Example 1.9 Determine if the following systems are time-invariant or time variant. @ y(n) =2(r) um" -) ati) ur) = = = a(— n) o Solution “ ue” we Given y(n) = T[e(n)] = 2(n) + 2(n — 1) i If the input is delayed by & units in time, we have +L y(n, k) = Tfe(n — k)] = 2(n - k) +2(n-&- 1) If we delay the output by & units in time then Yrr=3¢¢-8), unk) =2(n—k)-bata— kL y Here y(n, k) = y(n — k) . Therefore, the system is time-invariant. xC-Co-e) (ii) Given o¢C- het) ; y(n) = T[x(n)] = 2(-n) If fhe input is delayed by & units in time and applied to the system, we have y(n, k) = T{a(n — k)] =2(-n - k) If the output is delayed by k samples y(n — k) = 2(—(n ~ k)) = 2(-n +k) Here y(n, k) # y(n — k) , Therefore, the system is time-variant.Aifferentiation of X(z) IfX(z) = Z{x(n)}, then d A{n2(n)} = ~2 X(2) (2.16) | Proof: ry X(2)= D> a(nyern n==00 Differentiating the z-transform we have Multiplying on both sides wit | “eter! h 2 we get oo : 2M) _ > na(n)2-" n=—00 = 2(na(n)} Aowolution Theorem "X() = 2{e(n)}, and A(z) = Z{h(n)}, then 2{2(n) « hn)} = X(2)H(2) “ denotes the linear convolution of sequences, here a(n) « h(n)2.12 Digital Signal Processing Proof: We have ‘y(n) = 2(n) h(n) = x a(k)h(n — k) and ~ ~ ¥(2) = Z{y(n)}= Do [> oiinin~') a, n==co Lk=-09 = > > a(k)z7*h(n — k)27(n-8) n=—-00 k=-0o Interchange the order of the summation, Y(z) = > ax(k)2—* > h(n — k)z7("-4) k=-co n=-00 Replace (n — k) by in the second summation. Then Y(2)= > a(k)z7* > A(z k=—00 15-00 = X(z)H(z)Example 2.6 Find the stability of the system whose impulse response h(n) = (2)" u(n). Solution Given h(n) = (2)"u(n). Using Eq. (2.8a) we can find H()=— 5: |al>2 TheROC is lz] > 2. It does not contain unit circle. Therefore, the system is unstable. Fig. 2.5 , ei ROC of Example 2.6 eeRYN . The ROC is a ring or disk in t Properties of Region of Convergence he z-plane centered at the origin. The ROC cannot contain any poles. If x(n) is a causal sequence then the ROC is the entire z-plane except at z = If x(n) is a non-causal sequence then the ROC is the entire z-plane except a z= 00. . If2(n) is a finite duration, two-sided sequence the ROC is entire z-plane except at z = Oand z = oo. If x(n) is an infinite duration, two-sided sequence the ROC will consist ofa ring in the z-plane, bounded on the interior and exterior by a pole, not contait- ing any poles. . The ROC of a LTI stable system contains the unit circle. . The ROC must be a connected region.Wor each timpulee — oresponce eight pelos cohethes the cosecponeling systems Py caercel NY gdeeble’ oP in 9 a ohne ? am hinys auenr by hen? ® stn oe Gwen hent = a een) Ain to) for neo $0 the gytlem» te non coura! TOO steubility : © ‘ & them] < 2 ne+O : : ” n ry Es uta: N-0 2 oon Es naw: Syd —~ t te aye 142 2 2B “$0 the Sytem Ps stable Aeleamine_—__ % hin) = ginnx : KM £0 for ney S80 YAO eyetem Fe pons cautal “ha Stability , © EGhim< nz- _ © * g | ang), ns® edd veuec of ns | sioox) =1 a * : E thinr) +0 = ns-0 Soa . system. fe anctable o him z eum) the system fe Son n nk) Ay(n—k) ; ane the system is time-variant. v Example-1.9 Determine if the followin, @® yn)= 20) tae » o vi Solution is t bug — rey1.24 FIR and IIR Systems Linear time invariant systems can be classified according to the type of imp! sponse. They are 1. FIR system 2. IIR system. ulse fe FIR system Definition If the impulse response of the system is of finite duration, then the system is called a Finite Impulse-Response (FIR system). An example of a FIR system is forn = 1,2 forn=1 1.134) forn =0,3 otherwise Mn) = ConeDiscrete-time Signals and Linear Systems 1.63 TIR system Definition An Infinite Impulse Response (IIR) system has an impulse response for infinite duration. An example of an JIR system is h(n) = a"u(n) (1.135) 1.25 Stable and Unstable systems An LTI system is stable if it produces a bounded output sequence for every bounded input sequence, If, for some bounded input sequence +(n), the output is unbounded (infinite), the system is classified as unstable. Let -(n) be.a bounded input sequence, h(n) be the impulse response of the system and y(n) be the output sequence. Taking the magnitude of the output xs h(k)e(n - k) (1.136) Ike 00 we have {y(n)| = We know that the magnitude of the sum of terms is Jess than or equal to the sum of the magnitudes, hence ‘ u(r) < SF lala) Ile(n - &)| (1.137) ae Let the bounded value of the input is equal to M, the Eq. (1.137) can be written as y(n) s MS h(k) F (1.138) ks-00 The above condition will be satisfied when a YS [hI < 00 (1.139) kano So, the necessary and sufficient condition for stability is DY |h(n)| < 00 (1.140) Example 1.11 Test the stability of the system whose impulse response Nn) = (1) u(n) (BIE AU Dec’ 03)1,64 Digital Signal Processing Solution The necessary and sufficient condition for stability is oo > [h(n)| < 00 ne=00 Given h(n) = (1/2)"u(n) Ye ln(n)| “> |U/2)"u(n) © 71/2)" a) eee se 2 < 00 "7 Hence the system is stable,Example 1.12 Determine the convolution sum of two sequences z(n) = {3, 2, 1, 2}; A(n) = (1, 2 1, 2} Solution Step 1 The sequence z(n) starts at n = 0 and h(n) starts-at no = —1. Therefore the starting time for evaluating the output sequence y(n) is n =m +n2 = 0+(-1)= ah Step 2 Express both sequences in terms of the index k. 3 xi) 2 2 ’ 2 O12z23k 2-10123 > k Step 3 Fold h(k) about k = 0 to obtain h(—) 2 W-k) 2 2 hl+k) 1 1 1 2 2-10 1k 32-10 As starting time (o evaluate y(n) is -1, shift h(A) by one unit to left obtain h(-1 - k)Discrete-time Signals and Linear Systems 1.65 From Eq. (1.123) w= > alkya-1-4) kano Multiply the two sequences 2(k) and h(—1 — k) element by element and sum the products. = y(—1) = 0(2) + 0(1) + 0(2) + 3(1) + 2(0) + 1(0) + 2(0) =3 Increment the index by 1, shift the sequence to right to obtain h(—) and multiply the two se- ‘quences z(k) and A(—k) element by element and sum the products wo)= | F_a(e)h(-k) (2) + O(1) + 3(2) +2(1) +1(0) +2(0) = 8 cmd w= YL z(k)h(1-k) we / = 0(2) + 9(1) +2(2) +1(0) +2(0) w= FE zwe—H = 3(2) + 2(1) + 1(2) +201) 2 E zAG-w bio (0) + 2(2) + 1(1) + 2(2) Wa = SS 2(kNA-K) 7 soe = 34094-2(0) + 1(2) + 2(1) +002) w= E z(K)K(S—) © 25(0) + 2(0) + 1(0) + 2(2) + 0(1) + 0(2) +0(1) Site Mee ase Fig. 1.58 Operation on signals z(n) and “a in) to compute convolution y(n) = {3,8,8,12,9,4,4}1.66 Digital Signal Processing To check the correctness of the result sum all the samples in -r(n) and multiply by the sum of all samples in h(n). This value must be equal to sum of all samples in y(n). : In the given problem 5.2r(n) =8, Sh(n) =6 Eu(n) = 48 Ea(n)-ZA(n) = Ly(n) proved. Therefore, the result is correct, Method 2 Tabulate the sequence x(k) and shifted version of h(k) as shown below ‘The starting value of n = -1 k ~4[-3|-2|-1/o0 [1 }2 [3 |a}s|e]a 2(k) | % 2 —| h(-1-k) eae (0 - k) 2 |) | ala h(1~ k) 2 [ashe |a n=2 | h(2-k) alrlalg [n=3 | n@-4) 2 | ai |i }a n=4 | A(4—k) alaleh n=5 | A(5-k) Blarj2ia w(-1) = 301) =3 y(0) = 3(2) + 2(1) =8 y(1) = 3(1) +2(2) + 1(1) =8 y(2) = 3(2) + 2(1) + 1(2) + 2(1) = 12 u(3) = 3(0) + 2(2) + 1(1) + 2(2) =9 w(4) = 1(2) +2(1) = 4 (5) = 2(2) = 4 y(n) = {3, 8, 8, 12,9, 4, 4) Method 3 aa Given x(n) = {3, 2, 1,2}, h(n) = (1, 21,2) in) Step 1: Write down the sequence x(n) and h(n) as shown,Discrete-time Signals and Linear Systems 1,67 Step 21 Multiply cach and every sample in (1) by the samples of a(n) and tabulate the values. Step 3: Divide the elements in the table by x0) drawing diagonal lines as shown, Ae te AK 26 4 2, 4, Mo Ay ays Ary Step 4: Starting from the left, sum all the elements in each strip and write down in the same order BOBS AIOE LALA LAA D+ 24 @ 3, 8,8, 12,9,4,4 Step 5: The starting yalue of 1. 1, mark the symbol | at time origin (n = 0). y(n) = (3, , 8, 12, 9, 4,4}Discrete-time Signals and Linear Systems 1.95 Example 1.24 Determine the impulse response h(n) for the system described by difference equation y(n) + y(n = 1) — 2y(n - 2) = a(n — 1) + 2a(n — 2) Solution Given y(n) + y(n — 1) — 2y(n — 2) = a(n — 1) + 2x(n — 2). Since M = N = 2, the homogeneous solution includes an impulse term. The total response is given by y(n) = un(n) + pln) For input x(n) = 6(n), the particular solution yp(n) = 0 = y(n) = ya(n) ‘The homogeneous solution can be found by equating the imput terms to zero, that is y(n) + y(n—1) — 2y(n- 2) =0 (1.239) Let the homogeneous solution yj,(n) = A". Substituting this solution in Eq. (1.239) we obtain the characteristic equation nano} — aan? = 0 AM2[\? + A — J = => 4+A-2=0 ‘Therefore, the roots are 1, ~2 and the general form of the. solution to the homoge- neous equation is n(n) = e(1)” + 2(-2)" + Asn) (1.240) From the difference equation y(0) + y(-1) — 2y(—2) = 2(—1) + 22(-2) y(1) + y(0) - (1.241) Substituting n = 0,n = 1 and.n = 2 in Eq, (1.240) we get y(0)=ata+A y(1) = ey ~-2ey y(2) = e1 — deg (1.242)1.96 Digital Signal Processing from which ¢, = 1;¢2 = 0;A = -1. Substituting these values in Eq. (1.240) yields y(n) = u(n) — d(n) =u(n-1) (1.29)Step mesponse: Given causal system: Yin) +4 cn-t) -agen-a) = xen-+ ) + axtn-2) FO step esponse: + y=kuen) XEN) =uen) and the partiular solution pte Substituting XEN) & Uypin) in given system Yen) tycn-1) ~ayen-a) = xen-I+ axcn-a) Kun) + kucn-1) ~arutn-a) = ULN-1) 4 gutn-ad- 709 n=2 whene none of the terms vanish, we gee KUCR) + Kutg-t) ~akulg-2)= Ula+) +aQUca-2) KO) + ken ake) = cn 4at)- K+tk-ak =3 K=0 t+ Yptni= Kutn) = 0-utn) © “Ypin) = 0 —ad The toa! stesponse: ytn)= Yrin) + Ypen) - 62! we know that Yon) ty trmpulse nesponse : a 0 Gin) 2 Oy l-1) + Ge-a) 03) Ststipute tn bayer) 6 1D Inca (‘ 4 Yom = CC) 4 fg (-2) 410) - ¢u) Tom éy¢a): yw > ater Yurd = cr- alr Ton causal system Equation : Yi) = -yio-1) Hay eo-ay + xCo-1) # AHLO~?? v. ywro yun = yer) + ayll-2) 4 xee-1) + antl- “yo t Comparing Yytoy & ye’: (140270 a> Cre 762 C1 alas] => ¢reay alas! -3Cage! Ca = Hg Cas -033y4 Cre Cae —¢-00394) Cr =0:33y = 0.334, ¢a> 0-334 fubsHtute fn equationcy) a)‘. he step mesponse fe “ n YU = 9.33u0" ~9.33U¢-a) +0 i: 4 Vie yen) = 0: 3340 a eny -o0-33y ¢-d) uen) 4ouinree Example 1.37 Determine and sketch the magnitude and phase response of y(n) = len) +2(n — 2)] Solution Given y(n) = 5le(n) +2(n—2)] mp (1.408) ¥(e#) = > y(n)ensen x [2(n) + a(n - 2)Je~4" =3 ys 2(n)e~Son 4. > a(n — 2)e~50" ae 3 1 “ : = 5(X(e) + eX (e)] — Xe) ~2jw = z [i +e ' . H (ei) = VC) _ Lee _ 1+ cosy jsindy (1.405 ) 2 21.136 Digital Signal Processing (e) is antisymmetric with a period of We know | (c?~)| is symmetric and 2H ral0 Su ST 2z in w. We only need to know He”) over the inte (1 + cos 2u)? + sin? Qu = z Tp cos? uy + 2.008 2u + sin? Quy V2(1 + c05 2) = ave 2eos?w 1 = =V4c0s? w = cosw H(e*) = where C410) LHe) = tan? (3) = tan“! (4as*) = tan7}(-tanw) = -w -w for H(e%)>0 =-w+n for H(e%) <0 Jol x6 | 2/4 | 2/3 oh | is an/4 | om —0.707 |:-1 0.707 1 (411) LHe, (1.412) H(e) | 1 | 0.812 | 0.707 0.5 -0.5 jae») | 1] 0812 | 0.707 | 05 | 0 Zuo) | 0 | -2/6 | -x/4 | -n/3 | -7/2 EZ ° ° =a ro 2 Ao n® md Fig. 1.71 Magnitude response & phase response of the example 1.37. 2 , (0) Given X(z) = mea Dividing, X (z) by 2 we have =o C2 C3 “7-17-24 723 X(z) P4245 (2.36 Digital Signal Processing where C, C2, C3 can be evaluated using Eq. (2.60). X(2) wpe att a =0- 0 =O = Be Ne Ie =3) X() yee = G9 gO & Mae =A) x 24245 c= -9O) = Teale! Substituting Ci, C2, Cg values we have (i) In case when the ROC is 2 < |2| < 3, shown in Fig. 2.8, the signal x(a) is two sided. The poles z = 1 and z = 2 provide the causal part and the pole z = 3 the anticausal, part. Fig. 2.8 ROC of example 2.21 00) Thus, by inspection x(n) = u(n) — (2)"u(n) — (3)"u(-n - 1) Gi) In case when ROC is |z| > 3, shown in Fig. 2.9, the signal (7) is causal andl the three terms in Eq. (2.71) are causal terms. Ime) z-plane Fig. 2.9 ROC of example 2.21 bili). Fig. 2.10 ROC of example 2.21 biil) Therefore, a(n) = u(n) ~ (2)"u(n) + (3)"u(n)fee em avanmUL 2.97 jiy In case when the ROC is |z| < 1 shown in Fig, 2.10, the signal -(n) al the terms in Eq.(2.71) are anticausal terms. Therefore, a(n) = ~u(—n = 1) + (2)"u(=n ~ 1) ~ (3)"u(—n — 1) = [-1+ (2)" — (3)"Ju(-n - 1) is anticausal