INTRODUCTION 1.1 Signals play a major role in our life. In general, a signal can be a function of time, distance, position, temperature, pressure, ete. and it represents some variable of interest associated with a system. For aareie nan clectrical system the associated signals are electric current and voltage, Ina mechan aaa ne anociated signals may be force, speed, torque, et. In addition to these, some examples of signals that we encounter in our daily life are speech, music, picture and video signals. A signal can Ntepresented in a number of ways. Most ofthe signals that we come across are generated naturally Homever there are some signals that are generated synthetically. In general, a signal carries informa- tion, gnd the objective of signal processing is to extract this information. Cenal processing is a method of extracting information from the signal which in turn depends on Be type of signal and the nature of information it carries. Thus signal processing is concemed with representing signals in mathematical terms and extracting the information by carrying out the algorih- mic operations on the signal. Mathematically, a signal can be represented in terms of basis functions in the domain of the original independent variable or it can be represented in terms of basic functions in 2 transformed domain. Similarly, the information contained in the signal can also be extracted either in the original domain or in the transformed domain. A system may be defined as an integrated unit composed of diverse, interacting structures to per form a desired task. The task may vary such as filtering of noise in a communication receiver, detection of range of a target in a radar system, or monitoring the steam pressure in a boiler. The function of system is to process a given input sequence to generate an output sequence. It is said that the origin of digital signal processing techniques can be traced to the seventeenth cea tury when le difference methods, numerical integration methods, and numerical interpolation met ne — ‘veloped to solve physical problems involving continuous variables and functions. There n tremendous growth since then and today digital signal processing techniques are applied i= almost every field. The main reasons for such painges of ch wide applications sous advantages © eee plications are due to the numerous 1.1.1 Advantages of Digital Si i ignal Processi Analog Signal Processing (ASP) nema uees Digital circu ; re rctit do not depend on precise values of digital signals for their operation. Digital cites ges in Component values, They are also less sensitive to variations in empera™">ageing and other extemal parameters. Digital processing is, stable, reliable, flexible, predictable and repeatable. Ina digital processor, the signals and system coefficients are represented as binary words. This enables one to choose any accuracy by increasing or decreasing the number of bits in the binary word. Digital processing of a signal facilitates the sharing of a single processor among a number of signals by time-sharing. This reduces the processing cost, size, weight and maintenance per signal. Also DSP ‘can save both filtered and unfiltered data for further use. Digital implementation of a system allows easy adjustment of the processor characteristics during processing. Adjustments in the processor characteristics can be easily done by periodically changing the coefficients of the algorithm representing the processor characteristics. Such adjustments are often needed in adaptive filters. Digital processing of signals also has a major advantage which is not possible with the analog techniques. With digital filters, linear phase characteristics can be achieved. Digital filters can be made to work over a wide range of frequencies by a mere change in the sampling frequency. Also multirate processing is possible only in the digital domain. Digital circuits can be connected in cascade without any loading problems, whereas this cannot be easily done with analog circuits. Storage of digital data is very easy. A signal can be stored on various storage media such as magnetic tapes, disks and optical disks without any loss. On the other hand, stored analog signals deteriorate rapidly as time progresses and cannot be recovered in their original form. For processing very low frequency signals like seismic signals, the analog circuits require inductors and capacitors of very large size, whereas digital processing is more suited for such applications. 1.1.2 Disadvantages of DSP over ASP ‘Though the advantages are many, there are some drawbacks associated with processing a signal in the digital domain. Digital processing needs ‘pre’ and ‘post’ processing devices like analog-to-digital and digital-to-analog converters and associated reconstruction filters. This increases the complexity of the digital system. Also, digital techniques suffer from frequency limitations. For reconstructing a signal from its sample, the sampling frequency must be at least twice the highest frequency compo- nent present in that signal. The available frequency range of operation of a digital signal processor is primarily determined by the sample-and-hold circuit and the analog-to-digital converter, and as a result the frequency range is limited by the technology available at that time. The highest sampling frequency is presently around 1 GHz reported by K. Poulton et al, in 1987. However, such high sam- pling frequencies are not used since the resolution of the A/D converter decreases with an increase in the speed of the converter. A variety of analog processing algorithms can be implemented using passive circuits employing inductors, capacitors and resistors that do not need any power, whereas 2 DSP chip containing over four lakh transistors dissipates more power around 1 watt. Moreover, active devices are less reliable than passive components. But the advantages of the digital process- ing techniques outweigh the disadvantages in many applications. Also, the cost of DSP hardware is, decreasing continuously. Consequently, the applications of digital signal processing are increasing rapidly.Classification of Signals and Systems Cc "GC; Applications of DSP ications of digital signal processing that are often encountered in daily life are listed Some selected ap} as follows: apt a 1. Telecommunication Echo cancellation in telephone networks, adaptive equalisation, ADPCM transcoders, telephone dialing application, modems, line repeaters, channel multiplexing, data vvmmnication, data encryption, video conferencin, cellular phone and FAX. 2, Military Radar signal processing, sonar signal processing, navigation, secure communications and missile guidance. 4. Consumer electronics Digital Audio/T¥, electronic music synthesiser, educational toys, FM stereo applications and sound recording applications. 4. Instrumentation and contro! Spectrum analysis, position and rate control, noise reduction, ddata compression, digital filter, PLL, function generator, servo control, robot control and process. control. 5. Image processing Image representation, image compression, image enhancement, image resto- artien mage reconstruction, image analysis and recognition, pattern recognition, robotic vision, satellite weather map and animation. 6. Speech processing Speech analysis methods are used in automatic speech recognition, speaker setrcation and speaker identification. Speech synthesis techniques includes conversion of wst- ten text into speech, digital audio and equalisation. 7. Medicine Medical diagnostic instrumentation such as computerised tomography (CT). X-ray scanning, magnetic resonance imaging, spectrum analysis of ECG and EEG signals to detect var jous disorders in heart and brain, scanners, patient monitoring and X-ray storage/enhancement 8 Seismology DSP techniques are employed in geophysical exploration for oil and gas, detection of ‘underground nuclear explosion and earthquake monitoring. 9. Signal firing Removal of unwanted background noise, removal of interference, separation of frequency bands and shaping ofthe signal spectrum. CONTINUOUS-TIME AND DISCRETE-TIME SIGNALS 1.2 Signals can be classified based on their nature and characteristics in the time domain. They are broadly classified as (i) continuous-time signals and (ii) discrete-time signals. A continuous-time signal is @ mathematically nee function and the function is defined continuously in the time domain. On the other and, discrete-time signal is specified only at certain time instants. ‘The amplitude of the ignal between two time instants is not defined. Figure 1.1 shows typical continuous-time Discrete-time Signals—Sequences A discrete-time sij Ocserete sine sign at Yalue defined only at discrete points in time and a diseret-ime system Prodi rete-time signals. A discrete-time signal is a sequence which is a function 5 — cfined on the positive and negative integers, that is, x(n) = (x(n)} { XD, x0) 2D el nese the tT e nis an integer indicating the sample numbers as counted 0. The negative values of n correspond to negative time. The function’ of up-arrow represents the sample at m = from a chosen time origin, SeenAo 5 t (@) Continuous-time signal Arn =r Tr 0 T ar a (©) Discreto-time signal Fig. 1.1 Continuous-Time and Discrete-Time Signals ris referred to as a sequence of samples, or sequence in short. Ifa continuous-time signal x() is sampled every T seconds, a sequence x(n7) results. In general, the sequence values are called samples and the interval between them is called the sample interval, T. For convenience, the sample interval T's taken as 1 second and hence x(n) represents the sequence. The important sequences are the unit sample sequence, the unit step response, the exponential sequence and the sinusoidal sequence. 1.2.1. Unitimpulse Function ‘The unit-impulse function is defined as 30) =0,140 ay and j S(de =1 ~ ‘The Eqs. 1.1 and 1.2 indicate that the area of the impulse function is unity and this area is confined to ‘an infinitesimal interval on the -axis and concentrated at ¢ = 0. The unit impulse function is very useful in continuous-time system analysis. It is used to generate the system response providing fundamental information about the system characteristics. In discrete-time domain, the unit-impulse signal is called a unit-sample signal.Classification of Signals and systems Itis defined as ' a(n) { , n=0 0, n#0 ‘Similarly, the shifted unit-impulse sequence 8{n — k] is defined as (13) 1, nak 0, n#k which is shown in Fig. 1.2(b). ' am-4) -{ (4) ‘The shifted unit-impulse sequence 8{n + k] is defined as an += ff a sy 0, n#-k which is shown in Fig. 1.2(b). 1.2.2. Unit-step Function The integral of the impulse function 8(4) gives, t 1 #>0 [ama={ A a6 Since, the area of the impulse function is all concentrated at t= 0 for any value of # <0 the integral becomes zero and for t > 0, from Eq. 1.2, the value of the integral is unity. The integral of impulse function is called the unit-step function, which is represented as 1 t>0 = ut) ff peed an The value at ¢= 0 is taken to be finite and in most cases it is unspecified. The discrete-time unit-step signal shown in Fig. 1.2(c) is given by on fe n20 0, n<0 aa Similarly, the shifted unit-step sequence u(n ~ k] is defined as un -k) = n2k {i nck as) which is shown in Fig. 1.2(4). The shifted unit-step sequence u{n + &] is defined as nek += 7 {a n<-k (1.10) which is shown in Fig. 1.2(d).Linear Time invariant systems. C207 ——____— DISCRETE CONVOLUTION 5.3 amouton is & mathematical operation a and x,(")- . x= n)* 20) = Y Wx (n-b (5:10) 53.1 Representation of Discrete-Time Signals in Terms of Impulse ‘An atbitrary signal x(n) can be resolved into a sum of unit sample sequences. The signal x(n) shown in Fig. 5.7 can be expressed in terms of unit sample sequences as. x(n) = J x(K)S(n-k) 2 = a(-2) Hn+2)+a(-1) H+ 1)-+2x(0) Hn) +x(1) Hn ~ 1) +42) Hn — 2)+.x3) Hn - 3) =(1) Hn+2)-+(-2) Hert 1) +(2) Hn) + (1) Hn = 1) + (-1) Hor = 2) + (1) Ht - 3) Hence, any arbitrary signal can be represented in terms of the weighted sum of impulses. 8 (0) 22510 1 2 3 (o) Fig. 5.7 _ Representation of a Signal as Unit Sample Sequences 53.2 Response of LTI Systems to Arbitrary Inputs (Convolution Sum) Consider a Hinear system whose behaviour is specified by the impulse response h(n). Let x(n) be the “*Put signal to this system and corresponding output signal be y(n) as shown in Fig. 5.8. 7 2b fo the output signal y(n), resolve the signal x(7t) into a weighted sum of impulses, and the Tr itd time shift property of the LTI system can be used. ie x(n) =n), then y(n) =h(n) =HTKn)] “OY anbtrary signal x(n), a(n)= Y x(k)8(n-k) eightage to the impulses at n=. the system to the signal x(n) is given byzat) paisguinecnne yn) = HEX) uf z ssn] = FM uisa-wl= Y xWka-*) ~ ) by using the time invariance property. Note: H[&m — 4)]= ~ i" ce Alo) eee tae neuer | ome © Fig. 5.8 ‘The above expression which gives the response y(n) of the LTI system as a function of the input signal x(n) and impulse response /(1) is called the convolution sum. “The various steps involved in finding out the convolution sum are given below 1. Folding: Fold the signal h(k) about the origin, i.e. at k = 0. 2. Shifting: Shift h(-K) to the right by 1, if m, is positive or shift A(-K) to the left by 1, if n, is negative to obtain /t(n, ~ 4). 3. Multiplication: Multiply x(k) by h(n, ~ &) to obtain the product sequence y,(k) = x(K) h(n, ~ k) 4, Summation: Sum all the values of the product sequence y,(k) to obtain the value of the output at time n =n, 5.3.3 Properties of Convolution ‘The convolution of two signals x(n) and h(n) is given by y(n) = x(n)eh(n)= Y) x(E)h(n—k) or y(n) = h(n) « x(n) = Y hyx(n-k) (1D In the first case the impulse response /(n) is folded and shifted, and x(n) is the excitation signal. In the second case, the input signal x(n) is folded and shifted. Here h(n) acts as the excitation signal. Commutative Law Convolution satisfies commutative law, i. x(n) + h(n)=h(n) + x(n), which is shown in Fig. 5.9. Associative Law [x(n) + hy (n)] # h(n) =x(n) * (h(n) * h,@)] ‘Take LHS of the above equation. Consider x(n) to be the input signal to the LTI system with impulse response h,(n). The output y,(n) is given by y(n) = x(n) # hyn) ‘This y,(n) signal now acts as the input signal to the second LTI system with impulse response (7). ‘Therefore, en) = y(n) * bln) = (x(n) #hy(n)] * h(n)er the RHS of the equation, which indicates thatthe input xn) is applied to an equivalent fow consi nH and is given by sy h(n) = h(n) * h(n) qutput ofthe equivalent system tothe input xn) is given by a y(n) = x(n) * h(n) = x(n) + (h(n) * A) xn) a Lm x(n) + An) n(n) yy Leh xin) Fig.5.9 Commutative Property since convolution satisfies commutative property, the cascading of two systems can be interchanged asshown in Fig. 5.10. y(n) y(n) 1 T nya | mm Pe a hyo || ma HO | yey = mieneboten 0) patey mi) [20 Fig. 5.10 Associative Property IfN linear time invariant systems are in cascade with impulse responses h(n), h,(),-..t,(n), then the equivalent system impulse response is given by Rn) = h(n) « h(n) «~ * hyn) Distributive Law x(n) « (h,(n)+h,(n)) = x(n) * hy(n) +x) * h(n) ‘The distributive law states that if there are two LTI systems with impulse responses h,(n) and h,(n) aud are excited by the same input x(n), then the equivalent system has the impulse response as A(n) = h, (rt) + hl (n) In general, if there are N number of LTI systems with impulse responses h,(n), h(n). h(n) excited bythe same input x(7), then the equivalent system impulse response is given by h(n) = Sao ia This is Nothing but the parallel interconnection of individual systems. sequences ae 1, -lSns1 and h(n)= 0, otherwise Examp5.25 Soluti F jo a "The convolution of two finite duration sequences is given by yn) = y x(Oh(n-k) or y(n) = Y x(n hk) a Kem| x(n) n(n) Fig. E5.25(a) ‘Step I Plot the given sequence, as shown in Fig. E5.25(a). Step 2 To find the convolution sum y(n) When n=0, O)= DY x(kh(-k) Sox) he (1)-+2(0) h 0)+x(1) A (1) +> =0+(1) (1) #(1) (1) +(1) (1) +0 ---=3 When n=l, x= Y xAWhU-b tx) h (2) +x(0) h 4x0) h O)4-~ O+(1) (1)+(1) (1)+0---=2 ¥(2) = zy x(k) h(2-K) na = -¢x(-1) A 3) +2(0) h 2) +2(1) A (+ = 0+(1) (1)40--=1 When n=3, y@)= Y, xHhG-K=0 = yD= Yo xWkC1-b a $x(-1) h (1) +200) h C1) +21) h (2) 4 }+(1) (1) +) (1) +0 - When n=-2, 2) = Y xWAC2-H x1) h 1) +x(0) h -2)-+x(1) h (3+ =0+(1) (1)+0-ee When n=-3 yr) X= Ya h(3-b=0 : ms ‘The convolution signal y(n) is (See Fig. B5.25(b)] 2 Yn) = O,mS—3 and n=3 , it “Taso es Hn) = Fig. E5.25(b) Note: For the convolved signal, the left extreme and the right extreme can be found using the left and right extremes of the two sequences to be convolved. That is, ath, yaxth, where x, A, and y, are the left extremes of the signals x, ht and the right extremes of the signals x, h and y respectively, Alternate (Graphical) Method The given problem can be solved by using the graphical method as shown in Fig. E5.25(c). We know that Xn) =D) x(k) h(n) & y respectively. Similarly x,, hand y, are Whea n=0,(0)= >) y(k)= Sach =3 = = When n=1,y(1)= ) (= > xwaa—H=2 = ra When n=2, (2) z y= > 2x(h(2-k) =1 = rae S a{k)h(3-k) =0 a Wea =3,y3)= 5 90) eee When m=-LyD= ¥ y= > ACW =2 he When n=-2, 2) = x y,®= > x(k) h(-2-k) = 1 me When n= 3.43)- Fy (= F xyhe3-H=0 me = When these ee Sequence values are plotted in Fig. B5.25(c), we find that the result is identical tothe ult shown in Fig, B5.25(b).x(t) mh) | "| f 01% Ae ek ‘When n=0 pear ce) 1 1 Ut eek 01k Product Sequence, I-A) mk) alk) = x(k) XK) When n=4 mt—K) 40) WR) 1 1 012% +2. 8 A =) shit -#) Product Sequence, to right By one unit ath) = (4) (1 ~) When n=2 ne-k) 4 2) i | | 1 o12345 k 012 4 {2 K)s=oshift (1K) Product Sequence, to right by one unit srl) = (H)h12 = When n=3 | nee 012348 Biota k n3— K)seshitt (2 — W) Product Sequence, to right by one unit alk) = (8) (3 ~K) Fig-E5.25(¢) Graphical Method (Contd.)mer toric eee | a Whenn=—1 i-th) UK) H{-A-K) I | 1 1 S-2tot 8 “21012 k IN-A=K) shit ho (-K) Product Sequence, tof by one unit yh) = x) {=~ K) When n= -2 n-2-k) a(h) h{-2-K) | | i “32-101 k “2-10 7k 2-H) shift n= (1-4) Product Sequence, to left by one unit valk) = AK) H-2—K) When n= -3 -3-k) x(k) h-3-K) j 43-210 12k 241012 k 1-3-K) shift h =p (-2-4) Product Sequence, to left by one unit ¥-alk) = aK) {-3 - Kk) alr 3 2 1 a -3 sl 0 4 2 43 0 -2 Fisos) Dittse! pocsing 1h) x(3 #) Ax8-") When n=3 | — ri ial Product Sequence, y,(k) = h(k) (3 ~ A) = Kem hit x2 —H) 10 HE BY OE unit x(n) a a 2 ees . Ao mat ey 4 ee 2 Fig. E5.27(a) Solution Let y(n) be the resultant convoluted signal. Then ron) =D x(n W) ‘From the graph, x,=0.x,=2,4,=-1,h,=2 ‘Therefore, the left and right extremes of y(n) are found to be y= ath,=04(-1)=-1 yaa th=242=4 When n=-l = yep= Doh -» +.x(0) h (1) +x(1) A (-2)+2(1) h (-3) 4 20+(1) (+) ()40--=1ins sta ns | When n=0 . yO= Lancer 4+x(0) h (0) -+4(1) h (1) +4(2) h (-2) + = 0+(1) (2)42) (D405 4 When n=l y= Daa ' 5 (0) (1) 41) fe OY a02) fe (1) #403) h (2) + i = 0+(1) (-2)+(2) (2)+3) (It =5 | When n=2 - Dawie-w 4+.2(0) he (2)+2(1) he (1) +x(2) h (O) + = O+(1) (1) +2) (-2)+3Q2)+ LwrG-w -+.2x(0) h (3) +2(1) h (2) +2(2) h (+ +(1) (0) +(2) (-1) +3(-2)+0 ~ = Dawka-» | = -#2(0) h (4) +x(1) hB)+x(2) AQ) 8 = 04(1) (0)+@) (043-140 --=-3 | ‘These sequence values are plotted in Fig. E5.27(b). Fig. ES.27(0) xample 5.28 Compute the convolution y(n)=x(n) * h(n) of the signals vy =fnon and nn =fr?4} Solution ‘The sequences of the given two signals are plotted in Fig. E5.28(a). From the graph, = 2, 4,22, h=-3,h,=0 Hence the left and right extremes of the convoluted signal y(n) are calculated as i +h, ‘The convolution signal y(n) is given as yn) = L Mx(n—B) i5) Prenton x(n) Fig. E5.28(a) O-+x (-2) h 2)-+x 1) h (1) +2 (0) h (O)-+x (1) h (-1) +x (2) h (-2)+404--- + (1) (0)+ (1) (0) + (0) (4)+(1) (-3) +(1) (-2)+0 += -5 When n=1 - yd) = = k)x(1 x(—2) h (3)+x(-1) hr (2) +2(0) A (1) +x(1) h (0) + x(2) h (-1) 4+ ++(1) 0)+(1) 0)+@) (0) +(1) 4)+(1) (-3)+04--=1 When n=2 - y(2)= 2 W)x(2-8) <+x(-2) h (4) +2(-1) A (3)+2(0) h (2) +21) h (1) + 2(2) A (0) +- = 0+(1) )+(1) (0)+ ©) +1) 4)+() (0) +--- = 4 When n=-1 5 yCD = = hk) x(-1- 8) Sb x(-2) A (1) x1) A (0) +x(0) h (1) +2(1) A (-2) + x(2) h (3) ++ = 0+(1) (0) +(1) (4) +(0) (-3)+(1) (-2)+(1) (1) + 04---=3, When n=-2 - y-2)= D a)xC2-#) set x(-2) h (0) +x(-1) h (1) +20) he (2) + x(1) h (-3) 4x02) ha) = 0+(1) (4)+(1) (-3)+(0) (-2)+(1) (I)+(1) (0) 40+. =2 When n=-3 : 9-3) = D nwxC3-¥ Srv x(-2) (1) +x) A (-2) +x(0) h (3) + x(1) h (4) +.4(2) h (-5) 4--- = 0+(1) (-3)+(1) (-2)+ (0) (1) +(1) (0)+(1) (0) + 0- 5ns nt tenn Cm 737 + 24 14 ial os jp==eiayo a4 0 | a4 “56 Fig. £5.28(b) When n=-4 4) =D neyxca—W ssta(-2) h (-2) +x(-1) h (~3)+2(0) he (4) + (1) A (-5) +202) h (-6) + = 04(1) (-2)+() (1) +0) +(1) (0) +(1) (0) + 04--=-1 + x(-2) (3) ball) h (4) +2(0) fe (5) + (1) h (6) +412) h (HT) + O+(1) (1)+04- ‘These sequence values are plotted in Fig E5.28(b). SOLUTION OF LINEAR CONSTANT COEFFICIENT DIFFERENCE EQUATION 5.4 A discrete-time system transforms an input sequence x(n) into an output sequence according to the recursion formula that represents the solution of a difference equation. “The general form of a difference equation is w yn) =-Y ap y(n—b) + Y Bx(n-h) (5.12) =I io where NV is the order of the difference equation. The output at time n is a weighted sum of past inputs, the present input and past outputs. The solution of the difference equation consists of two parts, i. y(n) =», (2) +y,(0)322 >) ipa siqa! Processing From this equation, it is clear that the present outP previous inputs. Comparing this equation t sample response /t (7) at time { Jay, OSNSM so 0, otherwise ut sample value depends only on present ang fo that of convolution, we can recognise b, / 4, as ht (F), the value of the unit k. Hence h(n) is given by h(n) As this is obviously of finite duration, it represents an FIR system. Example 5.36 A DSP system is described by the linear. difference equation 4) =0.2.x (n) = 0.5 x (n= 2)+0.4 x (n= 3) Given that the digital input sequence {—1, 1, 0,—1} is applied to this DSP system, determine the corresponding digital output sequence. Solution ‘Taking z-transform of the given linear difference equation, we get Ye) = 0.2 XG) 052° (+04 9X (2) ‘Therefore, He 72 =02-0527 +0427 The given input sequence is x(n)={—1, 1, 0,—1) and its z-transform is X@=-l4et-9 Therefore, Y(@)=H (@)-X(2) 0.24 0.224052? -11z? 4042-44 0.525 — O.4e-* ‘Taking inverse z-transform, we get the digital output sequence yn) = {—0.2, 0.2, 0.5,-1.1, 0.4, 0.5,—0.4) si eee Yo)+ HOD + ED) = x(n) +x(—1) Solution Gi : 1 jgiunon | Givea YO) + GAD 4 EMM ~2) = (0) +x(0=1) Taking z-transform, we get 3 4 1 ¥@)+=z a + GOVO+ ECV) = X@+271X(2) The poles are at z= The zeros are at zTesponse for the systems given by the following difer= Example 5.38 Determine the impulse ence equations. (a) y(n) +3y (n= 1)42y (n= 2) 82x (n) =x (NN) (b) y (nd =x(n)+3x(n— 1) 4x (n= 2) +2 (n= 3) Solution (a) For the impulse response if x(7) = 6(n), then y(n) =/(n) h(n) +3h(n — 1) +2h(n ~ 2)=26(n) — Gn = 1) Taking z-transform, we get H (o[l 43244227) =2-2" 2-21 : Taking inverse z-transform, we obtain h(n) = [-3(- "+ 5(—2)"Jun) (b) The impulse response for the given difference equation is h(n) = &n)+3Hn— 1) — 480 — 2) +260 - 3) Solution (i) The given difference equation of the system is 6 y(n — 1) — 0.08 y(n — 2) +.x(n) ie, y(n) = 0.6 y(n — 1) +0.08 y(n — 2) =x(n), To find the impulse response h(n) For an impulse input, x(7t)= &(n). Hence, y(n) Therefore, h(n) — 0.6 h(n — 1) +0.08 h(n — Taking z-transform, we get H (z)[1 — 0.627! +0.082~ Hence H@)= 1 ea 1 (1-062 +0.08z7) (1-042 +1-0.022 Ay> ee Taking inverse ¢-transform, we get the impulse response as (nd = 200.4) a(n) ~ (0.2) mn) 7 To find the unit step response s(n) For the unit step input, x(n) =u(n) Therefore, the difference equation becomes Ye) = 0.6 y(n = 1) +0.08 y(n = 2) = a(n), Taking z-transform, we get roli-0.627 +0.0827] ‘Taking inverse z-transform, we get the unit step response as s(n) = y(n) = dn) $04 ur) + L(02)uo) Alternate Method To find step response from the impulse response using convolution We know thatthe step response s(n) isthe running sum of the impulse response h(n). That is, sn) =hin)euin) =F hun) = 2 Here, u(n — k)=0 for k>n and u(n — k)=1 for k ) x(n)h(m—n, (mod(N))), m =0, = N-1L Therefore, from Fig. E6.17, we have 3 (0) = Y) x(nyh(-n, (mod(4))) = 1 = 5 yQ) = ¥; x(mpA(1—n,(moda))) = 4 = 3 (2) =) x(nph(2—n, (mod(4))) = 8 = 5 ¥)= Y x(n)h(3—n,(mod(4))) = 8 =isreteand rastFourter transforms €_379 Me im [sya eet ] x mie et kiN a fewgtin eaten ewgromr | % von . = (13) . nowy! = wr ) =n d x "Wie where [1+ Wt WA eon Wg aN is already determined WK), =X(K) + 4K) (14) k k xy@=ue( 4) and y@=ni( 4) Substituting the values of X,(A) and ¥,(K) in Eq. (14), we get k k WW, = 2, +O. =ux(£)enr( 5) FAST FOURIER TRANSFORM (FFT) 6.4 ‘The fast Fourier transform (FFT) is an algorithm that efficiently computes the discrete Fourier trans- form (DFT). The DFT of a sequence {x(n)} of length NV is given by a complex-valued sequence {X(k)} Not X(kK)= Y xine P™™, OSKEN-1 (6.17) =} Let W, be the complex-valued phase factor, which is an Nth root of unity expressed by W,=e 20 (6.18) Hence X(k) becomes Net X= Y x(t, OSKSN-1 (6.19) = Similarly, IDFT becomes Net xin)= +S. xcomgt ,0SnSN-1 (6.20) Nin __ From the above equations, it is evident that for each value of k, the direct computation of X(k) involves N complex multiplications (4NV real multiplications) and N— 1 complex additions (4N —2 real additions). Hence, to compute all V values of DFT, N* complex multiplications and N(W ~ 1) complex additions are required. The DFT and IDFT involve the same type of computations. Ifx(n) is a complex-valued sequence, then the N-point DFT given in Eq. (6.17) can be expressed as uA X= Xt 1X = Y, [sad sono = jin) n=O Equating the real and imaginary parts of the above equation, we have Nt X= © [eines Brn 5 (npsin 2K | (62) ial N W386 _) Digital Signal Processing Example 6.19 Given x(n)= (1, 2,3, 4,4,3, 2 1), find X(K) using DIT FFT algorithm. Solution We know that Wy; =¢ Hence, we = of 1. (7) Wi=e \*/ =cos n/4— jsin 1/4 = 0.707 - j0.707 AG} =cos n/2-jsin n/2 v A} Wp =e NSP = cos 3n/4— jsin 30/4 =-0.707 - j0.707 Using DIT FFT algorithm, we can find X(k) from the given sequence x(x) as shown in Fig. £6.19. ‘Therefore, X(k) = (20, -5.828 ~ j 2.414, 0, 0.172 -j 0.414, 0, 0.172 + j 0.414, 0, 5.828 + j 2.414) 2(Q)21 0a yo p< X(0)=20 re ZL (Q)24 Oy — fp 2 iy =-5.828-2.418 1 o © M2)=0 (9) =0.172-/0.414 _o x(6)=0 0 X(7) =-5.828 + 2.414 Example 6.20 Given x(n)={0, 1,2, 3, 4, 5, 6,7), find X(k) using DIT FFT algorithm. Solution Given N=8. Hence, Ww, = -0.707 -j 0.707 Using DIT FFT algorithm, we can find X(k) from the given sequence x(n) as shown in Fig, E6.20.(4), we ot Te) ee - Orel Re [X(-") (-bl (6) To prove aati = un We know that _X(K) = x [x@l= bs; x(ne (6) le” < Sx), since = Also, ol pak xeb= Same ™ 0 Wet =| xe s=0 [XC] <¥ bo), since |e’ | © n= ‘Comparing Eqs. (5) and (6), we obtain IX (K)I= IX (~ 4). ‘Determine the DFT of the sequence 1 = where & =0, tan =