308 FOURIER ANALYSIS OF DISCRETE-TIME SIGNALS AND SYSTEMS (CHAP. 6 Solved Problems DISCRETE FOURIER SERIES 61. We call a set of sequences {¥,[n}) orthogonal on an interval [N,, N] if any two signals ¥,,(m] and ¥,[7} in the set satisfy the condition mek mek (6.114) ¥ vatmpvsin| = (2 where * denotes the complex conjugate and @ #0, Show that the set of complex exponential seque’ [A] = ke O 1 NOI (6.115) is orthogonal on any interval of length N. From Eq, (7.90) we note that wes aal z 6.116 a ae fone) Applying Ea. (6.176). with a= ewe obtain is NENt2N, grain | y= eatery ent ¥, | qo mean -° otherwise (enn) since ¢ re Since each of the complex exponentials in the summation in Eq, (6.1177 is periodic with period NV, Eg, (6.117) remains valid with a summation carried over any interval of length N. That is, iB k=0 \ 10 otherwise 6.118) Now, using Ei, (6.178), we have Eo vgn|Weln) = Semen sn gemaes sin = Seem tensn {Noms (e109 0 me where m,k-<.V. Equation (6.179) shows that the set (647%: k= 1) is orthog. onal over any interval of length N. Equation (6.114) is the discrete-time counterpart of Eq, (5.95) introduced in Prob. 5.1CHAP. 6] FOURIER ANALYSIS OF DISCRETE-TIME SIGNALS AND SYSTEMS 309 6.2. 63. Using the orthogonality condition Eq, (6.119), derive Eq. (6.8) for the Fourier coefficients Replacing the summation variable & by m in Eq. (6.7), we have alm] =X cgemer/von (6.120) Using Eq. (6.115) with N= Ng, Eq, (6.120) can be rewritten as *S cotaln] (6.21) afr] Multiplying both sides of Eg. (6.121) by p(n] and summing over n = 0 to (Ny ~ 1), we obtain Yalelet= ED ota te) Imterchanging the order of the summation and using Eq. (6.119), we get ¥ sinlvela) ¥ col E valnlven]] (6.122) ‘Thus, = E slolvete) i E alae sort Determine the Fourier coefficients for the periodic sequence x{n] shown in Fig. 6-7. From Fig. 6-7 we see that x{] is the periodic extension of (0,1,2,3) with fundamental period Ny = 4. Thus, a= 3 and By Eq. (6.8) the discrete-time Fourier coefficients cg are 12 1 cong Dalen) = GO 14243) ee) Sats pnn- 243 ati noe 27/3 Ms 2nd a a Ely i 13 sad 1 ong Lali 5 --t Note that =e, = eF (Eq. (6.179)310 FOURIER ANALYSIS OF DISCRETE-TIME SIGNALS AND SYSTEMS (CHAP, 6 Fig. 67 64. Consider the periodic sequence rl] shown in Fig. 6-8(4). Determine the Fourier coefficients c, and sketch the magnitude spectrum ley]. From Fig, 6-8lu) we see that the fundamental period of xf] 1s Ny= 10 and 2, = 3/5. By Eq, (6.8) and using Eg, (/.99), we get ay ad gman SREY gy 9 10 Sing kx 10) The magnitude spectrum Jeg! is plotted in Fig. 6-815}, an)CHAP. 6] FOURIER ANALYSIS OF DISCRETE-TIME SIGNALS AND SYSTEMS 31 65. Consider a sequence x(nl= n— ak] fa) Sketch x{n]. (b) Find the Fourier coefficients ¢, of x{n] (a) The sequence x[n] is sketched in Fig. 6-%a). It is seen that x[/1] is the periodic extension Of the sequence {1,0,0,0) with period Ng 4 o128 © w Fig. 6-9 (®) From Eqs. (6.7) and (6.8) and Fig. 6.%a) we have t 0) 1 Hk (+> 4 ls -ne/05 and a= Z Eabe since xf1] [2]=x[3] =, The Fourier coefficients of xin] are sketched in Fig. 6-96). 6.6. Determine the discrete Fourier series representation for each of the following se- quences: 4 fa) xln]= cos sn 16.7. lw ) FOURIER ANALYSIS OF DISCRETE-TIME SIGNALS AND SYSTEMS (CHAP. 6 ‘The fundamental period of x{n] is Ny =8, and Q) = 27/N,= 7/4. Rather than using Eq. (6.8) to evaluate the Fourier coefficients ,, we use Euler’s formula and get, F ed cenn, wer rainy =} giton cos pn = 5(e Thus, the Fourier coefficients for ala} ate c= $, €)=¢-se¢2¢>= fy and all other 6, = Hence, the discrete Fourier series of al) is rola d xf) = cos ny 2, From Prob, [.16(4) the fundamental period of xin] is Ny = 24, and My Again by Euler's formula we have 1 ! AL] = 507 5 eH 4 Cee gaan ty a fe niton gettin — eran g Lefatin Thus, ¢ . =i, U4) and all other the discrete Fourier series of x{m] is From Prob. 1.16()) the fundamental period of [1] is Ni, =8, and Again by Euler's formula we have alm} = (erm 4 itm 6 5 a Leo aetn Henee, the discrete Thus, cy = Fourier series of xf) is aahe =e and all other ¢, an! an 4 gia a, 274 a Let _x{n] be a real periodic sequence with fundamental period Ny and Fourier coefficients c, = a, + jb, where a, and 6, are both real. (a) i) fo) Show that a, —a, and b . = —b, Show that cy, is real if Ny is even. how that x[n] can also be expressed as a discrete trigonometric Fourier series of the form oy 92 x(n] =cy+2 LD (a, coskQgn ~ 6, sin kQon) a, (6.123) if Ny is odd or + fey 2 t2 EK (ay cos kya ~b, xf] nkQyn) (6.124) if Ny is even.CHAP. 6] FOURIER ANALYSIS OF DISCRETE-TIME SIGNALS AND SYSTEMS 313 (a) If xln]is real, then from Eg. (6.8) we have phe LM E afnpeitton E afnjeer ont No neo Thus ge tb y= (a HIB) =, — IP and we have rr (6) IL Ny is even, then from Eq. (6.8) phd aoe xf] enh /20e 7800 1)" x[a] = real (6.125) (©) Rewrite Eq, (6.7) as anda Eee nege Eger ran mo If Ny is odd, then (Ny — 1) is even and we can write xlr] as Noma stn meat Ee (eqertter seg, get en) Now, from Eg. (6.17) and eM kia eo thlton Thus, omens ata] =eot = (cgettor + efeittony cut? YL ReCay +Jb,)(cosk Ayn +Jsin kgm) (4, 608 KMynt ~ By sin kM) IL Ny is even, we can write x[n] as D7 meet E (eset eg get 0M) ey eatin314 FOURIER ANALYSIS OF DISCRETE-TIME SIGNALS AND SYSTEMS (CHAP. 6 Again trom Eq. (6.17) ea set and ef Ite gone and ethan wm elmo)" then afm]=ep 4 (Meat EE 2Re(eyer") nav E (acon kn by sin kB) 68, Let xn] and x,[n] be periodic sequences with fundamental period discrete Fourier series given by jy and their xinl= X det" oxfn}= Leer y= mo oy Qe No Show that the sequence x[n]=.x,(n]x,{n] is periodic with the same fundamental period Ny and can be expressed a x[a]= See aye = where c, is given by XL deem (6.126) Now note that alm = Ny] =xyfe + Nyeo[e + Ny] = bebe fa] = at] Thus, x{n1 is periodic with fundamental period Ny. Let s[n]= E gente Qy= Then one Dane £3 ‘ ' 5 dein ste since and the term in parentheses is equal to e,_,.CHAP. 6) FOURIER ANALYSIS OF DISCRETE-TIME SIGNALS AND SYSTEMS 318 69. Let x,[n] and x,[n] be the two periodic signals in Prob. 6.8. Show that 1 Eoainin Ese. wen Equation (6,/27) is known as Parseval’s relation for periodic sequences. From Fa, (6.126) we have E alrletedese= “Steen 1 ONG mao Setting k= in the above expression, we get EE atninde E dyews Dodie No nce mao io 6.10. (a) Verify Parseval’s identity [Eq. (6.19)] for the discrete Fourier series, that is, beet a wy Biel Ea (b) Using x[n] in Prob. 6.3, verify Parseval’s identity [Eq. (6.79)], (a) Let alnl~ 5 een and a[nJ= E dyer ‘Then d,- Ki = xt[n] ewistton = e iL atnet (6.128) Equation (6.128) indicates that if the Fourier coefficients of xl] are cy, then the Fourier coetficients of an} are c* . Setting xml =axfn] and x{n] =.x°(n] in Eg, (6.127), we have d, =e, and e, ~ 7, (or eg =f) and we obtain wy, Laleletd= YE cst (6.129) ls ano es er or wy 2 tind= E teak (6) From Fig, 6-7 and the results from Prob. 6.3, we have Nyt edger er? = Oe Peo and Parseval’s identity is verified,316 FOURIER ANALYSIS OF DISCRETE-TIME SIGNALS AND SYSTEMS (CHAP. 6 FOURIER TRANSFORM 6.11. Find the Fourier transform of x[n} = -atuj-n— 1] a real From Eq. (4.12) the z-transform of x(t] is given by i Xx = Pe ae Thus, Xe’) exists for |al> 1 because the ROC of X(z) then contains the unit circle. Thus. 1 T~ae X(Q)=X(e@)= yal (6.130) 6.12. Find the Fourier transform of the rectangular pulse sequence (Fig. 6-10) afm] =uln] —u[n—N] jing Eq, (J.90), the z-transform of x{n] is given by x)= L \zi>0 (6.131) Thus, Xe) exists because the ROC of X(z) includes the unit circle, Hence, wuny cae 0 grt 4 gfN/2_ gly (9) = Xe) = ie = sia 7y soe ay2 SM ON/2) 1 . sin( 21/2) (6.132) Fig. 6-10 6.13. Verify the time-shifting property (6.43), that is, x[m— ny] re 8 X(Q) By definition (6.27) rw Flxln— no] gleCHAP. 6] FOURIER ANALYSIS OF DISCRETE-TIME SIGNALS AND SYSTEMS By the change of variable m =n ~ ng, we obtain Flxln— nj = Lo afmpeserren ee SL xfm]e = eX) Hence, “x(Q) afnon] oer 6.14. (a) Find the Fourier transform X(0) of the rectangulur pulse sequence shown in Fig. 6-11(a). LaIeL _L @ wo Fig. 6-1 (b) Plot X(Q) for N,=4 and N, = 8, (a) From Fig. 6-11 we see chat a[n}=x,[n +N] where xn] is shown in Fig. 6-11(b). Setting y+ in Bg, (6.132), we have sin O(N, + 9) sin( 172) x(n) Now, from the time-shifting property (6.43) we obtain sin 00% + )] + 33 sin 0/2) (O18) ON XO x(a () Setting N, = 4 in Eq, (6.133), we get sin(4.5Q) X(9) = asm)318 FOURIER ANALYSIS OF DISCRETE-TIME SIGNALS AND SYSTEMS [CHAP. 6 which is plotted in Fig. 6-12(a). Similarly, for N, = 8 we get sin(8.50) Sin(0.50) which is plotted in Fig. 6.12(b). amy 7 7 7a w » Fig, 612 6.15. (a) Find the inverse Fourier transform x{] of the rectangular pulse spectrum X(Q) defined by (Fig. 6-13(a)] fh {O\ — (ae Verify the multiplication property (6.59), that is, 1 silella] > 7290) 8 X(0) Let xl] = x,Ln}x{n} Then by definition (6.27) x(M)= EY x laletnder By Fg, (6.28) x(a) ei da xn]324 6.27, 6.28. FOURIER ANALYSIS OF DISCRETE-TIME SIGNALS AND SYSTEMS [CHAP. 6 Then xMa= 5 al (oye *as)sinte ne Interchanging the order of summation and integration, we get x(a) =f x40 | Elle] as L ; 1 JMNM 8) dd 5X0) @¥() Hence, alm besln + ge X( 9) ex(a) Verify the properties (6.62), (6.634), and (6.636); that is, if x[n] is real and alm} xfer] tf] (0) = A(2) + jB(0) (6.140) where x,{t] and x,{n] are the even and odd components of x(n], respectively, then X(-9)=¥*(0) x,[a] Re{ X(M)) = 412) x [a] rj Im{X(0)} = jB(2) If x(n] is real, then x*[1] = xLn], and by Eg. (6.45) we have xt[n] ox 0) from which we get X(Q)=X(-0) or -X(-M) = (9) Next, using Eq. (6.46) and Eqs. (/.2) and (1.3), we have aL on} od] xe] 2 x(- 9) = (0) = 4() ~78(9) (6.141) Adding (subtracting) Eq. (6 /41) to (from) Eq, (6.140), we obtain x,[n] —A(2) = Re X(0)} eof] JB) = j Im NCO} Show that \Ql X(0) Now, note that an) ~ ulm) — fn 1) ‘Taking the Fourier transform of both sides of the above expression and by Eqs. (6.36) and (6.43), we have = (se) X(0)CHAP. 6] FOURIER ANALYSIS OF DISCRETE-TIME SIGNALS AND SYSTEMS 3 6.29. 6.30, Noting that (1) = 0 for = 0, X(O) must be of the form 1 X() -49(2) + lOlsz Where A is a constant. To determine A we proceed as follows. From Eq. (1.5) the even component af ula is given by un] Then the odd component of u(x} is given by uf] ln] ~u,{n] =e] ~ $= tol] and From Eq. (6.636) the Fourier transform of an odd real sequence must be purely imaginacy. ‘Thus, we must have A=, and uln}es Mer Verify the accumulation property (6.57), that is, Yelk] 7X (0)5(0) + XID) isa From Eq. (2.132) Saal =a{a}euln] Thus, by the convolution theorem (6.58) and Eq. (6/42) we get Eafe} aay] aca) + i | IDls= = a X(0) EN) + Sg X(M) since X(2) (0) = X(0) ACM) by Eq, (1.25). Using the accumulation property (6.57) and Eq. (1.50), find the Fourier transform of u(r] From Eq. (1.50) uln}= Sate] Now, [rom Eg. (6.36) we have {nm} Setting xLk] = lk] in Eg. (6. a[n]=4[n]+x(2) ), we have and X(0)= 1326 FOURIER ANALYSIS OF DISCRETE-TIME SIGNALS AND SYSTEMS [CHAP. 6 and FREQUENCY RESPONSE 6.31, A causal discrete-time LTT system is described by n-2=s[n yin} ivf) + (6.143) where x{n] and y[n] are the input and output of the system, respectively (Prob. 4.32). (a) Determine the frequency response H() of the system. (b) Find the impulse response A{r] of the system, (a) ‘Taking the Fourier transform of Eq. (6.143), we obtain ¥(2) — eM YLD) + LePPYED) = XO) or Pay) = X(Q) “Thus, H(Q) = x(2) (b) Using partial-fraction expansions, we have H(Q) = Taking the inverse Fourier transform of H(Q), we obtain alm) = [20)"- Gy" abn) which is the same result obtained in Prob. 4.32(6). 6.32, Consider a discrete-time LTI system described by tft xin] + y (6.144) (a) Determine the frequency response H() of the system. (b) Find the impulse response hf] of the system (c) Determine its response y[11] to the input x[n] = cos (a) Taking the Fourier transform of Eq, (6.144), we obtain ¥(M) ~ be MYM) = XM) + fe XD)CHAP. 6) FOURIER ANALYSIS OF DISCRETE-TIME SIGNALS AND SYSTEMS 307 Thus, AD : «an 3m) th) HQ) = 5 (©) From Bg. (6.137) =79(0-TJe rm 1% 4 n4(04 Te Taking the inverse Fourier transform of ¥(G) and using Eq. (6.135), we get 1 gn] = sete I arn tty Lae = cos{ Zn — 2tan 6.33. Consider a discrete-time LTT system with impulse response /4) atm = Find the output yl] if the input xl] is a periodic sequence with fundamental period No = 5 as shown in Fig, 6-17. From Eg, (6.134) we have 1d the filter passes only frequencies in the range [0 < =/4, only the de term is passed through, From Fig. 6-17 and Eq. (6.17) “ 3 Lala]328 FOURIER ANALYSIS OF DISCRETE-TIME SIGNALS AND SYSTEMS (CHAP. 6 Thus, the output y[r] is given by yin]=2 alle 6.34, Consider the discrete-time LTI system shown in Fig. 6-18. (a) (by (co) (d) @ () ca) Find the frequency response H() of the system, Find the impulse response A[/] of the system Sketch the magnitude response |H(0)| and the phase response 4(). ind the 3-dB bandwidth of the system. From Fig, 6-18 we have vbr] oxtn) extn 1] (6.145) Taking the Fourier transform of Ea, (6.145) and by Fq. (6.77), we have y(n) x) (2) Ite IAG @ 3/2 4g Ay ay #7 co5| lols (6.146) By the definition of Al] [Eq. (2.309) and Ea, (6.145) we obtain Alm] =6[n] + 8{n— 1] Osnsl on Mt) incre From Eq. (6.146) 1091-20CHAP, 6] FOURIER ANALYSIS OF DISCRETE-TIME SIGNALS AND SYSTEMS 329 nip and (9) = which are sketched in Fig, 6-19. waa a 7 a = . Fig. 619 (a) Let 5, be the 3-dB bandwidth of the system. Then by definition (See. 5.7) 1 |H( 2566) = SHH D)loan we obtain Pose) and 5 a We see that the system is a discrete-time wideband low-pass finite impulse response (FIR) filter (See. 2.90). where a is a constant and 6.35. Consider the discrete-time LI system shown in Fig. 6-21 O) 7 rr "nt re 0 Fig. 6-22. Transformation of a low-pass filter to a high-pass filter 6.37. Show that if a discrete-time low-pass filter is described by the difference equation vin] =~ Sagvln-k] + Spel 4] (6.153) then the discrete-time filter described by x a y[n}=— X (-tfayfn kp + D (1 dypxla ak] (6.154) oy mo is a high-pass filter. Taking the Fourier transform of Eq. (6,753), we obtain the frequency response H,pp412) of the low-pass filter as App ( (oy (6.155) rel) Fey (6.155) If we replace © by (O1~ m) in Eq, (6.755), then we have ea “ Sbjenn S o(-nteae Hype (Q) = Hype Q— 2) =~ Se (6.156) Faye re S (tye which corresponds to the difference equation LC vtayla-k] + Lo (- testa kd vl332 FOURIER ANALYSIS OF DISCRETE-TIME SIGNALS AND SYSTEMS [CHAP. 6 6.38. Convert the discrete-time low-pass filter shown in Fig. 6-18 (Prob. 6.34) to a high-pass filter. From Prob. 6.34 the discrete-time low-pass filter shown in Fig, 6-18 is described by [Eg, (6145)] yin] a(n] sain 1) Using Eg, (6.1541, the converted high-pass filter is described by vin] =xtnj—xfn— 1] (6.157) which leads to the circuit diagram in Fig. 623. Taking the Fourier transform of Eq. (6.157) and by Eq. (6.77), we have H(Q) <1 2 = 7208? sin sin (Ala (6.158) From Eq, (6.158) JH()|=2'sin| \Q\er [or=0y/ ver us 8) oe ay neo which are sketched in Fig. 6-24. We see that the system is a diserete-time high-pass FIR filter Fig. 6.23 6.39, The system function H(z) of a causal discrete-time LITT system is given by (6.159) where a is real and |a\< 1, Find the value of b so that the frequency response H(Q) of the system satisfies the condition H(Q)\=1 all (6.160) Such a system is called an all-pass filter By Eq. (6.34) the frequency response of the system is pis 7 (6.160) HD) = HDL =NALS AND SYSTEMS 333 CHAP. 6] FOURIER ANALYSIS OF DISCRETE-TIME SI vain)! ‘Then, by Eq. (6.160) H(M) which leads to Ib+e + eos 2 ~ jsin 2) = [1 — acos 0+ ja sin 9) 4a? ~ 2acos (6.162) or 146? + 2bcos 0 and we see that if b ~ ~a, Eq. (6.162) holds for all © and Eq. (6.160) is satisfied. 6.40. Let Aln] be the impulse response of an FIR filter so that Aln]=0 nn <0,n2N Assume that Afn] is real and let the frequency response H(12) be expressed as H(2) =|H(M)le*™ (2) when Alm] satisfies the condition [Fig. 6-25(a)] hn] = h[N~1—n] (6.163) (a) Find the phase respons (b) Find the phase response 6(2) when h{n]} satisfies the condition [Fig. 6-25(4)] Aln] = -A[N- 1-1) (6.164)FOURIER ANALYSIS OF DISCRETE-TIME SIGNALS AND SYSTEMS [CHAP. 6 (a) ) | wat ven «) w Fig. 625 aking the Fourier transform of Eq. (6.163) and using Eqs. (6.43), (6.46), and (6.62), we obtain H(A) = Dy eH or LAC Dye — FC AYIe MRE HN OH Thus (2) = #2) ~ (N=) and a ny= =H N-1)0 (6.165) which indicates that the phase response is linear Similarly, taking the Fourier transform of Eg. (6.164), we get H(M) = -H*( Ay eat or VHC Vie) = [HEALEY eH OF Thus, HD) = 2 ~ AM) — (N= 1) and 99) == —(v-1)0 (6.166) which indicates that the phase response is also linear.CHAP. 6] FOURIER ANALYSIS OF DISCRETE-TIME SIGNALS AND SYSTEMS 335 6.41. Consider a three-point moving-average discrete-time filter described by the differenct jon yln] = ala] ta[n— 1] 4x[n—2]) (6.167) (a) Find and sketch the impulse response /i{n] of the filter, (b) Find the frequency response H({2) of the filter. (c) Sketch the magnitude response |H({2)| and the phase response 6((2) of the filter. (a) By the definition of fil) [Eq, (2.30)] we have (0[m] + afm ~ 1} +aln—2)) (6.168) * Hed cree is sketched in Fig. 6-264), Note that A(n] satisfies the condition (6 163) with (b) wy) @ wo Fig, 6-26336 FOURIER ANALYSIS OF DISCRETE-TIME SIGNALS AND SYSTEMS [CHAP. 6 By Eq. (1.90), with a =e", we get gy =e be )-54 5 SWAT —n 8in30/2) 1 sin(302/2 3 sin(/2)_ Hi Me™ (6.169) where HAM) = (6.170) fe) From Eg, (6.169) sin( 30/2) | 1H(Q))=[H,(Q)\= 4 aay-{ 72 when H,(02) >0 an (9) ase when H,(12) <0 which are sketched in Fig, 6-26(b). We see that the system is a low-pass FIR filter with linear phase, 6.42. Consider a causal discrete-time FIR filter described by the impulse response ln] = (2,2,-2,- 2} (a) Sketch the impulse response A{n] of the filter. (b) Find the frequency response H(Q) of the fi (©) Sketch the magnitude response |H()| and the phase response 0({2) of the filter. (a) The impulse response ln] is sketched in Fig 6-27a). Note that f{n] satisfies the condition (6.164) with N= 4 (b) By definition (6.27) H(Q)= YS Almere =2(1 eM) +e ade PA/2(giM/2 gL) 4 JQ 72 (gIM/2 gH) 2 30 je | (6.171) where (©) From Eq. (6.17) LH(Q)|= 11, 2)1= sin H(M)>0 a) which are sketched in Fig. 6-27(h). We see that the system is a bandpass FIR filter with linear phaseCHAP. 6] FOURIER ANALYSIS OF DISCRETE-TIME SIGNALS AND SYSTEMS 337 me’ Fig. 6-27 SIMULATION 6.43. Consider the RC low-pass filter shown in Fig. 6-28(a) with RC (a) Construct a discrete-time filter such that hal] = hh nr i (nT,) (6.172) where h,(¢) is the impulse response of the RC filter, A,{n] is the impulse response of the discrete-time filter, and 7, is a positive number to be chosen as. part of the design procedures. (b) Plot the magnitude response |17,(w)|of the RC filter and the magnitude response |H,(aT,)| of the discrete-time filter for T, = 1 and T, = 0.1 (a) The system function H,(s) of the RC filter is given by (Prob. 3.23) 1 HAs) = (6.173) sel and the impulse response 4,(0) is At) etu(t) (6.174) By Eq. (6.172) the corresponding h,{r] is given by Adn)=e-"uln] =(e %)"uln) (6.175)338, ) FOURIER ANALYSIS OF DISCRETE-TIME SIGNALS AND SYSTEMS (CHAP. 6 a) sn) ) Fig, 6-28 Simulation of an RC filter by the impulse invariance method. Then, taking the z-transform of Eq. (6.175), the system function H(z) of the discrete: time filter is given by from which we obtain the difference equation describing the discrete-time filter as vin] yl = 1]=2[1] (6.176) from which the discret By Eg. (5.40) ime filler that simulates the RC filler is shown in Fig. 6-28(6) Hw) = H)\s-i0= Then Hw) Hier By Eqs. (6.34) and (6.81) HoT.) = H,(2))s0 From Eg. (6.149) H(oT,)| a [eet 26 Meos(wi,)]”CHAP. 6] FOURIER ANALYSIS OF DISCRETE-TIME SIGNALS AND SYSTEMS 39 From 7,=1, Hy) For T,= 0.1, 1 Igwt))= a MoT) = 8a 36M tw] The magnitude response |F/(w) ofthe RC filter and the magnitude response [1/(w,) of the discrete-time filter for T,=1 and T, = 0.1 are plotted in Fig. 6-29. Note that the plots ae sealed such thatthe magnitudes at w =O are normalized 0 1 “he method utilized in this problem to construct a dvcetetimespntem to simulate the contiuousctime system i Kowa the ipubesneenance method Wig Haat) 6.44. By applying the impulse-invariance method, determine the frequency response H,( 2) of the discrete-time system to simulate the continuous-time LTT system with the system function 1 “GFDGr2) H(s) Using the partial-fraction expansion, we have HAS) = a 54d Thus, by Table 341 the impulse response of the continuous-time system is h(t) = (ef =e *)u(r) (6.177) Let hin] be the impulse response of the discrete-time system, Then, by Eg. (6.177) hla] =h (nT) = (0% = eT ulm]340 6.48, FOURIER ANALYSIS OF DISCRETE-TIME SIGNALS AND SYSTEMS (CHAP. 6 and the system function of the discrete-time system is given by Hz) (6.178) (6.179) (6.180) then the impulse-invariance method yields the corresponding discrete-time system with the system function 17,(2) given by Hdzy= (6.187) A differentiator is a continuous-time LTE system with the system function [Eq. (3.20)] H,(s)=s (6.182) A discrete-time LTY system is constructed by replacing s in H,(s) by the following transformation known as the bilinear transformation: 24-2! “Ris (6.183) tor, Again T, in Faq. (6.183) is a positive number to be fo simulate the differen chosen as part of the design procedure. (a) Draw a diagram for the discrete-time system (b) Find the frequency response H,() of the discrete-time system and plot its magnitude and phase responses. (a) Let H,(2) be the system function of the discrete-time system. Then, from Fas. (6.182) and (6.183) we have HAz)= (6.188) Writing Hz) as ap , 42) 5 (qri|-e 9 then, from Probs. (6.35) and (6.38) the discrete-time system can be constructed as & scade connection of two systems as shown in Fig, 6-304). From Fig. 6-30(a) it is seen hat we can teplace two unitdelay elements by one unitdelay element as shown in Fig 306).CHAP. 6] FOURIER ANALYSIS OF DISCRETE-TIME SIGNALS AND SYSTEMS 341 wo 6-30 Simulation of a differentiator (6) By Eq, (6.184) the frequency response H,(1) of the discrete-time system is given by 21- 2 HAM) =F mF 2 sinfn/2 (6.185) Note that when © 1, we have HAD) (6.186) if 2 wT, (Fig. 6-31). 631342 6.46, FOURIER ANAL OF DISCRETE-TIME SIGNALS AND SYSTEMS [CHAP. 6 Consider designing a discrete-time LTI system with system function H,(z) obtained by applying the bilinear transformation to a continuous-time LTT system with rational system function H,(s), That is, H(z) =HAs\len aes wa (6.187) Show that a stable, causal continuous-time system will always lead to a stable, causal discrete-time system. Consider the bilinear transformation of Eq (6 183) t : (6.188) Sofsine Bq, (6188) for 2. we obtain Lens : (0.189) 1=(7/2s Setting 5 je in Eg, (6.189), we get [| so) mr | (at ‘Thus, we see that the jw-axis of the s-plane is transformed into the unit circle of the z-plane. Let care" and so jw Phen from Eq. (6.188) 2rsin Per?+2reof '/1 =r? + Drees roa (6.1910) dren ‘ sin nt 6.191b Treas ‘ » From Eq, (6.1914) we see that if r <1, then <0, and if > 1, then o> 0, Consequently, the left-hand plane (LHP) ins maps into the inside of the unit circle in the z-plane. and the right-hand plane (RHP) in s maps into the outside of the unit circle (Fig, 6-32). Thus, we conclude that a stable, causal continuous-time system will lead t0 a stable, causal discrete-time system with a bilinear transformation (see Sec. 3.68 and Sec. 4.68), When r= 1, then o = 0 and a 2 sin 2 Tirana” Tt tan (6.192)CHAP. 6] FOURIER ANALYSIS OF DISCRETE-TIME SIGNALS AND SYSTEMS 343 ‘plane plane hci ; Fig. 6:32. Bilinear transformation, n=2 1% 6.193. or tan 5 (6.193) From Eq. (6.193) we see that the entire range —» a] 1 =a ,oo- Talal” x3) Poy at j | [3 . Let x{n] be a sequence of finite length N’ such that xin] =0 n| =F[k] ~ WeGlk) k (6.217) wwe N where Flk]}= Lf {n| wer and (6.218a) ano 2 21 N Glk]= Lo gina, azo) (6.218) no 2 (c) Draw a flow graph to illustrate the evaluation of X[k] from Eqs. (6.2174) and (6.2170) with N= 8. (d) Assume that x(n] is complex and Wg* have been precomputed. Determine the numbers of complex multiplications required to evaluate X(k] from Eq, (6.214) and from Eqs. (6.2174) and (6.217) and compare the results for N = 2! (a) From Eq. (6.213) py fla]=sL2n]=0,n <0 and [S| =s[N]=0 Thus sln]=0 Similarly glnl=af2n +1) =0.n<0 ‘Thus, eln]=0 (b) We rewrite Bq. (6.214) as x{k]= ¥ x[n]ngrs ¥ alge -"S “stamens 9 38 ‘em ser (6.209) But CHET we CMO IN) = RIN = Wy (6.220) With this substitution Eq, (6.2/9) can be expressed as xceg="E “sles = “Yelm SPUR] +WAG[R] k= O1 NI (6.221)«© a) FOURIER ANALYSIS OF DISCRETE-TIME SIGNALS AND SYSTEMS (CHAP. 6 where ree ‘pean, wy G{k]= Salm peer, HOS Note that F[K] and G{k] are the (N/2+point DFTs of fl] and gin), respectively. Now WEN = WEWe? = WE (6.222) since WE Ce HEM MIN gow (6.223) y X[k) = F(R] + WECK] keOb Zod fee Sfmar-mgetsy bento The flow graph illustrating the steps involved in determining X[& J by Bas. (6.2/7a) and (6.2178) is Shown in Fig, 637 To evaluate a value of AK] from Eq. (6.2/4) requires N complex muluplications. Thus, the total number of complex multiplications based on Eq. (6.2/4) is N?. The number of complex multiplications in evaluating [AJ or G{x) is (N/2P. In addition there are N multiplications involved in the evaluation of WSGL&]. Thus, the total number of complex multiplications based on Eqs. (6.2174) and (6.217) is 0N/2)4N = N2/2¢N. For N= 2!~ 1024 the total number of complex multiplications based on Eq. (6.214) is 0 and is 10/2 + 1924 = 10"/2 based on Eqs. (6.2172) and (6.277), So we see that the number of multiplications is reduced approximately by a factor of 2 based on Eqs. (6.2174) and (6.2176). ‘The method of evaluating X[k} based on Eqs. (6.272) and (6.217b) is known as the decimation-in-time Jast Fourier transform (FFT) algorithm. Note that since N/2 is even, using the same procedure, Fik) and G{&] can be found by first determining the (N/4)-point DFTs of appropriately chosen sequences and combining them Fig. 6-37 Flow graph for an 8:point decimation in-time FFT algorithm.cHaP. 6] FOURIER ANALYSIS OF DISCRETE-TIME SIGNALS AND SYSTEMS. 353 6.57. Consider a sequence alm} = (IL-1. 111-1) Determine the DFT X[k] of x[n] using the decimation-in-time FFT algorithm, From Figs. 6-38(a) and (4), the phase factors W! and W# are easily found as follows: wWe=; and weal Wie 1 ijt = eB Next, from Eqs. (6.215a) and (6.2156) Fln]=x(2] = (xf0}, [2], 214), xf6]) = (1-1. 1.1) alm] =a(2n +1) = (sft) 013], (5). 017) = (= 11-1) ‘Then, using Eqs. (6.206) and (6.222), we have F(0] 1 1 1 ft ° FU) 1 j 1 ifr} fase Fai}"| 1 - 1 -ajf-1 o | F[3) 1 i 1 i 1 2-32] Gro] 1 1 1 poy oul} jor Fm i[-1)_j0 G2] 1-1 1 iff [ola 63} 1 ioc ilf-al [o Ios Rete Rew «@ Fig. 638 Phase factors ¥" and Hf354 FOURIER ANALYSIS OF DISCRETE-TIME SIGNALS AND SYSTEMS [CHAP. 6 and by Eys. (6.2172) and (6.2176) we obtain [0] = F[0] + #EGT0] = 0 X[A) = (0) - WEGTO} ~ 0 XQ] FU] + WE] =2472 X15] = FE] wg ett]=2 442 X[2]=F[2] + #212] = 14 [6] = F[2) ~ W3G[2] =4 XD] =F) + ZED] =2- 72 X[7] = £3] -wecis] = 2-;2 Noting that since xf] is real and using Eq. (6.204), X{7], X16}, and X15) can be easily obtained by taking the conjugates of X{1}, X(2h and X{3], respectively 6.58. Let x[z1] be a sequence of finite length N’ such that x(nj=0 0 n<0,n2N Let the N-point DET X{k] of x(] be given by [Eq, (6.92)] wo X{k]= Eoxfnfge Wye k aa Suppose NV is even and let N N pli] =x[n] 0] Osne> (6.2252) N N alm] = [s(n] aint SJ Ose 5 (6.2256) (a) Show that the N-point DET X{k] of xf] can be expressed as, N X2K)=P[R] k=O ta pat (6.226a) N XQ2E+=O[k] k=O Fat (6.2266) wart N where Plkl= DL pln| Wars Mong oh (6.2278) wey N Qlk]= EL ainjwers Ot z 1 (6,227) aso (b) Draw a flow graph to illustrate the evaluation of X[k] from Eqs. (6.226a) and (6.2266) with N = 8. (a) We rewrite Eq, (6.224) as Xk)- Yates YL xlnyge (6.228) Changing the variable m =m +N/2 in the second term of Eq, (6.228), we have oyna wet X(k]= Dafne eye De sm (6.229) 7CHAP. 6) () FOURIER ANALYSIS OF DISCRETE-TIME SIGNALS AND SYSTEMS 355 Noting that [Eq (6.223)] Eq, (6.229) can be expressed as atk]= lens ne (6.230) For & even, setting k = 2r in Eg, (6.230), we have wart wt 7 XPr= Lo plapyer= > in Eq, (6.230), we get warn wart N XPrei= Lo alm Co alan, $01 FT (6.232) Equations (6.231) and (6,232) represent the (N/2+point DFT of pla and gin, respec: tively. Thus, Figs. (6.232) and (6.232) can be rewritten as y X[2k) = P(A] scl N XQK+N-Olk)— k=OL St wt N where Ptel= plein, scl wart y k= Lo alalwyry ke OM Sad ‘The flow graph illustrating the steps involved in determining X[&] by Fas. (6.227a) and (6.2276) is shown in Fig. 6-39, The method of evaluating X[k] based on Eqs. (6 227a) and (6.2276) is known as the decimation-in-frequency fast Fourier transform (FFT) algorithm, 10) xo i xn 121 me x0) XIol 4a) xu as) xBI x6) IS] a xP| Fig. 6.39 Flow graph for an 8-point decimation-in-frequency FFT algorithm.356 FOURIER ANALYSIS OF DISCRETE-TIME SIGNALS AND SYSTEMS [CHAP. 6 6.59. Using the decimation-in-frequency FFT technique, redo Prob. 6.57, From Prob. 6.57 ala] = (LL a By Faqs, (6.225a) and (6.2256) and using the values of W'7 obtained in Prob. 6.57, we have N pln] = x(n] tx] 5 ={(1 1.4 DEED C1 1) = (0,2,0,2) N alr} (at =a[n= 4 ])ve ={ (Ls Dee ye 2,0, j2.0) ‘Then using Eqs. (6.206) and (6.212), we have P(o}} fa 1 1 yf o] fo Pt] toy ifo2, | =i Pl2] 1 -1 1 -1lf ol7]o Py) fot joc! ijl-2) [ua a} fa 1 1 ipo) fra ott] top pifo|_f2-y2 ea] 1 1 1-1 /f2} 7 }2+i2 a}, | 4 joo ijfo) [2-22 and by Eqs. (6.2262) and (6.2266) we get X{0] = P[o] =0 X[4] = P{2] X[1] = Of0] =2 +52 X[S] = Q[2] =2 +12 X{2]= Pl) = = X[6]=P(3]=74 XP] = Of = 2-32 X(T] = Q[3]=2-32 which are the same results obtained in Prob, 6.57, 6.60. Consider a causal continuous-time band-limited signal x(c) with the Fourier transform Xu), Let x[a] = T.x(nT,) (6.233) where 7, is the sampling interval in the time domain. Let X[k] =¥(k do) (6.234) where Sw is the sampling interval in the frequency domain known as the frequency resolution. Let T, be the record length of x(¢) and let wy be the highest frequency of x(t), Show that x(n] and X{k] form an N-point DPT pair if T, wm oul ie and N= (6.235)|ALS AND SYSTEMS 387 CHAP. 6] FOURIER ANALYSIS OF DISCRETE-TIME SI Since x(1) = 0 for ¢<0, the Fourier transform Aw) of (1) is given by [Eg. (5.3) X(w) =f" (ayer de= fxr) ede (6.236) Let T, be the total recording time of x() required to evahiate XC}, Then the above integral can he approximated by a finite series as X(w) =r D xts,ye where 1, =n At and T,= Nas. Setting w~, in the above expression, we have X(w,)= ar E x(de (6.237) Next, since the highest frequency of s(4) 8 ayy, the inverse Fourier transform of X4u i given by [Eq (5.3231 Xu) et dus (6.238) Dividing the frequency rane ~wy Sw Sey into N (even) intervals of length Su, the above integral can be approximated by ae ry SE Xaver where 2ujy~ Nu. Setting ¢~1, in the above expression, we have Ba x)= 5 5 (6.239) Since the highest frequency in x(1) is wy, then from the sampling theorem (Prob. 5.59) we should sample x(1) so that where T, is the sampling interval. Since T, ~ Ar, selecting the largest value of Ar (the Nyquist interval), we have Ar aN and ow (6.240) Thus, 48 a suitable even integer for which ee ee (6.241) 7, Bw 7 From Fg. (6.240) the frequency resolution Aw is given by sy a 2e DEN Be on ON ONT, (oon)358 6.61. FOURIER ANALYSIS OF DISCRETE-TIME SIGNALS AND SYSTEMS (CHAP, 6 Let (, =m At and a =k Aw. Then 2m ln 1 tyeee = (MAN(K Be) = nk x m= Sek (6.243) Substituting Eq. (6.243) into Eqs. (6.237) and (6.239), we get Xe dw) = Eo ata(n dre 12r/8i (6.244) bw 82 ; and a(n) = 52 x(k) erm (6.245) Rewrite Eg. (6.245) as Then from Eq, (6.244) we note that X(k Sw) is periodic in k with period N. Thus, changing the variable k =m — in the second sum in the above expression, we get dw fiver x x(n) = S2{ Exe awyeermh se E Komaw errr ao! : = SEL xb de) eres (6.246) Multiplying both sides of Eq. (6.246) by At and noting that Sw Ar = 2m/N, we have 1st x(n AN Ar= EX (kaw) eter ek (6.247) Now if we define afm) = Sacre) = Tx(n) (6.248) X[K] = x(k Aw) (6.249) then Fs. (6.244) and (6.247) reduce to the DFT pair, that is, x)= Sates ko Wa Ainl= 2S xt wet n= 0h gm (a) Using the DFT, estimate the Fourier spectrum X(w) of the continuous-time signal x(t) se tu(t) Assume that the total recording time of x(t) is T,=10 s and the frequeney of x(t) is wy = 100 rad/s. (b) Let XLk] be the DFT of the sampled sequence of x(1), Compare the values of X[0], X{1], and X10] with the values of (0), X(Xo), and XQ0 Ae). ighestCHAP. 6] FOURIER ANALYSIS OF DISCRETE-TIME SIGNALS AND SYSTEMS 359 (a) From Eq. (6.240) 100010) sigs Thus, choosing N= 320, we obtain du = B= 5 =0.625 rad Are so 0031s and Then from Eqs. (6 244), (6.249), and (1.92), we have X[k] = E are(marye eres 1S emwung =e 250° 716) (6250) which is the estimate of X(k Ae) (Setting k = 0, k= 1, and k= 10 in Eq. (6.250), we have 0.031 y[0) = MW 7 page * | x O81 _ gusset (> gos joms ~5* 0.031 X{10) = = 0.159610 (0.0896 = 70.189 From Table 5-2 XC) 0 ut) 9 Xo) = and x0) X(Bo) = X(0.625) = pas = OAK X(1 Aw) =X(6.2 i san Aw) = X(6.25) = Fae ae ~ OSB Even though x(¢) is not band-limited, we see that XTE) offers a quite good approxima: tion to X(w) for the frequency range we specified360 IS OF DISCRETE-TIME SIGNALS AND SYSTEMS [CHAP. 6 Supplementary Problems 6.62. Find the discrete Fourier series for each of the following periodic sequences: sll) in(0.17) os] barn) + Sin’2.4zrn) 09,9 Ole (a) al (b) ala] for ale Ans. (a) tn () ale) = 219") = 0.1 (co) alm] = C1 =J0.8) 0" + U1 + 0.5) 6", 05 ~ 04 6.63. Find the discrete Fourier series for the sequence x[] shown in Fig. 6-40, . igen oe Ans. stn}= ¥ cget™4".04 = X= asin SF Ja ssn }e] ig. 6-40 6.64, Find the trigonometric form of the discrete Fourier series for the periodic sequence x[n) shown in Fig, 6-7 in Prob, 6.3 3 7 zo Ans, x{n)= 5 ~ cos en —sin =n ~ cos rn tn of the following sequences: 6.65. Find the Fourier transform of (a) xln)= alla <1 (B) sla} = sin Qn) 1) <= (0 aln)out=n= 1] Ans. (a) X(Q) (by XCD) = ~jr[ BEM = 24) HEA = 2), 94 [pl = 1 (©) MQ = 7 6M) ~ quillsCHAP. 6] FOURIER ANALYSIS OF DISCRETE-TIME 6.66. 6.67. 6.68. 6.69, 6.70. |GNALS AND SYSTEMS 361 Find the Fourier transform of the sequence x1] shown in Fig. 6-41 Ans, X(Q)~ s2sin 1+ 2sin20 + 3sin3M) Fig. 6-41 Find the inverse Fourier transform of each of the following Fourier transforms (a) X00) = cos.) &) Ma=)0 Ans. (a) x{n]= $6ln— 2] + $4ln + 2] ) atae((CD/" neo io no Consider the sequence y{/) given by nye islel even (te Express y(A2) in terms of X(0). Ans ¥(Q)= 3X(O) + 4X 2) Let 1 nis? ale] {0 Inl>2 (a) Find ylnl=atn)+ a(n (6) Find the Fourier transform ¥() of yf) 5(1 = ini/3) nls 0 inl> 5 sin(2.5) | Sincosay | ins (a) tnd = { Ans (o) ylnl={ (6) Yny= Verify Parseval’s theorem (Eq. (6.66)] for the diserete-time Fourier transform, that is, Er (xfn]? = sf x(O)F do Hint: Proceed in a manner similar to that for solving Prob. 5.38,362 6m, 6.72. 623. 6.74, FOURIER ANALYSIS OF DISCRETE-TIME SIGNALS AND SYSTEMS [CHAP. 6 A causal discrete-time LT system is described by vim] tele t+ fol 2}= ata) where x[n] and y[n] are the input and output of the system, respectively (a) Determine the frequency response H(12) of the system (b). Find the impulse response fil] of the system, Ce) Bind yin) if cla) =u] 1 Ans. (a) HD) = seta (bd hin) = 12% fe) vlad = Gala Consider a causal discrete-time LTI system with frequency response H(Q) = e( H(0)) + j Im{ H(M)) = 4( 9) + 1B() (a) Show that the impulse response hin] of the system can be obtained in terms of A(Q) or BLD) alone. od HS) and. {nt at o Re{ H(2)) =A(2) = 1 + cos (a) Hint Process in a manner similar to that for Prob. 5.49, (b) Ans H(Q)= 1b e" Ala) = ian] + ale =U Find the impulse response Ain] of the ideal discrete-time HPF with cutoff frequency 2, ( vind =00,2.3,3,2.1,0,0) Consider the sequences x] and k(x] in Prob. 6.80, (a) Find the 4point DET of xl), ile], and y(n) () Find sla] by taking the [DFT of YI] Ans, (a) (X10), X(), X21 X13) = (4.0.0.0) [H(0), HU, HQ), HBD=B.~ 3.1.30 (¥(0), YU], Y12}. YET = [12,0,0,0) (6) yn =13,3,3.31 Consider @ continuous-time signal x(1) that has been prefiltered by a low-pass filter with « cutoff frequency of 10 kHz. The spectrum of (1) is estimated by use of the N-point DFT. The desired [requency resolution is 0.1 Hz, Determine the requized value of N Gesuming & power of 2) and the necessary data length 7) Ans No 2" and T, = 13.072 s