A Signal combedefined as O ' which conveys information about the behaviovY or nature of a physical cane «raping fro ane et (domain) anole (ange + |Asignal may also be defined a: Ex} monochromatic pleture > brighness € (2sraHial yorviables) video signal > colovy alntensity €4(20 & 1D>4ime) sen) | Spain| at sagt eta ‘eeu foefion: guy 05 oti sidan (rani dt =") Js prenunder the 61) ig alunys zero 5 sHattodta, suffix nmean rth derivative | stifting ais [2 xe06-D6 5 i annsen-¥] Sampling» (Camgia-anat =, 20 § xn tena) = 2000) ° nee Scaling, gate et) Sw) ¢ $i) en | Time Reversal Set) = SU) Cow zevengo] | Ser) = SC") fedtionbis) «| sth = = dow” Sto} = vtrD-YEml) vnitimpoise & 7 unitstep function | f’6t-t8U = vl) ¥ 6{nw) = Ufn) feo a fonction of 2) & sutat)2 1 fieoars@ra)) wibere xi 9 Tents of 9029 C4 aontrwonstime DT Discrete tine pn2 7 NN wet i a0 vit 1 0; to : 10 _(t +20 ver{i 3 £40 Avitseee ube dult) | lp 20 i aso ¢a i 20 ate o 3440 Pte sa) = Pott) L f 0 jemi) Ito 4 4 4 hh Sanlt)= | 1; 440 sgnit)= ve) + 4-veN} sgntt) = Ut= vet) edd function vey = 2 [it samt) 15 yee 0} othewwise lei(t)or a(2)= = alll 5 AK % Z f) [ ' 1H — * CT utt also called Heaviside Unit fl): ute) =o = (Paver Be swe ftupdt =(* (stave on, pen {? :m20 err on 5 no + Sante} OF ne 4, 1:5m20 un): 1 ono - amibt 2a oy The unitstep-function is not (ontinvovs 4 differentiable at 0. ‘a rotherratically, avg-valve,vlod=—4 1 jt ae on / 300) soa {? pn20 1 mo 7 + -an)= nvr) bl iti clin fein) 0 nso «+ poms Ta. rote) 2 z=Teor gin i en Qs sinc funchor Tris compressed version of salt)- ica eas di in The sinc & sinc? functions ave defined interms of an independent variable t+ ‘sine(t=0 + galt) . orto « hnris function hos important vole in probability : & communication theory: mys B- Be e . 5 Sm gn” ayer A +|Tis sitels 0, sitn) rest, sito)=7™2 + | Ttonv only afew terms inabove equation ave needed fora good arr | pment sins sa trial $19, ; General form of complex exponential 59 alas pependingon voluesofc &a,we further classify complex expone’ + |(CRTERPORETRGTETORA + fh c 24 ove Re ; BE aims ca? jfor ann [a=€ 4 P=Reai] xt): ceet ( a rowing exponentiaisg — * 04a 3 deen exon f ic + ard wing exPOr . ve > decaying exponential Sid * vo, 3 Frouing treating exponent =O > const sig + aél 3} growing alternaxing exponential 5:.| Sinusoidal Sigrals= | ; . ais Asin(aento), NEZ"M miner Sexes Asin wats), T* FF DU ams Asin( a wre complex. j ath C&A ve OMAK. yin ; ; cele CzIehel® azorjwo ingeaonte) aed sur) = {cl fal ws cjg pdt UOT “AGM [el a” feos (aang) +3sin(aantg] xt)? yqert {costuait 09 t3sincwot +81] + 4002 simaldl sequence x growing expo 570 3 ghnusoido)sig multiptied by Towing E1403 sinusoidal sequence x dteaying.exP2. jo 40 3 sinvsoidalsig routtiplied bydecaying expoClsveal & 3 iS purely (Cr exponential (Dt exponential [Distinct signals for any valve of wp « Identical sigrals for value of Mo seperated by multiple of a7. *| Periodic for ony valve of wy + Periodic only if 20= a0 far some aivrof integers. N>0 a m- Fundamental Freq; wp + fader r2e [Hem dove lThe terms continuovstime & piscrete time ave eva ines if tis Le continuous variable: But if is adiscrete vavlable that is att is defined at discrete time, then xi) iS OTS: ‘A TS xfn) nay be obtained by sampling a, CTS xtt): 2. [army &. Dl Signals $- ———— ithe terms andiog & vigital ave Tfa Ts xtt) cantake on| |-co & DmYde te,then CTS xIt) 1S cal Ha. OTS x10) con take on only finite no-oT distinc vale, then this DTS 2") is called acca : four combinatlons ave possible: + @nalog, CTS » analog DTS » digital cts + digital OTS Concept of continuous time & analog ave. notsame- concert of discrete time & digital are not same: ralog sa is not necessarily ct and pigital sig is not necessarily diseretetime. ‘Analog 6/g arn be converted into digital one using quantization. at) ‘thevector of signals is termed as mvitichannel gig. xH)= i; #. [Ex: o1pof glead Ecc is 3-channel $/9: {otdemensiora) signa) 11 asignal is functionof move than one independent variable. JEX: + Picture > 2D + Bock Quhite TY picture 4 3D Tagg? imiensityor oone TUE) = | Toeeeno 4.) Tyvecty.t). ith Teas —furdayrental peoried 1) The veciprocalof fundamental pevied i called the furdemental frequency A perlodic signal is one which «i ‘ereats over over tii contains all the time (010 +=) ETS avis aid tobe geviodic with period T if there isa eve non-zero vue! i for which T= peviod of the signal & @ve const: o Thismeans Jf condition is satisfied for t=T, then itis also satisfied for 7=270 3 + defines the duration of one complete cycle of =tt)+ i evetasting nature eg. of continuous-time periadic. signal: ADTS Sequence xtn) is periodic with period N Af there is a eve integey N for which, Any sla) 24t) [or x(n] for which vo value of + for x) satis the condition of periodicity Vs catled, Fundamental period: TS, oT, 2) KeTt [1a] he fondcmnental period Tp of st) & N of stn isthe smariest ve integer forwhich above equation holds good: Tee ame ua xa FETS fe MADD A FHT. Tdentitythe \endividuol time periods,( 1,,Tar--- fOr CTS @ NuoNa.--- for DTS] Toke the ratio of theiv respective periods, [te Tn... for cts @ MMe for 075] rats ove rational, signal is peed: (aime tvein ons)? li) Sf votios ave inrational, signal is aperiodic: Ihe overall fundamental time pevlod, for C8) foros, lsum of two cT periodic signals may or may not be periodic \ in case of cT,al| sinvso oveperlodic but not ali oT smuscidal sigs ave periodic \same is the case of Euley function ¢™ g ¢*”. lAdditionof pc component does not affect the timepeviad of signal: Avguatut of the signal , here 2crLt)> one pevisdof xtt) T> period of cut)Note ~~ vy a oR oR tii] xeotey) 2 E ‘mereases Energy RMs value of Acoswttd) is A H a With offset ams vive of vt AcosCut +) is ait # ‘ Thesum of two ov move arbitrary ginusoldsis periodic under certain conditions. {tordition: i the ratio of the time periods of the sinvsolds is a vationa| numbex)- ENERGY & POWER SIOINALS 6° Energyof signal: el for OTS, x Ag: power of sgn: brs Ea pare veal & @ve quantities, that is why Formvlaincudes modulus: A signa is vefevred to as on Eni signal if only if the total energyis finite. te. ftetm ko. 0] IA signal Is Teferved t0 as an Power signal if aonly if the power at thesignal is finite te signal having 0 duration with Feit yolve —¥ non-zero const: as *>% and signals which ave perlodic ave in general power signa): All perlodlc signals ave but not vice versa a gait ration cgnals & Goode signals ave Eigysiaasbut rot vice-versa A signal is referred 10 as gyal if both enegy & power lof signal are equal 40 «lag ¢>40, if amplitude > co, it is ‘neither energy nor power signal Power & Energy signals ave ‘mutually excusive teans that a signal can not both be an lenergy signal & 0 power signal a a fine. a proctiat signals have finite enerles ad therefore these arve energy signals. physical generation of power signal isimpossible in practice becoz it vequives lo duration & energy: sum af energy signal a power signal is power signal rime shifting tne inversion does not affect signal energy only timescaling affects the enev94- ont : aol “7 ferarl 9 time compression bf ax) ER je tir aor) 2) redoees Eneray For a4 time expansion yal lrwer is not affected by time scaling, vevensal or shiffting-lao —‘“—sSSSCSCSSS “| A Sanat aut) (c15) ov x(n] (OTS) is am evensignal if A signal att) o x(n) is veferted 40 as an oddsignal if aioe Me syrometric obout the verticalaxls and Signals ave antisymmetric obout the vertialaxis. | a ee era ae ‘ fee eee thE even signals is atimes its oneside avea. and ev the odd signals is zero. Foy Signal con be expressed as sum of its even ddd ps, =0= tty tell) & $feromer wa ax a * Ere conjugate ov conjugate symmetric 3 if O48 coniugate or conjugate antisymmetric > if 1] mM erme cen) [RENE | re: «evenneven = eddeld= evensI9 ‘Even xW+2C0— [ap_axeny | * Even sig x odd sig = odd 1g, Reol | part 2 TTR | # odd sig + 08dsig = odd sig Yalved | oad MOKED acpap-aeny | M99 + even sig = det evensig = even Sg | SEE eds evensg det oddta = neither evensig bl : iy ner etd S19 Ben PCD [RETIN |* EE Time scaling & time reversal ave commutative. There \s vo effect of scaling on vtt)- The trme scaling on it) will result into magnitude sealing as cat) art!)The sig pt) isobtained from amplitude sealingof xt) 1s definelag azeongt 2! > Shows 4433 shows Signal gtt) /(r \sobtained by addition(gubtvaction of 24) [09 & xaW)/2ahrD Similar to signal additlon &subtiaction anysigna) can be decomposed interms of Gybitvaty signal: ) &xait) ave two CTS Then sig Ott) isobtamed by multiplication of these two, ath ae Let aut perivativeof 0) wrt time 1s defined gs the integral of xtt) Wt time is defined a5, ] UH = StH) 3512) +28CE-3) omn= (ixwat UH)= ut) -3ult-2)4 203) Ip® wi) BE] guys tht ie dies sxse| op 4s A system \s 0 quemtity which rmaps a set of dip sig toasttofop sig, (ex : ' Tis defined as an entity that manipulates one more sig fo accomplish a function, thereby yeilding new signals (oR) i System may algo be defined as set of ermers which produces expec olp with eveiable dp ex: evecrica yctem, mechanical system, etromechanical sistem eto A Continuous: time CcT) system Isone Muhich CT dip sjgsare transformed into CTOIP Si9S- yt = T x00} A iscrete time cpr) systern is one which transform discxete time dip sig into discrete time ofp sig, ytnj= T Lem} , Move over, a. CT sig can be processed bya OT system: ontheother hand processing lof DT sig bya CT System \s also possible becor DT Systems have Several ISlgnificant advantage over CT systems. ‘vihen system satisfy principle of Superposition & bomagene ity, itis called tinea system else, non-linear system: oe BOF Mr) & AUH—y Yale DT sistem IF tit) 9 Yt) & XL —> URlt) > CT system then TU) Fatt) > YD YAlt) CT SYS Bild 4A) —> YCrD + Yale) 4 OT SYS Qzxitty+ Dalt) —9 aL t bUalt) acTsys axiCn) + bx > ayn) + bYaCn) 4 OT sys- Non-lineor operators vi2- sin, os,tan, sini! og! 4avi!exp,tog, | |, medic, square, cute, NZ J, UC), S9nC), Sat), sinec), etc either on x oy makes system pon-lineay- lex: giv axth+b :additlonofconst, division yen) = wo A sy61s said to be time Tnveriant if a time delay or fimeadwance of the Lipsig lead to lan identical time shift inthe olp Sla> iF xt —yyud) then [zttats) —> Ulsts te) 7 A system is said to be ime variant if the characteristics of a system changes with tim. delay dip by te, wit) = 2xcatott delay olp by to, yet) = yte-do)= (do) x(d-todt 4 # HIF) Hence, sys.is time yaniantA,| Stic rim system = Ex} static ys: yued= £200 & ytt) ssingamn3 dmamicsys: yun= (Maer & y= 2 S. . initgiterf bounded dip results im bounded ojp.the sys-is said to bestabe «| marhemattically,. if, For alt for when, forat Ip B. & [Dwertible: yinzaat) Non-inwertible + yt} ro typ xU)=) 9" YD=0 * | system having Memory ave called as dynamic system: A system is causal if the response depends on present Lip, past sip but not on future Jip. Cavsal systems ave ponscrnticipative and Physically realizable . Response of a non-causal system also depends on future Lip along present Apast ip |All memory legs systems are causal (since the olp responds only fe the current Valve of thedip) but not vice-versa: Mathematically, for causal system, *| A system is static if the present olp depends on present Jip only put not on post or future Lip. * [In general. system without memory ave calledas static or memoryless system. * Jan differential terms ave corresponding to energy storing elements so they Jave dynarnic. Then system is Said to stable. [systems for which time scaled time veversed ,time Shifted forms of the Jip ldefining the Tes ponse ave generally stable. A consequence of the homogeneity of lineay system is that a zevo Jip yield a ZeTO O}p- Stable sys: Ywn= e Unstable sys; yi) txt4) & yu) = (taoat yu) <2) dips xi eure) (Geunded dp) YWd=d x0) dp xen= et) +2 oonded ve op > yin =e" (Gourded e1P) oP. ye) = 80 6 (ube [2nvertbe & Noninertible Systems > A system is said tobe mvertible if the dip of sys. canbe recovered from the ofp- eal A system i$. said to be invertible if distinct Jies lead to distinct olp.te two different Liles for a. given sys, should not produce Same ol?» / TF a sys: is invertible then an Inverse syctem exists ‘That, when cascaded with the, original sys. yeilds anojp, equal 40 the sip 40 the first sys gine TE], zey27 Py) uh fii] a [yy 20g ul {Mer Yr faunjy = x LUD = VU) > yet = UU) & AU)=-VID Yu's -UUA) We} y)=0 ereSteer. steel uy) te) xm) : ain 3 xed vwtoy= yond-yeo) warience axe Known OS 0 may Feasons + e approximated with these Syslems with two basle properties vit linearity ond ti LIL systems. These systems ave preferred beccz of tw Tony physical processes though Not absolutely LTE can be . | properties: The mathematical analysis becomes easier: ithe LTE Sys- ave always considered ust lrmpols ‘is impulse wesponse- response te Jip is impulse, the olp w= THO} xu) ih nis T4600} eqral ov the superposition integra, - of its resi it impulse 1 convolution int rsepresent a CT LTT gysteminterms Tt ig represented as | method methods to perform cr convolution: a. Graphical method _b- Analytical obtain the limits of yt? __. [+ eommotative property: D4 BD = welt) somof lower limits £34 sumof upperlimits pistibutive property: ctefAuh that) =Xk0e nwt change the axig frem 210% associative property: oars oud Ne POPES: +r utr system Is memory es sh) £0 for £40 tp (ERAT) « line aip of a causal sys-depends onlyor ‘the present & past values of the dip: « [the cmvolution sum, y= z bid xn-¥) + bro xts) + gE pod xtnai} ea I fora causal sys TK 40 future dip present ip fest diP «Jasystem is hemoyy less fits ofp at anytime depends only onthe yalue xin) + lp > 2 Sp vm magnitude, [pra}gmcsnite) form Gor: [eA | Zanoxtnd tapas € Innat [xtmaal = E,!n0 hee eo , / «+ [r¢ impulse vesponse 1S absolutely sane nn yon Is bounded in Tagnitude: Hence system ig stable. this conditl sulficiant t» guarentee jhe stabilityof DT LT system. =| cT UTI system is stable if jawlemavt then ays xt) SE. Ww yl) app of cascaded sys Is hme +l For cTUTESysinve’ : aut) [riety gs | 3 (siamo 80 ogThe unit Step Yesponse sit) ovS(n) Is vesponse — a y=6t") (OVS ponding to unitstepIP xu)=UU) oF an=6r] XIm)= bn} : Sit Ut) xh str= UO) ASH) = duu) = Gl = hl) at Aa nit) GIt)xhtt)= hi IS0, ja yids 2b ahh) = £ atv HaADde © yin XHDESI = =frosende efourier sevies is an approximated process where any general (periodic ox aperiodic) enn expressed as sumof harmonically related sinusolds- (oA) * foutter series is an approximated , pracess by Which any non-standard ig 1s omvertedinto a, standard sg. (09) ae 7 mot by using 0 non-sinusoidal periodic furetion embe expressed os am @ SPO", Kini finns eration ITt gives vg frequency domain vepres 7 . when The approximation of agiven function by F-Sgives_a smooth function even Whi ‘the function being approximated hag dis continurttes: | % Continuous Time Footer Series (TFS) piffevent forms + 4. Trigononetvic FS 2. felax form 3, exponential FS Series Type Coefti 7 Gomversion A periodic 4g xtt) can be expre asco sumof sine or cine font ° ave mnlegral muttipleof wor A One cnten Tao: Fundamental trequercy Javernox-amplitude of nih bn = 4(tn- Cn) nwo! nth harmonic of Wo horonic component: fo eT = tag! (BP tvs wt) Las F dyAn= Qn- ibn On&bn : dn & Oy are alsorea) ‘dn > magnitude spectrum n= oo aan (Paxundt dan = dnt [Congugate] ant3bn 2 6n9 phase spectrum E z ion FS. # ( RRUTIETORREEOIOE Gms dovdb Nes BE 49 ln z = fl B2) = tn" (ba)=-6y @L Cay? On “We The mognitede spectrum of EFS is even symmetric & Pbase spectium is oad symmetric. [Therefove, cy ig even conjugate or conjugate syrmmeitic-Exponential €6 |s more compact sinusoids: int a LTT gye-regponge to exponential fg i simpler than sys-vesponse *0 Mathematical rraniputation 1s. much easier: ere apenas ; The knowledge of symmetry conditions vesutt in veduced calculations: ition: b=-*(24%) if n= even: Inf n=0dd? aye e fxtnodounndt bon=£ fansininadit Jon will consist of condition; atty= xe) londition: xuy=-2t-t) My: oye 2(% ena | 00: [HED an! G0} bot bne$ ay smbawndt |" FS expansion of odd periodic Fs exponsionof anevenr (9 containg : pevidic sg Contains nig a x r + ee [THA * FS expansion of an odd & half wave symmetry contains only ddd sine tevms- 8 expansion of even & half wave syometry containg only odd cosine terms: the displacement of a cig on value axis does not change harmonic content of the sa, only de value of the sig changes: : The spectrum of Fouriex sevies is always discrete. uy = 2(t-Ta) > even hormonics exist. Qt One $ (wrosnuetdt Lf. slg wit) satisfies cevtain conditions. its Fouriey Series is qurvanteed to converge point wise ak all points where xtt) is Comtinudus: These Conditions are knowNas __ |ivichlet conditions, — (0 |The sig xed) ost bi over the vange of time period, (i [1 should have fini ‘s inoneperiod. aes] lit shoud have jnimna none peviod or within ahy finite ntervalof- #Joondition ch is. net ¢ sufliclant conditions: (MOTE | revodcity isnot necessary onditlon forthe convergence of Foumer Series & Fouriev form Icornplex exponential & sinusoidal ave acting like eigen function of uTI system ls the €S coefficient of jp zt) 1s Gy then the yy) yt) les coefficient of ofp ytt) will aty= F cq eM then op nea therefore, the Fovtier series efficient « the olp OF LTT sys: is given ag Gy} (nuts)‘ TF the response of osystem for an dip sg is same as the dip sg exert for 0 . Scale tnultiple,then this Jip sig is called Elgen function of the system: : A phasor sig econ be defined as igenfuncin ofany genera TZ YS. ZF the Lipof a sys: ig Ynterms of phasors, then the ~respense evaluation can be simplified from a nore difficott convolution to simpler snvitiplication: this is the basic principle of Fourier Series: Linearity pare gyt) | AGn+ Bdn [urenossmeT] [Time shifting aot) imurte cy ie Frequency shift amet Cram Lhad sealing utat) 5 490 G [teied=%6] (Time Reversal ed) ¢ ‘a . “1 conjugation a(t) é bifferentlationin aa) g gtx) “? oN Fe at “ak (3m) Gy & (3TH) Cy cae t ' : ' Finitevaloed& ch . ‘ Integration intime| (p xnae ita] zm [Periodic convolution] sot) xytt) = {zeuaren T Gdn Imuiti plication lt)ytt) F cn dnem me fry of veal ous a xt) Is teal Cok Anlelnls Len eben Pavseval's 4 (Tranfat Zaha? power Theorem : - yee cna FS coefficients of x10) 9 £5 coeftient of yUt) Real (neither [enerally (omplé : Beer <6 |eenna ad) [ore ch fanronto foie temg i Real A Symmetric (c5)<> Real Even fcvein nate} ee oetenn Acne oe o> OS a TaginaTy Qe lonlg gine ant-Symmneice> Jing ltoddinnatore) | bn #0 tems (eas) palwave | n=0 eat ane ealzeven <> Reaidéven | stmmetty [for nseven oat cosine terms Real &odd <> Jmggodd [Even & Half Cne Real even Se only odd cosine sang even <> Jig & even |Ua¥e SYAMENY Jen= 0 n= even nen de hing &odd <> Realacad [ood & Aa — en=Imglodd Yay, only od sine juavesymmetty |q203 aeven meas ee | etms' there ig mo convergence issue with DTFS: Discrete Peri FS coefficient Cy? orFg aU) > Cy cyis peviedic with n ee Cy Cuan in vepeats every an interval dlong.A sale. cy Axim} + ByCn) un tN) Linearity «|Time shifting ey stitting | GSP) aru + Joonjugation Eo * [Time reversal aon " _fxfB + [Time scaling sf Oe + |mottiplication’ |” = gh + |fve difference | 2-xEn- «_|Bvaning som/ 2 x Accumplation Line et 5 xfxJ Yr) (oR) EY) + | Periodic. omiojution II ; hears Periodic signals are vepresented as sum of complex exponentials. baneitas diary time fourier ‘yepresentation of periodic Sig takes aform of © Series, ete time fourier vepresentation Is of Finite series as a result of this, Fourier serles (DFS) representation of a per & fordamenial frequency -2o= 47 1s given by, Fourier series representation of discrete & periodics/g is also discrete & peri fourier series for discrete time periodic sig is a. finite sum defined signal itself over one period, the series always comverges- + ferladic with N jodic Sig xt) with fundamental iedlc enfiraly by thevalve of ; 3) ko + xl qoute eC eevitclth eiedicwithna \ Fo ies of Common $ Imsoz| + L Squove wwe fon evenn 2. Sawtooth A Ja V's 2 ann i Triemgulay A ot Gnz0 wave ae — ven —- 4. Full wave aA aA fore rectified 7 nO-4™) for even ¢nz0 :0ddn is. Half wave A A eet vectified, 7 ar (en) are AT AT gine ( TKO & Rectangulay Ar aT wave [: { = +] © te am % calution of 1! imitation is Fourie txfor thal is arglale for aperiodic also: olution of 2rd (imitations Laplace txform by using ett. 7 “A pImpube |... | wee au a i tin | | %. te “eo ow Limitation of Fourier Series &- orn _—_ Tr canbe generalized onlyfor periodic dip & thos not applicable generally for aperiodic one « Tt can not be used for unstable or even marginally stable system:) Mi) +@) +o £ £ * | Fuurley Transfer is extension of Fourier series for apeviod! ans FORM (cr FT] Y Time domain Sige ic signals Fourier Tvansferm \s applicable for both aperiodic & peviodle signals: ConrinvousTtine fourter Tr Fourier Transform provides a frequency domain description of estan | he fourier Transform: = [€T(xity] = x(w) = ( xuye” dt | continems aren deo aperiodic ontinvous F Out, KEnverse fourier Transform: (TFT = &F (wur] = 20) = é { awe aul IInterms of Frequency'$' Hz: = oa xW)= (ran eat @ ue fare dt sig xt) Should be absolutely imtegvable ive: 1 (Prenat o> then [x(w)] <2 signa) xt» should be deterministic over any finite interval Irt should have firite no: of rnavima & minima. over a. finite interval: Tr should have finite no-of finite size discontinuities over finite intervals the conditions ave sufficient not necessary: It means that there 1s class of sig that violate either one or both conditions, possess fourier transform: lex: power sig, periodic sig etc. a Fourier txfovm \s defined for allstable sigs, {/ wldt {@ Periodic Sig which ave neither absolutely integrable nor squave integral over a” hntinite interval cab be considered to have Fourie txform if impulse fonctionsare nals [permitted inthe tiform: + [fourier ‘Transform Fourier Ty form Pairs xh) x) x(w) wth) XG) XW) s i * L angs) | andtia) reat sinc) sine($h) | haat angie). & | ans(us) fie sinc’) | (4p) “| gy” ten i sindt) | vet | rect on) | eostanmpon | _atsans_ | _ad_| cosuot [Gus e4tstoV2|mtataaapedtoe ansantye ones (gen) + Gna Sinwgt fase) 86H ef Ffarony-s010 Esintenpyow| _amB__. anp sat, i ele Sant) oamp) [So Heng) . vw | Sam a ee (hy samp | (30)? Harpy sé) | _. D aWeL Gen) 1S ey-x ang, Taian? | Gay ch Fests) Fhe) eat aa aa reviele ig ayant | aR tut Ty (ate) =sh “; en em even aft: ee 1), ined Paean 2 a. Bed ay > EES ara sontt) yt ye fnaneste-H: =k PpaeMyet ot on onsslst samy | UMM Sa gee an - . ston) Sa (ar - 28) an 2 gaye dat ae LES amptw-wa | Esewn 42 6-4) EF be *!)sama smamn » Properties of Fourier Transform >. PROPERTY xt) x) Linearity, aalt)+ pxatt) aA) + Bead) vite) Ral) Time gcath + x(4 " Ua) a x(f) 1A) (3) Time shifting | xcasay oH yy) ow) Frequency shitting | ain @? #4 Xtts4) x(wz4md) Time Reversal xt) xe) x(-w) comvolution SU HUN), x) AU) x9) HWW) multiplication | x0 hit) XH" HY) 3, (xu Hw] modvlation att) (og(2nat) 4 [xgaa)+ xG-A))_ | 4 [awsard) ax(u-mm)] derivative xing) Jans x) uxt) freq differentiation - yanz-ctt) 4 xq) = x) arrx'(w) tepation (fund wo + just fy) s HOB) congugation ott) xtet) xe) correation siege =n ged xu)" xwy"w) Autocorrelation sup xxntt) me xaytyy = pl? | xtwoxtea)= ol Duality 2a) > XG) 2) > x(W) x= 3, {xu eu (Similarity) XY x) XP eo anxtw) dF acted, {en 5 dy aa 2% arate) = Pe so [anZCO = FTERI] nal AreaundeT Hor: a= (Tnat = Lf xwewe x00 we sya? Ania under 228): X00) = |? xinat [oitbsphenomeror] Parseval’s This sig energy conbe velated to its Fourier spectrum Xiu) as~ Theovern es (Pputat = (iaotay = 2 (2 xwoPdw + |Plancherals © cout dt = (xy = ayy logtwdt = (xarytwag = 2 (xe) yw dw theorem L Srnrinas = 3,60 Signal type «£25 Fourier Tyforn TRICK: 1. Reals Rays att) E> x7) =xC0) Real eer Cr {conjvgate oyrmmeti ed os oad 2 Real 4 EVEN C5 yu)2yhn) 2x10) XUEAC-H= XW (Rea) + even) Imaginary odd odd), even ren [0 3. Real+odd 4 mgt odd BwWsxetsextt) AW)Important Resvtts: ! wi, ie 1s WET, angus | id veetety AL gine (2) 1 ) Sn) = z e Fr 1 Langu) Slates Ly gt sine.) Bs arrrectio) i) fr 7 Samit) ae Tro? a fi EE twl “+ yew) A Ae sanw ah to né mo are uw) vip ven Es newyrd, (wild ostmyt efor [otwenayr6lw-woI) gt) Fr. Tim vw) Sld-de) 4 6(tnte) AfE> acoso tb PEAY to c arsine (SE) arect() 0 Aarrvect (2) Ww ssrect eevenfunct” Aarrrect (Ye) * fe % a) Way * eal pos (Alas aa » “U2 >” ee at sine |) %, Z wi yro)= XWHUW) ac [Fr x10) >) a a x S net YA, ? OB Frat an ae + |shiftingin time domain only changes the phase spectrumof signal by -Wts whichis linear function of w- [é¢- Unear phase]. ont oe at tee Frequency shifting: moltiplication of si by € ‘rifts the spectnvm of x¢t) by w inright direct”? |Scaling property: compression imtime domain teads to extension in freq. domain & vice-versa: Reversal property: reversal of time also tever se its fourier tx form - tt isalso called as ‘inversion property of time & frequency. I Fw]*is Energy spectral Density of SI8MAl for energy sig. rime diflerentiation property : only valid if 470 as j>o 94) must satisty pivichltt’s condition (& [28% | dt com nerefore nd one teone relationship bla We SU"? | w iy [ai witl destroy the De component of the orisinal spectrum 1mioz! + in |The phase vesponse LHIw) must be Fourier Tromsfovmn of Periodic Signal 8- Drew e__---e—— 2 nun A perledic 61g xct) canbe Yepresented in Fourier Sevies as xine 2 ind ae e rons [Eco] = Fenlangqwenwy) 3 [x(a an Ey 6(w-m0) ie & bal |Thus,the Fr of periodic sigis a impulse train having strength of amrcr & xatedat WEniy, Fourier Transformn of Tmnpulse Traing- xm) A period impulse train with fundamental period uth. Ts is giv = Fete Wye +t ig IS GIVEN AS, xit) Zoe Ts) 5 Wo’ a faye consider, for onetime period, atts sit) 3 Tp4teh Games . FScetficient, cned (ene ™™y 2% gery ome = L mek Ly, wer dt = 4 (ee, wear = £ Frof periodic sig, mta)= an ¥ cy gqw-nuy) = a £ % 6lw-nU) = F own) ne neo 'S no = Froth * x00) KK(w)= Wo F Stw-nue) Thos Prof impulse train oy bed \galgo aimpietvain. — --- ne ae ee Distortionless Transmission [aertication of Fr e~ : Distortions ITrangmission through a ur system is said to be, a) wins Weta-te) dlctortioniess if he olp is xepica of Lp with saling 2% a 7 lin amplitude & possible time shi{t+ ; ee oe dain cnition fo astriness rmision: (= WEN] Seay yan fokirg Fr yc) = Kern) 4 29, = How) 16?" [Ht] 22 ¥™ » Freq, Response of tts lequivalent condition in freq: domain: non The magritude vesPonse |H(W)! myst e be const. [THOl= k= const] “~ > linear function of frequency. (agnitude Response) for 1 Tinea function of w, but it should also pass through wo. I1n several application sych as signal arnplification or sq sig temission over communication, Icharme|, we vequive thet the olp waveform be 4 veplica of djp waveform: 2paramedersassatiated with Hlw) + phase delay & group delay: Lx¢ amplitude gpectworn's not const, then that distortion is Known as amplitude distort Xf phase spectsum is not linear fi vunction of f1eq: then that is known ag phase. disttntion of delay distortion: For linear phase sys, both the phase dernge the grovP delay should be const. Is otw)= Zita) =-wty, phrase delay tp ~StW) = at, groupdelay dy=-40u) - 4, js tp: tgs to =const, 7 Ww+ oe: . = 1. HT does not change the domain of asa, [impulse 7 UF . sponse 2 HT does not alter the amplitude spectum of a sig. Phase delay t~ The time delay experienced by single frequency sig When the Sig passes through a Sys.ig vet it isgi He ' ef erred to. as phase delay Git isgivenas [Tpw! a Grsoup delay 3- en ithe time delay experienced by group of frequency (or small band when on dup sig that contain components with different frequencies (0! elated) through a sys: is referred elay &it ls given as, passes through a sys: i Qs group delay & restore of the ‘ Tyr = =F evan] he Me grovp delay af each Freq, equals to Ove o : dw phase at that freq. : Tf glu) 15 const tithe components are delay by same interval- Tdeal & Practical Filters ¢- lan ideal filter atlow distortionless trmission of certain band of frequencits & SUPPTESS the / frencining completely: The viel be iss filter is one that allow al) components below wu to pass without ‘ distortion ond Suppress Components above w =W,- a 1 The ideal tow pass filter has linear phase of slope to which results in atime delay ofto | 40 Oll its dip components below we Tad|sec. , US Y= x(t-to) , The sig xt) is tymitted by the Sys without distortion, out with firme delay to - for this filter the freq: response Hlw) is given as, rtp rect $5.) & amie) a hw) cutuy= 1% entred at W=Wo) ' st harmonically) “f -rmpulse response of an ideal LPF. All idea! filters ave non-cavsa] unstable therefore physically unrealizable: HILBERT TRANSFORM 8- Ihe Hilbert trangforin is an operation that shifts the phase of xtt) by-M% while ‘the amplitude spectrum of the signal remain unattered » lAn ideal Hibbert trforrh cHT:) iS an all pass go" phase shifter HiT ig used in no-of applications such as reprecentationof band pass signals, phase shift modulators, genevation of SSB. . |Signdls & 1t$ HT have same Energy Spectra density or ower spectral density & Avtoconda” J HIw) Wy alt xt) rt yl) 5 vo at ry) | BF tty etZs ko then Sip aL> xt) A, xtH& ZU) ave orthogonal 40 each other:SEA # | Energy Spectra} Density (ES0) §- *| The enevay perunit bandwidth 1s Known as Esp & is denoted by ix(w)- The ED of on Energy signal His, [i,t = Ix(wnl™ The ESD is clefined for energy sig 4+ Power Spectral Density (PSD) 3 * |The power spectral density (PSD) has the same relation to power Signals AS ESD has to| energy sigs . the power spectral density of a power sig xUi5, where, Xplw) = Fr[aq(ai] & x7lt) ae 3 Yee 0 + otherwise Properties of PSD: (| Area under PSD gives total power tee e S,tw)dw tid} Ag power do so, s,tw)20 forallw- wi THis an evenfunction of w, [S,=S70) (™] sytw) is veal- HICORRELATION &- + |Tt provides a measure of the srmiiatity blw 2 wWayefosm as the functionof search parameter, lAnapplcation of correlation is signal detection in @ radav,where q signal pulse is tumitted in order to detect a suspect target: If a target is present, the pulse will be ‘reflected bya. 74 the target is not present, there will be no veflection pulse just noise: By detecting the preserce or absence of the veflected pulse we confirm the presence or absence of the target: ‘ +] tn digital commonication,tne important thing is that 2's & Os yn the data strearn be distinguishable from each other so the receiver carYProduce the bitpattemn that was tymited. on + |Ttigthe similarity OF Coherence b/w two signals ce: computing the Correlation blu the two lelgnals is to measure the degree to which the two ggs ave similar. « livig of two types : dl Auto crvéatlon function. cAck) Civ Cross Gowvelation function cece) 1.|AUTO CORRELATION FUNCTION (ACF) ¢- « ler ques covvelation of function uithits shifted version: Ace of am enewsy signa, [R= (ewaanrat = (Meare) esdinat #foatebal 2feanaceett 20 | PxeeRrine a EO} RELOAD 9 mete 2 1A] Wn . + van 9; 7 The convution of signa withits reversal glves ace, [xuoW CH= Ba AcFat origin indicates either energy o¥ powerinthe s/g- Stet xu» be energysig, [enerayinxa)= (*xuplat = Reto) do Let xi be power sig, | rowerin set) = tim {> ratet = Ral) Teo 2 (w | pourier Tnsformof Ace Is Known as Esp] PSD: Oot we have RetO= xKVAxCC) Thos, Saw) = (° ae rat , Fr {Rata = Ho)KCw) oo fs? Way if seu is eal & even xUo)= (w) a ation: This is Know? r- Khim ergpetO} = xlw)x*tw) = [snl =) inom ag wine cane Ag under PSPC Rett) 5 Solu) gu0= J, (Pura = a 2. [CROSS CORRELATION FUNCTION Cee );~ Trigthe similarity blu) signa} & shifted ve lave two energy Sigs The cross correlation bl Reyit) = fo wuld rodk = (Px arounde = 100 xYCT) exsion of other signal Suppose x&ytt) givenas ay unde if both are reals denoted as Rxgtt) 4 + [Similarly Cv0ss commelation blw yit) zu), isdenoted by Ryxtt) a givends, Byztt) = JP uurec-tdt = (Pup reyattnat = 9crdK 20) to to Properties of ccF s~ ida bas Eis . [Inge s eos | iy FFrisan even fonction. [Raut = Rao} [Inge s eos | R10) Rul) Gite fe xtt) & yer) ave Cald to orthogonal: prot: Ryyle) = Six gta nat > Rayiod= $" Ar) yundt If Reyto)=0, it glelds, Fexunytneo bee sty &yttyare orthogonal NOTE: | Time domin > __Freqvencydomim_ | Timedomain > Frequencydommain Iz. cntinvouge Periodic —> fouvierseries (eres) “| 2 continuous Aperiodic 4 ontinvouss Aperiedic > fourier Tranetotmcet)| 2, discrete —> — Periodic bees aie > pisnete Famiesseries(OFs)) 3, Periodic =—> — diserete > piscrete + Aperindic > oisete time al di ‘; Taform COTET 4 aperiodic Continuous $f [Dmitatlon of Fourier mane forma its Solution i FT 1s not defined for non-obsolutely integrable signals thesefore solution to this lieritati on is Laplace txform:% H ay ae ee ee ee Drscrere Time fourre Trans form [OT FT] The extension of DT fourier series for DTaperladic sigs ges OTFT: MThe freq: domain description of oT sig 1s continuous function of that can takeany Nave over continuous interval trom -> to +0 & itis periodic function of frequency with period a7 . The corr describes thespectrum of discrete sig & formalize that discrete sigs hove pevicdic. spectra: The freq: range for 0. discrete Signal is unique over (Tm, 7) oF (0, a7): Definition s- - Definition ¢- arn ose OTET of discrete time sig x(n], |DTFT §2fm{ = x(n) = zzime (gad) purbers) the teform xia)ig freq: domam descriptlonofxtn}- try PET x(n) , o Disrele, aperiodic” erlodic, Cortinvous hhe inverse prer which maps the frequency domamn description x¢a) back inte time is Rivenas. Prorergxcay}= 2) = 2, fA IP gn NOTE: | notation of orFr of xtM + xu, XO), X50), x(e4), x (ed) EF Nature of Fourier Spectrumn & Periodicity of DTFTs- (2), the 41eq.domain ‘representation is continuous function of | (a) which can take lany value over aconitnuous interval ‘rom = to ten, Also,x(=.).s periadic fonction of 2 with period of an lit follows that, [X(atam) = XC) . aren = KAI] jinn gianm Lan at nee oof: xcasam) = SP xT” _ Fain. we nee convergence (Existance) of OTFT $- Ithe pret x(a) is said toexists if C0 isabsolutely summable, 2 3 ad + as By definition, ycny= F ttm é"” with [é°"|=2, wie dbtar ron IThis is sufficiant condition for existorxe of pret representation: Thes con guarantees its uniform convergence. . . Some sequence ave not absolutely Summable, but they are square summable- forthe signa) where OTFT converges, not uniformly, the condition is, | Z Lat? Leo . nso pret of x{nJ=1 foral\n & x)= 2” exists even ‘though they violate both conditions 1a this is possible becoz §() is permitted: There ave cignals that are neither absolutely summable nov square symmable (finite energy), but still have OTET. df we allow OT FT tocontain impulses. ore r lex: Mugen akg aga RTS aDTET of some Basic Signals ¢~ w ott, . + TFT is notdefined, ay 1 Vat : OTE Is defined @ lalhL! OTFTIS (| Sod ES ae db i For oxagy athe spectrum of oTFT is conlmuous & periodic with27 [The amplitude spectrum 1x¢n)| 1S eventunction & phase spectrum £xmis odd ., {SPectum of 1+ _ aé* ) togma; ra 28 ctgn verges ES Tm Ira W)] aM 2 Mugratd™ven-) | gm QF eat lait T-eacosa. ta?” ' 1 W)}Dlserete time gate pulse: i xone{* 3 e ° ® putn+h =m 3 te MN Naa xia) ="5 ee" 5 z {e Nl any insh aap mba eM en He cay 2 eM erent |: Bem in) 7 eet) _ ay) en K(2) son (2) Let N=4 iu becomes a 9 point edt “pulse, ona) “X(a)= singsa) SMOSA) Fa ei Wo |piscrete time triangular pulse: ae lagn)= Gfna2] + A6n40) + 36m} + 260n-i) +4 [n-2) ne by Definition, yin)= ¥ apn) 2” 7 wo 2 Lill) | Unit step function ven): let urn) bea) [New gc) =vfn)-vem-) [Taking eT, pelea) Woting that (1-é)=0 for 2:0, xt) most be of the form as, x) = Agtay + LL, , | Sar ( Gla) ree I Acconst- Te determine A we proceedas Foyous. vgtd= 4+ 5) tn) = un) -Uetnd aad = F Lecnas] +9 gpm 36tr}4 28Gn-7 #60014 G9" = & cs (viiy | trpuise fonction grr: [a Bs 100) =F sinew” 2 OM yeg =1 $0 [SLAP eT hapa) 30) 3400 Jaco, 3 Sa) 9 Se, Pat RO aM Votn}= UMA] -2 9 FTf uty = Agta) _ =m)“ The FT of an odd Yea] ceqpence must be prety: Thos we must have vg 2S gpa) + Lf$f] Properties of pTFT S- : + Linearity: aautmapratny > dain) B20) @ iftime shitting: ecnanaa EES oH XE Gi) Frequency shifting: @fPactny Ey crn gto) [Iniis doa of me sifting] ww Modulation property: fm cosngn Ka? Ke(F) thus we got relatlonship blu» continuous-time frequency" continvovsstime freq: (w)= ete-time frequency (2) (Sampling intewar)(7] 2-Tt is Sarnpled e Pa, xelot) 6(2-97) a) jz X(T) Qaiscrete time frequency'a| 4 [Energy Spectral Density (ESD) §- energy pec ‘Alc to Pavseval's theorem, total energy Exinasigxt, Ex= terelates energy per onit time (txts) to energy per urit Bw (mary Ithe energy per unit BW is known as Esp & is denoted as $,(0)- the ESD of an energy Sig xn) isgiven as, [S.cay= | xca)l2 Powe’s Spectral Density (Psp) s- oun Psp hag the Game relation to power sig 4s ESD has to ‘energy $19" The PSD of a power s/9 x(rd is given as, [Sca)= xa? Relationship btw ipa olp.Psp of anurt sys: [eyca)= [near Sx) + [CORRELATION ¢- (eee 4. | Avto- Correlation Function (ACF) §- lit gues covrelation of a sequence with its shifted version. If zim) isan energy Keequence the ACF of OT sig x(n) isgiven as, |R,ud= ¥ xpnixfn-wl Fxceazon Properties of ACF 8- is tin | Ace is am even foretion of time R Rx(H) = RCD Ui} geoys ees % Com — Eneraysig 0 fat) = Ex = lim, El pools £ f Tred a oe 8 FE pny Power sig dust HW GID RCH) = aimee) Gv] peley ETS Seca); Ace Re(HIA eso mnakes FT PalsSUMMARY g~ Tt dives conrelatlon of a. sequence withshified version of anathey sequence: The cce of 1wo energy sig xtrd 4 yrn) is definedas Rayli = F xtryinend = F xCmwd seo) nee peo Its even function of Frequenty Ges [Ras] = Ryx CH Gontinvous Time CcT) Time domain Frequency domain Discrete Time (OT) . Frequency + domain Time domain Fourier ‘ : ' ‘ ‘Cominuovs Time f Discrete Freq. Discrete Time | Discrete Freqwency Aperodicintime APeriodicin freq, 4 Peviedicintime + APeviodicinfreg, periodic intime | Periodicin Freq. Series, aa d fair Transform continuoustime continuous freq. |piscretetime {continuous fred, lAperindic intime itetedicin freq,wt w Gis} wy) = = vi iii) FERRE) DISCRETE FOURIER TRANSFORM(OFT] & [EE Fast Fourteramustoam CeeTD.. ® DISCRETE FOURTER TRANSFORM (DFT) > IThe Discrete Fourier Transform (DFT) 1s basically fourier Transform of sampled Sig repeated peviodically-(oR) Sampled version of DTFT spectrum IS DFT: The no-of poims in DFT is KNOWN AS N-point DFT: Ithe fourier sevies describes periodlc signals by discrete specta.hercas i a Kdescvibes discrete sigby pevlodic spectra» As aresult sigthat ave both aise ; Periodic income domam ave alSo periodic g discsete in the other This is the basis for the formulation of the DFT Definition :- N-point DET: consider a sampled 6 periodlc sig xf, its DFT isgivenas: DET Axim} = xCK) = Fxg eye? | He oats oe GD te , = FT discrete The sigste) & xn makes DFT pair as, "(dane ) xy PE 0 (rec) : reve Inverse DET (DFT): a rey <2m_ oz perlodof sigztn) ( zm: 2 Exte n Ait Ne Pe Both xtr) & xfH] ave discrete & peviodic. DFT relates Samples of time domaln to samples of frequency domain- Nj we rr ra n] Where, nyse isthe phase factor WOO Ear My GAME EHO We . ot tude Factor eee + Symmettys wy he ay Periodicity: wh = my Properties of DFT 2- i ee Linearity $Ufornpeint sigxn a0) , zit) + Baln) PET hifi BARIKI lrimestifting: #Cntne) PEt Et x50 & 0rH) Sh a x00 (cirewlar shift] hiftin yedomain: up MO PFT, rund [aso ctted moduiation] eam SS xt) lcongugate symmetry: — ac¥fn) gPFT, y¥ Ry . For realyolued 203: xt) =xCK) & xt FH) = XO) algo Hnownas even conyugate- se xruais periodic with N, Hence, x*EH = x"[n-H] also kmawnag conjugate symmetric TF tw ig real then xMA)= X* (w-K) & bcoz of this property, we need compute only att of the DFTs fov real =ctn)- The other halt ts conjugates Time Reversal: x60) 22S xe4) wr) xtwen) «PEL xIN-H] Lineay Convolution! sind ¥ Xai) PELs wicvaraCi) with period = Nit Na-1 Periodic convolution: xan} aatn) PET» aifhdnate with period = No (sane) Pavseval’s Relation: = g= # xc? 218 lana no Nikoi Centra) ordinates: xtoae F'xgm) sxfoextnecteyy ected * xCAT§ te x(n] xed mA, HO= 4 Ev = L fran . waeven ()]pvatity Property of ver: fe) «Es uti) where nis thet per, The notation Xi) it HO) > NAC) aga function of time index: ‘computed to obtain another N-Point sequence, we get DETTE XM] = NXEMy (OD DET Sper (xtm]} = N LEM) y Method r: xt} =4x100,x10), 4027, x(3)} ‘Method D: Rio) aa X10) s apoyext} ATR} 1089) baad xa) Required tOcalculate ela XIE afo)-20134 2187-204] wi Lad 5 gin XT)= conjugate of x1 Pag] sta) 14262] = one GU) Seay = feerentn eas) -vtn 4-40] ee inderg K isthe treq-index. cates that x1K) is Craluated Asin wat iteret the duolity propesty is the fovouing Let xf devote the N-rint per of on Neroinrt Sequence xn} The DET xTK) itself Is an N-point sequence, Jf the pFTof x1 is ‘xte) x xi] xa Sy PAST FOURIER TRANSFORM (FFT) > computed Tesults- The less storage FET algorithins reduces storage ¥ Fast algorittrm seduces the problem of calculating an n-foint DFT tothat of calculating many smaller size DFTs The computation iscarried out seperately on even-indexed & ‘odd inilexed comples toveduce the computational effort: All algorithms atrocate for ‘required, themore efficient is the algorithm: Many equivements by performing computations in place by storing esutts in the same memory locations that previovsly held the data. H] 3 stages in a g-point DIT-FFT S- 4 Typleal buttevfiy for bIT- FFT Algorithm §- *) mt kio—Lr_-—jomvre My Ne, ramen baw) fxrcr—{ ber DENS. x1) 4 point }-x14) wt ombine I pers FD 8 2-sints }—] X10) as oFTs xt) stagen. ‘stage TL 3 |DET Algorithms for N= 24.0% ooo 11, aaa po x HON OKO) GEG] ' ae xt Nxt ee fas 00 fe} XI] 00 oot £10, lo xP. xD oO} yor x05} W ot xt XG) 0 oat} xT) rede we £157 Aarau Aw (To) 800 fi) ol 1) oto on KI] 100 e121 76) v0 «Tm=| OLE Algorithm for N=2,4,05 |e te, a] fo) Xe) (OS xt) Feature IN-Point DET [N-point FET ‘ sototion of N [wa Bunetitsfgege JAtgorithrm equationsin n | m stages Unknowns otabuteslies = = mottiplication | N per equation | per butterfly Addition N-1 per equation |2 pey butterfly Hotalmutiplicainy §— N2 Mog, fl ' |e Total Addition Nari) N logalta CHF] a *IThe procesg of sampling 1s a bridge blu continuous time discrete time domain+ * lsampling is a. process of converting a contingous time signal’ under certain condition the ‘ortinuaus time sig con be completely vecovered From its sampled Sequence: FF [SAMPLING THEOREM - «|accordingto sampling theerem, A bard limited sgn! xt?) of finite eneray can be completely reconstructed ffomits samples token uniformly at vate” wt? Wm “martha, Ov §,2R5m sample)sec: (oR) + [nother words, the minimum sampling rate Is, 45° 25m He. * [To Prowve Sampling Theorem, w ~+flet xt) be continuous thme finite energy sig oP loose spectrums bard ited to ly rfler —_—™ (ie. Xw)=0 for [wl >wWm )- 4 602) +] The dip xc) is sampled at the vate of fs Hz by i moltiplying xt) bya. periodic impulse tram 6,4) 7 ee with period T= Vfg Cl | The sampled sig 40 Consists of impulses spaced gy 7” every Ts second whose strength (area) is equalto at Thar instantaneous valve of Lip sig xt. ~ {iis *HTherefore, — xstn= t-br(t) ingot xy = aun: £ o(a-nte) = Exonpouento Saning of: neo om Taking Fourier txform & using multiplication propertyot FT e Huw [xws a1 f go-nas] = 7 tO z sian) vsing convolution property of impulse 0, | This result means that thespectUM xu) consists xo) lot -- xtw) tepedting periodically with period 15= 27 cee, we have axgomned gpectrurn 2) Us bardliited to th. — * MThe sampling freq: cantake 3 passible values wrt: pone Ispeewurn width (aun). _ ‘~ Fechom aa (Way QUI Uli) Wg=AWry MD We 42h CAST: | Wer Rm = Let Ws=awm,, spechumof sampled 5/9 will be, xgtw) = e Exww- anwm) We see for Ws>AWm, there ls no overlap blw shifted spectrom ofjtuy Ts a 5 long as the Sampling frequency we is greater than the twice thes}g BW (2%), xtt) anbe ‘Tecovered by passing the sampled sig xs(t) through Ideal or practical iow pass fitter having BW biw Wry & WW, Yad sec:SEW] We gum sn )No overlap blu the shifted seta of xtw): . [8 %st4 trough Ideal LPF Xew)= 2 ! 74 X(W-anU9) 4) canbe recovered by passing the sampled Ty TOR Bm Om inimum sampling Yate ws=2Um (Or 4s aim) required toxecover, xit)iscaiied Nyauist rate: sem: SM!) We aWm'- — xslw)= & E x(w-non) ern ne © [There is overlap byw shifted spechum of x(w)- . -onsequently the sig xtt) Cannot be exactly secovered — E 0 Wm 20) from its sampled sig. J “even a * JAliasing o+ Spectrum Folding $- > !The overlap in shifted spect of original 51g is termed as aliasing: > [Inother words," The phenomenon of bigh frequency, components taken on the identity of low frequency component is Known as aliasing: 9) To avoid allasing, thecondition is- (Ws >2Wm Sfaliosing is an ivreversible process: once atlasing has occured then sig a@nnot be [eecovered back: for D> _[imporfant Resorts Ny mmina eo ciara [PORTER] comments Bard Pass Signals 5 sw | 8R=RUb | Squaveng doubles the HA. A Band Pass sigxxtt) with bandwidth 8 xttey [NRF We | Hoetfect of Himeshiftingon HR and an upper cutoff Frequency fy ca xean | NRE Re[Timecompression changes nA be recovered fromits sampled version xt) | NRFOH2 |ime expansion chemges nA “by Bad Pass iterng i Fe we nn=w — |efterentiation mero onne |” il were, [Ret (5) Fh fund | MAW. otegaton bas noetfect on HR Ae siege et z . toxmt) | NR=We |oovedtion uithitselt cbes not ctamge NR # [SAMPLING TECHNIQUES &- 2. - |Tdeal Sampling 8 (Instantaneovs ov Impulse Train Sampling ) sling sampling signal is periodic ae cht: recess impulse train: The area of each impuse In Fs sampled signal Is equal to r instantaneous valve of an Alp signal tt): tl Tt ik he Practical Sampling ¢- tt - Lap Reality, Sompling is done by using a pulse train in two di Herent ways. Cd Natural Sampling Gb Flap Samgling_-_L i. fl ee W |Naturay Sampling (Analytal Rtalization) $ + Jthi i 7 . 6 the one easy to Yealize Tratbe matically but velatvely ditficutt 40 ae) Sp ‘implement mactivatly. Fat oO * |Ihe vesuiting Sampled sig will {] — Sustain the natvrat variation “Py ofthe sig inthe pulse duration ee AA [& hence the name natural I A Sampling. : 1 | ation someting nati seaiaton : This is difficult 10 realize mathematically but velatively easy to implement practically. [The resultant sampled sig wit! hold the instantaneous values of the sig for the entire pulse evration: fog HOt re LA: ttre AA AAA ® Py “ | Dawn ~~ [ ae — Lt zy a -. ” == e °° = ees ale aU : . 3 id 4 ; mn Aperture Effect s- “ted oO effect +JAs seen in above spectrum, the higher 179: Ant fl 0 a 4 components ave experiencing distortion in comparison to the oviginal spectium » This Is called as Aperture effect *+| To decrease the aperture effect , cave must be taken to consider the pulse witha narrowpulse width */As pulbe width & isincreased, the widthot rstlobe ty Nels [esi ga] lof spectrum (4/t) 1S decreased. Hence 7 Jincreased pulse width in the flat lop sampling, teadstoatlenuation of high frequencies in veered ucion}. 4 |Reconstvuction of Signal 3- The reconstruction is accomplished by passing the sampled 1g x4lt) though an ideal low pass filtes (LPF) with cut-off frequency fc=Sm° no ofa io) uN -Cll ? “ Mr 7 > som © my since ideat LPF is not practically realizable so we use practical LeFwhose response given as- A For such type of practical) LPF, we cannot use critical sampling |ls=2um) but we have to Se over Sampling & seperation inover fi \ sampling is decided by the transition band of LPF: This seperation — 7 is also colled Grvard Gord Tone end it loutput of Sampler = Lip to LPF = nfgt4m) n=0.1, For practical LPF: (Gard Gand > transition Band] [4s = 24m4 G8 > Afmt 7B trin-vavve of sampling frequency :