【Sol.】 1) The trigonometric Fourier series with 0 = 2 /T0= 2 /2 = 1 x(t ) a0 [an cos nt bn sin nt] n 1 2) Find a0 1 a0 T0 教育部網路通訊人才培育先導型計畫 1 T0 x(t )dt 2 2 2 1 1dt 2 Signals and Systems 4.3 16 Example 4-9 (3/6) 3) Find an 2 an T0 T0 x(t ) cos ntdt 2 n cos ntdt sin( ) 2 n 2 1 2 0, 2 , n 2 , n 教育部網路通訊人才培育先導型計畫 n is even n 1,5,9,13,... n 3,7,11,15,... Signals and Systems 4.3 17 Example 4-9 (4/6) 4) Find bn 2 bn T0 T0 x(t ) sin ntdt 5) The signal x(t) is written by x(t ) 1 2 sin ntdt 0 2 cos x cos(x ) 1 2 1 1 1 (cos t cos 3t cos 5t cos 7t ...) 2 3 5 7 1 2 1 1 1 [cost cos(3t ) cos5t cos(7t ) ...] 2 3 5 7 c0 cn cos(nt n ) n 1 教育部網路通訊人才培育先導型計畫 Signals and Systems 4.3 18 Example 4-9 (5/6) 6) The trigonometric Fourier series for the periodic rectangular pulse train signal with period 2 is obtained. x(t ) c0 cn cos(nt n ) n 1 1 2 0, cn 2 , n , n 0, c0 教育部網路通訊人才培育先導型計畫 n is even n is even n 3,7,11,15,... else Signals and Systems 4.3 19 Example 4-9 (6/6) Plot the series in frequency domain. 1 2 1 1 1 x(t ) [cos t cos( 3t ) cos 5t cos( 7t ) ...] 2 3 5 7 Note: The phases are either 0 or , cne jn 教育部網路通訊人才培育先導型計畫 the amplitude and phase are combined in this special case. Signals and Systems 4.3 20 Example 4-10 (1/6) Please find the trigonometric Fourier series for the periodic triangle pulse train signal with period 2. x(t ) A 0 t A 教育部網路通訊人才培育先導型計畫 Signals and Systems 4.3 21 Example 4-10 (2/6) 【Sol.】 1) The trigonometric Fourier series with 0 = 2 /T0= 2 /2 = . x(t ) a0 [an cos nt bn sin nt] n 1 2) Set the period as [1/2, 3/2], then x(t) in this period is written as 2 At, x(t ) 2 A(1 t ), 1 t 2 1 3 t 2 2 3) Obtain a0 = 0 . (the average of x(t) is 0) 教育部網路通訊人才培育先導型計畫 Signals and Systems 4.3 22 Example 4-10 (3/6) 4) Find an 2 an T0 T0 x(t ) cos ntdt 2 3/ 2 x(t ) cos ntdt 2 1/ 2 1/ 2 1 / 2 3/ 2 2 At cos ntdt 1/ 2 2 A(1 t ) cos ntdt 0 Note: use the odd/even properties 教育部網路通訊人才培育先導型計畫 Signals and Systems 4.3 23 Example 4-10 (4/6) 4) Find bn bn 2 T0 T0 x(t ) sin(nt )dt 2 3/ 2 x(t ) sin(nt )dt 2 1/ 2 1/ 2 1 / 2 2 At sin(nt )dt 3/ 2 1/ 2 2 A(1 t ) sin(nt )dt 0, 8A n 8A 2 2 sin( ) 2 2 , n 2 n 8A 2 2, n 教育部網路通訊人才培育先導型計畫 n is even n 1, 5, 9, 13.... n 3, 7, 11, 15.... Signals and Systems 4.3 24 Example 4-10 (5/6) 5) The signal x(t) is written by 8A 1 1 1 x(t ) 2 [sin t sin 3t sin 5t sin 7t ...) 9 25 49 Using sin kt cos(kT 900 ) , we rewrite x(t) as 8A 1 x(t ) 2 [cos(t 90 ) cos(3t 900 ) 9 1 1 cos(5t 900 ) cos(7t 900 ) ...] 25 49 教育部網路通訊人才培育先導型計畫 0 Signals and Systems 4.3 25 Example 4-10 (6/6) The signal is expressed as harmonic form Fourier series and plotted in frequency domain. x(t ) c0 cn cos(nt n ) n 1 Only odd order harmonic (n times of fundamental frequency) components exist. 教育部網路通訊人才培育先導型計畫 Signals and Systems 4.3 26 Exponential Fourier series Assume that x(t) is a periodic signal with fundamental period T0, then x(t) can be represented as the exponential Fourier series, x(t ) Χ 2 e j 20t Χ 1e j0t Χ 0 Χ 1e j0t Χ 2 e j 20t jn0t Χ e n n Χ n n e j 2nf 0t 教育部網路通訊人才培育先導型計畫 , 2 0 2f 0 T0 Signals and Systems 4.4 27 Coefficients of exponential Fourier series (1/6) Review the trigonometric Fourier series x(t ) a0 (an cosn0t bn sin n0t ) n 1 Review the Euler’s equality e jn0t e jn0t sin n0t 2j e jn0t e jn0t cos n0t 2 教育部網路通訊人才培育先導型計畫 Signals and Systems 4.4 28 Coefficients of exponential Fourier series (2/6) By using Euler’s equality, we rewrite the trigonometric Fourier series and obtain e jn0t e jn0t e jn0t e jn0t x(t ) a0 (an bn ) 2 2j n 1 1 1 a0 (an jbn )e jn0t (an jbn )e jn0t n 1 2 n 1 2 1 1 jn0t (an jbn )e a0 (an jbn )e jn0t n 1 2 n 1 2 x(t ) ... Χ 2 e j 20t Χ 1e j0t Χ 0 Χ 1e j0t Χ 2 e j 20t ... jn0t Χ e n n 教育部網路通訊人才培育先導型計畫 Signals and Systems 4.4 29 Coefficients of exponential Fourier series (3/6) The relationship between the coefficients of exponential Fourier series and those of trigonometric Fourier series 1 2 ( a n jb n ), Χ n a0 , 1 ( an jbn ), 2 教育部網路通訊人才培育先導型計畫 n0 n0 n0 Signals and Systems 4.4 30 Coefficients of exponential Fourier series (4/6) Review the harmonic form Fourier series x(t ) c0 cn cos(n0t n ) n 1 By using Euler’s equality, we rewrite the trigonometric Fourier series and obtain cn j ( n0t n ) j ( n0t n ) [e e ] 2 c c ( n e j n )e jn0t ( n e j n )e jn0t 2 2 cn j n jn0t cn j n jn0t x(t ) c0 ( e )e ( e )e n 1 2 n 1 2 cn cos(n0t n ) e n 教育部網路通訊人才培育先導型計畫 jn0t n Signals and Systems 4.4 31 Coefficients of exponential Fourier series (5/6) The relationship between the coefficients of exponential Fourier series and those of trigonometric Fourier series: c n j n , 2 e Χ n c0 , c n e j n , 2 n0 n0 n0 The exponential Fourier series and trigonometric Fourier series are equivalent. 教育部網路通訊人才培育先導型計畫 Signals and Systems 4.4 32 Coefficients of exponential Fourier series (6/6) An alternative way to find the coefficients of exponential Fourier series. Multiplying both sides of the series equation by e - jk 0t and integrating over one period T0, we obtain T0 x(t )e - jk0t dt T0 ( Χ n e jn0t )e - jk0t dt n n Χ n e j ( n k )0t dt T0 Χ k T0 1 Χk T0 x(t )e jk0t dt 1 Χn T0 x(t )e jn0t dt T0 T0 教育部網路通訊人才培育先導型計畫 T0 e j ( n k ) 0 t 0, n k dt T0 , n k Signals and Systems 4.4 33 Example 4-11 (1/6) Please find the exponential Fourier series for the periodic rectangular pulse train signal with period T0. . T0 教育部網路通訊人才培育先導型計畫 T0 Signals and Systems 4.4 34 Example 4-11 (2/6) 【Sol.】 1) The exponential Fourier series with 0 = 2 /T0= 2 f0. x(t ) Χ ne jn0t n Χ ne j 2nf 0t , 0 2f 0 n 2 T0 2) Find X0 1 Χ0 T0 教育部網路通訊人才培育先導型計畫 1 T0 / 2 x(t )dt T0 T0 / 2 1 T0 / 41dt 2 T0 / 4 Signals and Systems 4.4 35 Example 4-11 (4/6) 3) Find Xn 1 Χn T0 T0 / 2 -T0 / 2 x(t )e jn0t 1 dt T0 T0 / 4 - T0 / 4 e jn0t dt jn / 2 jn / 2 1 1 e e [e jn / 2 e jn / 2 ] [ ] j 2n n j2 0, sin(n / 2) 1 k (1) , n (2k 1) 教育部網路通訊人才培育先導型計畫 n 2k 0 n 2k 1 Signals and Systems 4.4 36 Example 4-11 (5/6) 4) Rewrite the coefficients 1 Χ0 ; 2 Χ 2 k 0, k 0; Χ 2 k 1 (1) k 1 (2k 1) 5) The exponential Fourier series x(t ) Χe n jn0t n 1 (1) k e j ( 2 k 1)0t 2 k (2k 1) or the trigonometric Fourier series 1 2 x(t ) (1) k cos[2 (2k 1) f 0 t ] 2 k 0 (2k 1) 1 2 1 1 1 [cos(2f 0 t ) cos(6f 0 t ) cos(10f 0 t ) cos(14f 0 t ) ] 2 3 5 7 教育部網路通訊人才培育先導型計畫 Signals and Systems 4.4 37 Example 4-11 (6/6) Double-sided spectrums of the periodic rectangular pulse train signal. Double-sided amplitude spectrum Double-sided phase spectrum 教育部網路通訊人才培育先導型計畫 Signals and Systems 4.4 38 Example 4-12 (1/2) Please find the exponential Fourier series for the signal xc (t ) 3e j ( 2000t / 6) 4e j ( 4000t / 3) e j (6000t / 6) 【Sol.】 1) The signal consists of 3 complex exponential components with frequencies 1000, 2000 and 3000, respectively. The fundamental frequency of xc(t) is given by GCD(1000, 2000, 3000) = 1000. 2) The expression of xc(t) is in the form of exponential Fourier series. 教育部網路通訊人才培育先導型計畫 Signals and Systems 4.4 39 Example 4-12 (2/2) 3) Rewrite xc(t) xc (t ) 3e j ( 2000t / 6) 4e j ( 4000t / 3) e j ( 6000t / 6) 3e j / 6e j 2000t 4e j / 3e j 4000t e j / 6e j 6000t 3e j / 6e j 2f 0t 4e j / 3e j 2 2 f 0t e j / 6e j 2 3 f 0t , f 0 1000 3e j / 6e j0t 4e j / 3e j 20t e j / 6e j 30t , 0 2000 x(t ) e n n jn0t e n n j 2nf 0t , 2 0 2f 0 T0 Χ1 3e j / 6 ; Χ 2 4e j / 3 ; Χ 3 e j / 6 ; Χ n 0, n 1, 2, 3 Note that xc(t) is a complex signal and thus the symmetry property of Fourier series discussed later is not held. 教育部網路通訊人才培育先導型計畫 Signals and Systems 4.4 40 Example 4-13 (1/5) Please find the exponential Fourier series for the periodic signal with period T0. A sin 0t , x(t ) 0, 0 t T0 / 2 T0 / 2 t T0 A T0 / 2 教育部網路通訊人才培育先導型計畫 T0 Signals and Systems 4.4 41 Example 4-13 (2/5) 【Sol.】 1) The exponential Fourier series with 0 = 2 /T0. x(t ) Χ n 2) Find Xn n e jn0t 1 Χn T0 , T0 / 2 0 A 2 jT0 2 0 T0 A sin 0te jn0t dt T0 2 0 (e j0t e j0t )e jn0t dt T0 2 T0 2 A j0 (1 n ) t [ e dt e j0 (1 n )t dt] 0 2 jT0 0 A e j (1 n ) 1 e j (1 n ) 1 [ ] , n 1 4 1 n 1 n 教育部網路通訊人才培育先導型計畫 Signals and Systems 4.4 42 Example 4-13 (3/5) 【Sol.】 1) The trigonometric Fourier series with 0 = 2 /T0. x(t ) Χ n 2) Find Xn n e jn0t 1 n T0 , T0 / 2 0 A 2 jT0 2 0 T0 A sin 0te jn0t dt T0 2 0 (e j0t e j0t )e jn0t dt T0 2 T0 2 A j0 (1 n ) t [ e dt e j0 (1 n )t dt] 0 2 jT0 0 A e j (1 n ) 1 e j (1 n ) 1 [ ] , n 1 4 1 n 1 n 教育部網路通訊人才培育先導型計畫 Signals and Systems 4.4 43 Example 4-13 (4/5) Using e j (1 n ) cos(1 n) j sin(1 n) (1)n we obtain the coefficients 0, A 1 , 2 1 n Χn A j , 4 A j , 4 教育部網路通訊人才培育先導型計畫 n is odd and n 1 n is even n 1 see next page n 1 Signals and Systems 4.4 44 Example 4-13 (5/5) A Χ1 2 jT0 A 2 jT0 T0 2 0 T0 2 0 (e j0t e j0t )e j0t dt (1 e j 20t )dt A A j 4j 4 Similarly Χ 1 A A j 4j 4 教育部網路通訊人才培育先導型計畫 Signals and Systems 4.4 45 Example 4-14 (1/4) Please find the exponential Fourier series for the periodic impulse train T (t ) with period T0. 0 T (t ) 0 (t m T ) 0 m T (t ) 0 1 … 3T0 2T0 教育部網路通訊人才培育先導型計畫 T0 0 … T0 2T0 3T0 Signals and Systems 4.4 t 46 Example 4-14 (2/4) 【Sol.】 1) The exponential Fourier series with 0 = 2 /T0. x(t ) Χ ne jn0t , n 0 2 T0 2) Find Xn 1 Xn T0 1 T0 T0 T (t )e jn t dt T0 2 T0 2 0 0 (t )e jn t dt T (t ) (t ) 0 in t herange [ 0 T0 T0 , ] 2 2 1 T0 1 T0 (t ) T0 教育部網路通訊人才培育先導型計畫 jn0t e n Signals and Systems 4.4 47 Example 4-14 (3/4) Double-sided amplitude spectrum of the periodic pulse train T (t ) . 0 Phase is 0, thus the phase spectrum is not shown | X nX|n 1/ T0 教育部網路通訊人才培育先導型計畫 Signals and Systems 4.4 48 Example 4-14 (4/4) e jn0t e jn0t 2 cosn0t Using we obtain 1 T0 (t ) T0 e jn0t n 1 [1 2(cos0t cos 20t ...)] T0 1 [1 2 cos(n0t )] T0 n 1 Single-sided amplitude spectrum of the periodic pulse train T (t ) . 0 教育部網路通訊人才培育先導型計畫 Signals and Systems 4.4 49 Exponential Fourier series of common used signals Name Waveform X0 Square wave 0 Sawtooth wave A 2 A 2 2A Triangular wave Fullrectified sine wave Half-rectified sine wave Rectangular pulse train Pulse train 教育部網路通訊人才培育先導型計畫 A TA T0 A T0 Xn, n 0 Note 2A n A j 2n X n 0, n is even j 2A (n) 2 X n 0, n is even 2A (4n 2 1) A (n 2 1) Tn0 TA sinc T0 2 A T0 X n 0, n is odd except X1 j A A and X 1 j 4 4 Tn0 Tn 2 T0 Signals and Systems 4.4 50 From Fourier series to Fourier transform (1/5) Rewrite the exponential Fourier series of xE(t) xE (t ) j 2nf 0t Χ e n n where 1 Xn T0 1 T0 x(t ), xE (t ) 0, 教育部網路通訊人才培育先導型計畫 T0 / 2 T0 / 2 xE (t )e j 2nf 0t dt x(t )e j 2nf 0t dt | t | T0 / 2 else Signals and Systems 5.1 51 From Fourier series to Fourier transform (2/5) Let’s define a function X ( f ) x(t )e j 2ft dt Then the coefficients Xn can be expressed as 1 X n X (nf0 ) T0 The exponential Fourier series of xE(t) can be written as 1 j 2nf 0t xE (t ) Χ (nf0 )e Χ (nf0 )e j 2nf 0t f 0 n T0 n 教育部網路通訊人才培育先導型計畫 f0 Signals and Systems 5.1 1 T0 52 From Fourier series to Fourier transform (3/5) As T0 , f0 = 1/T0 become infinitesimal ( f0 0), Thus let f0 = f . Then we have x(t ) lim xE (t ) lim T0 f 0 j 2nft Χ ( n f ) e f n The sum on the right-hand side of the above equation can be viewed as the area under the function Χ (nf )e j 2ft f as shown in the figure. Χ ( f )e j 2ft f j 2ft f Area = Χ (nf )e Χ (nf )e j 2nft f nf 教育部網路通訊人才培育先導型計畫 Signals and Systems 5.1 53 From Fourier series to Fourier transform (4/5) Therefore, we have the Fourier representation (Fourier integral) of a nonperiodic signal x(t). x(t ) X ( f )e j 2ft df Similarly, we rewrite the coefficients Xn multiplied by T0 1 Xn T0 In the limit case f = nf0 = n f x(t )e j 2nf 0t dt lim X nT0 x(t )e T0 0 教育部網路通訊人才培育先導型計畫 X nT0 x(t )e j 2nf 0t dt j 2nft dt x(t )e j 2ft dt X ( f ) Signals and Systems 5.1 54 From Fourier series to Fourier transform (5/5) The function X(f ) is called the Fourier transform of x(t), and the Fourier representation defines the inverse Fourier transform of X(f ). Symbolically, they are denoted by X ( f ) F [ x(t )] x(t )e j 2ft dt - x(t ) F [ X ( f )] X ( f )e j 2ft df 1 Keep in mind that the Fourier integral is the nature of a Fourier series with fundamental frequency f0 0. - x(t) and X(f ) are a Fourier transform pair expressed as F x(t ) X( f ) F-1 教育部網路通訊人才培育先導型計畫 or x(t ) X( f ) Signals and Systems 5.1 55 Fourier transform pair (use ) If the angular frequency is used, the Fourier transform pair is expressed as x(t ) Χ () Fourier transform: Χ ( ) x(t )e j t dt Χ ( ) 2 Inverse Fourier transform: x(t ) Χ ( f )e j 2ft df Χ( 1 2 教育部網路通訊人才培育先導型計畫 jt )e d ( ) 2 2 Χ ( )e jt d Signals and Systems 5.1 56 Existence conditions of the Fourier transform Not all signals are Fourier transformable. The existence conditions for the Fourier transform of x(t) are The signal x(t) is absolutely integrable, i.e., | x(t ) | dt . The signal x(t) has a finite number of maxima and minima within any finite interval. The signal x(t) has a finite number of discontinuities within any finite interval, and each of these discontinuities is finite. Though the above conditions guarantee the existence of the Fourier transform for signal. If the impulse function is allowed in the Fourier transform, some signals ( e.g. impulse, unit step, sinusoidal, complex exponential signals) can have Fourier transforms (called generalized Fourier transform that will be discussed later). 教育部網路通訊人才培育先導型計畫 Signals and Systems 5.1 57 Example 5-1 (1/2) Determine the Fourier transform of y(t) = et u( t), > 0 shown in the figure, and plot its spectra. 【Sol.】 1) Compute the Fourier transform Υ ( f ) y (t )e j 2ft dt e t u (t )e j 2ft dt e ( j 2f )t dt 0 1 e ( j 2f )t j 2f 0 1 j 2f 教育部網路通訊人才培育先導型計畫 Signals and Systems 5.3 58 Example 5-1 (2/2) 2) Compute the spectra and plot them shown in the figures. Υ( f ) 1 1 j 2f 2 4 2 f 2 Υ ( f ) ( 1 ) 1 ( j 2f ) 0 t an1 (2f / ) t an1 (2f / ) j 2f 1/ /2 / 2 / 2 / 2 / 2 / 2 Double-sided amplitude spectrum 教育部網路通訊人才培育先導型計畫 Double-sided phase spectrum Signals and Systems 5.3 59 Example 5-2 Determine the Fourier transform of x(t) = ea|t|, a > 0 shown in the figure, and plot its spectra. 【Sol.】Compute the Fourier transform 0 0 Χ ( f ) x(t ) x(t )e j 2ft dt e at e j 2ft dt e at e j 2ft dt 1 1 2a 2 a j 2f a j 2f a (2f ) 2 a2 1.5 1.5 x(t) = Χ( f ) ea|t| 1 Χ( f ) X2(f) x2(t) 1 x(t ) 0.5 0.5 0 0 -0.5 -5 2a a (2f ) 2 2 -4 -3 -2 -1 0 t t 1 2 教育部網路通訊人才培育先導型計畫 3 4 5 -0.5 -5 -4 -3 -2 -1 0 f 1 2 3 4 5 f Signals and Systems 5.3 60 Example 5-3 (1/5) t Determine the Fourier transform of the rectangular pulse signal x (t ) rect ( ) shown in the figure, and plot its spectra. The rectangular pulse signal | t | / 2 1, x(t ) rect( ) 0, t | t | / 2 t x (t ) rect ( ) / 2 教育部網路通訊人才培育先導型計畫 1 0 /2 t Signals and Systems 5.3 61 Example 5-3 (2/5) 【Sol.】 1) Compute the Fourier transform ( use f ) Χ ( f ) x(t ) x(t )e j 2ft dt t rect( )e j 2ft 1 e j 2ft j 2f /2 dt 1e j 2ft dt / 2 /2 / 2 1 (e jf e jf ) j 2f sin(f ) sinc(f ) f sin(x) sinc( x) x 教育部網路通訊人才培育先導型計畫 Signals and Systems 5.3 62 Example 5-3 (3/5) 2) Compute the Fourier transform ( use ) Χ ( ) x(t )e jt dt t rect( )e j t /2 dt 1e jt dt / 2 1 (e j / 2 e j / 2 ) j Sa( 教育部網路通訊人才培育先導型計畫 2 sin( 2 ) 2 ) sin( y ) Sa( y ) y Signals and Systems 5.3 63 Example 5-3 (4/5) Define and plot sinc(.) and Sa( .) sin(x) sinc(x) x sin( y ) Sa( y ) y 教育部網路通訊人才培育先導型計畫 Signals and Systems 5.3 64 Example 5-3 (5/5) Double-sided amplitude spectrum Double-sided phase spectrum 教育部網路通訊人才培育先導型計畫 Signals and Systems 5.3 65 Example 5-4 (1/2) Determine the Fourier transform of the signal xb (t ) rect (t 1 1 ) rect (t ) 2 2 【Sol.】 Compute the Fourier transform 0 Χb( f ) e 1 j 2ft 1 dt e j 2ft dt 0 1 j 2ft 0 j 2ft 1 [e e ] 1 0 j 2f 1 (1 e j 2f e j 2f 1) j 2f 1 [2 2 cos(2f )] j 2f j (1 cos 2f ) f 1 cos 2f sin 2 f j2 f e j 2f e j 2f 2 cos 2f xb (t ) t 2 sin (f ) 2 j 2f sinc2 ( f ) 教育部網路通訊人才培育先導型計畫 Signals and Systems 5.3 66 Example 5-4 (1/2) Double-sided amplitude spectrum Hz Double-sided phase spectrum Hz b ( f ) j 2f sinc2 ( f ) sin( ) sin f 2 2 2 ) j j Sa ( ) If is used b ( ) j ( f 2 ( ) 2 2 教育部網路通訊人才培育先導型計畫 Signals and Systems 5.3 67 Example 5-5 Determine the Fourier transform of the signal x(t ) 1, t . 【Sol.】 Compute the Fourier transform Χ ( f ) 1e j 2ft dt lim T /2 1 e j 2ft T j 2f T /2 1e j 2ft dt lim T T / 2 T / 2 1 1 jfT jfT lim [e e ] lim sin(fT ) T j 2f T f The Fourier transform of x(t) = 1 does not exist. (The signal x(t) is not absolutely integrable) A generalized Fourier transform is introduced next to address this situation. 教育部網路通訊人才培育先導型計畫 Signals and Systems 5.3 68