KEMBAR78
DFT Notes With Examples | PDF
Scribd
Upload
63 views32 pages

DFT Notes With Examples

Uploaded by

4zfq8g84rk
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
63 views32 pages

DFT Notes With Examples

Uploaded by

4zfq8g84rk
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF or read online on Scribd
You are on page 1/ 32
MODULE- 1.3 DISCRETE FOURLER TRANSFORMS (DFT) To pexforsen preqnency analysis en a discrete time Sequence alm), we convent time -domalu sequence fo an equivatomt frequency domain represemtation. eplyiong Fourier Tramforrm om x(n), we qe x(oo) , Which is Comtinuoun And periodic Aumction of frequeny dt is mot a computationally comvemient ye presentation of the Sequence x(n). Representation o| a Sequenct stn) by samples Of tts } 4 | pe Spectrum X(w) is Know % the Discrete Fourier Tramsjorem (D PT ). FREQUENCY. DOMALN SAMPLING §— W.K.T aperiodic ferite energy signals hase Comtimuew spectta. Fourier Tromsfoven (F1) nf Aperiodie DIS sem) is given by @ . Xw)= AZ xme" —_@ M= -00 Since X(w) is periodic, with perioet on , oenty Sampler in Ma fardarmemtak frequemey vomae are obtcumed by sarmpbug periodically im frequemey at 4 a a spacing of Sw yadans between successive sacmples. We take N equidistont sacmplea im the interval 02we2o0T with 3 oclug Sw = 27, P N { 1 1 1 | ! | I ' I roy ed ee | 9 eee, > Substituting, ws atk in a @, @ —jon oT ee J kn x (2x) = Zz a(n) @ oN —o@ M=- 00 whene k= 0)1;2);--.,No4 Summation im equ cam be subdivided iat iengjenite mnucmloan 4 sumamatievs, where ede sum Comtalms N_ ferms. ee kn a x(He) = ot Z use N n=-N as vem kn Z a) elke * 2 x(n) e tse 400 IN+N-4 —j2 em = Ey ott =-~@ n=IN Th we change the index im the immer summation from ~ to n-LN amd imterchanoe the order of Summation, Wwe eet, N-l oe : x (ate) = S a(n-J2N) eon n=0} J=-00 Poy 20) 1,2) ...,N-4- -j jam Put eee -aNn) = Suet Ne ee jopedy wom =4 whew e@ Hd +o ott ° _ Joe eo x(2te) = 2 Z x(n a9| ene” aco L t=-@ L—6 Let xp(n) = eZ xtm-dn) —®& =~ 00 xplm) is obtalued by the periodic repetition of s(n) eveny N sowmples. dt is periodic voit fumdormentort period N. sxpln) com hence be expamoad im a Fourier Series Nol . + 42m winh= AOE e wEM __6) k=0 N= 0)1,2,.-.,N-t as with Fourter coef pices, = eu = + z xp(m) © wen —6 N=0 Ke Optpee.y Not Comparing equs © «©, CE = Ay (ate) SK=0,)1)...NO —@® Substituting equ® tn eq, be jatien apn) = y(AAeye No i N =o ( S ) ® Equation @) provides the recenstuction of the periodic signal pln) from the samples of the spectrum X(co}. Hower, to recody Klos) OY ln) from the samples certaim comditten needs to be satisfied. Since pC) js the periodtc extensien % cn) » it is clea thot x(n) Can be retovened from xpln) ip fhere is 10 mo_aliasing im the time domain i.e, iP oclm) is time ieited do dese than the period N of Xp (en). @ xp) - IT.. . ® p Zot : CT W2L “aplen) HO ALIASING “Tani, Comeiler 98, alm) ie a sequence mnonzre in the imterval o2M41-1. © When NZ (m) = tpl men om L, xlm) = 1pln) for penen 1[ & 6| Hewce x(n) Com be recovdud frorn peri). Whon N2L, '& cammot be rewred dus ft, Crane — domain aliasing. THE DISCRETE FOURLER TRANoFORM (DFT) $en gqemeral , the eqpoly spaced lreqpency contre x(2t- re) PROM hh, do snk wriquons vepren the original sequence, %(0) Who ~Cn) bos duvalio. 6 Imstead, the frequency sanples x(2%e) Covrespernal fo a periode Sequence Xp) of period N, Whe Xpln) is am aliased version of xm) glvew by xp(m) = E xln-n) = -00 When the Sequence x(m) has a fevite dusratien of senoth Len, theu xp) is serpy a periodic repetitien 4 x(n), where Zl) over a siengle periool is given by, Kplm) = pu, Oemel-1 0, Lemen-4 Hence, the freqrency sacmpler x (2m) , K=0,1,..,N uniquely represent the fimite duration Sequemee xlon), Since xl) = tpl) over A simate period ( padded by N-L 2evs), the original fimite duration sequence x(n) com be obtained from the prequemey sSaemples x (24K) from equ ©. Suamenarising . A fimite - duration sequence aln) 4 Jengtc | haa a Fourier fransforen X(wo) = z auton) @ , 0fWeeT m=0 Limtts imdicat x0) is ze outsicle the van eo Oe=mMm41-4, J + When we Sample Xu) at equally spaced frequencies Gop= SME) K=0)..yN-1 Whe NZL, the samples are X(e) = x (AE) = Same lefen N=0 ee —j20 em ty | Xk) = EZ. ume N Pee m= 0 arta) Upper index in the sum ha been imcreaved from L-4 to N-4 sionce x(n) =0 for MEL. dhe above relation is a fovemule. fer Cramsforrniiong O sequence Soctord | demotr LEN imto a Sequence op brequemncy sanyples b xc} of demos N. Since , the prequesncy Sampler ake obtained b evaluating the Fourier traastorm X(u) at a set of N Ceqpatty spaced) discrete prequemctey, the above relation is called discrete Fourier tromshvam (DFT) of xl). Let Wy = eV | then oe kn wef Xk) = ZB xm) wy ke o.u9, n=0 “i 8 She velotion which allows us dB recover the Sequence x(m ) from the frequency samples is called the Imuerse DFT ( IDFT), DET AS A LINEAR TRANFORMATION Wk, Nel DFT | amy} = X(kK)= & xm) we N=0 K=O,1,2,...,N-1 N-I - TDeT } xe = xn) = + = xe) Wee k=6 M=0,1)2,.-,N—1 —jot N Twiddle Factor whore Wy = © From the above equations we observe that the Computation. 4 each point of the DFT. requires N= Complex cmultiplicections and (N-1) Commplex. additions. Heave, N=point DFT veqpives N” Complex smattiplicsti¢ns and N(N-1) complex additions. _ q DET amd IDFT Caw be Wewed a dimear tawstor— mations om sequences { xem} and } xe}, respectivety, Let us defame Q N-poimt vector xy of the sigma Sequence alm), am N-point vedor Xy of Prequemuy Samples amd A NXN matrix Wy as [ x00) Xl) mye | 2 Xy=] xt) L x(n) X(n-1) -f— . : o(N-1) Wn = we? oe a oe wee wit wir) “tuo | (HDA (N-1)Cn-1) | wd ‘Jo Wr a Wy r 4 Ove eee ee N-) 4 Wy Wa Wy 2(N-1) 4 Wy wy! : » Wn ‘ ’ aTNe ne (ion) owe wat) wh From the above dlfienitioms, the N-poimé DFT cam be expressed im matrix om a Xu = WyXn : Whae Wy is the matrix % dhe Linear transformation, We obserwe trot Wy Is a symmetric omatrix. If we assume that tmvenrse of Wy exists dren IDET com be expresred im matrix fore n X= Wy Xi or] xy = 4 Wa Xy whee Wir demotes the Complex onjugati. Of, Wy. PROBLEMS 4) Compute the 4- poimt = DFT 1 a sequence aten) = | 4273, 4): Solm!— T- METHOD ¢ From the defn oi DFT wkt, Nol pet} xen) = x(k) = & x(n) wh” = j K=O; pN-| m= 0 Giver N=4, 2 kn a Kk) = Ex) Wy N=0 = rl) + 2G) WE + 20) Wy + aCe) Wy 3Ic [xe = 4 towP tawy tag | 5 kona Twiddle Factors 3 For N=H, Fiyst four twiddle faders. arF Wa=t Wy =-f 4 : 3 ° Wy=-j We =J POWERS -j J car Use Wy = e a and a Solve. let r=0, X(o)= 4+ 2 +344 = 10 2 3 Let k= 4, XC) = 4 + 2QWy £3Wy + 4 Wy 4 4 2(-j) + 3-1) + 4G) i} -2+2) uw Letts, X(@2) = 4 + 2wP + Swi +t 4we = t+ 2(-4) + 3(1) 44(-4) = -2 a6 4 Let Kea, X@) = 4 bP Wy + Oy + 4 Wy ree ne) makes cael) ety (e)) = =2-95 x xCe) = 4 10; =942)) -2, -2-95) | \o Teron «+ Xu = Xn Wy Xn & Xn > Row vectors oma) Xy = xy Wy [4 asu4ulf[4 4 4 4 4 WE Wig Wy Awd Wal vue Lowe We Wy! Fromm periodicity Propaty of Wat eee i = Wy and symmetry Propaty 5 we - we y -Creeu\fs t+ tt 4 -j -b J ieee ee la 4 j -bo Additiomal problems en DFT t- 2) alm) = fo, 412,38} for Nett ANS? X(k) = 6, 242), -2, ~2-9}} 3) atm) = | 4,2, 9, 4) doy NEY. ANS! XC) =f 6 -1-j, 0, -14)h 1d 4) ln) = Sen) + Sln-1) = S(m-2) ~ s(m-a) ~ } 1, 4, -4,-2\ ANS} X(r) = ( 0, 2-2) , 0, 2425} 5) Find the k-polnt rper % x(k) = i lo, ae +2], ~2 ,-2-2)}. Self — From the defn of IDFt, No} a rort 4 xle)} = x(n) = 1 Z. x(t) War” N “tee M = Oh Dp.-y Not for Neu, 3 er xm) = 1 = XC) Wy 4 20 xm) = [ Xo) + xX) Wy + x(2) wy” +x(3) wa?” | ax) = + [ 10 + (9424) Wy” + C2) Wy 31 - + (2-2) Wa | 7 = O14, 2/3. TINIDDLE FACTORS$ Tor Ne4, 4 ° oO] —2 +] Wy = 4 Wy =-4 \ “48 3 ° 2—+——— 4-0 Wy = J Wy =-) j [ lo + (-242]) Wye + (C2) Way? + (2-9j) Wa | [ to + Gat2})Cj) + C2(-2) + C22) | x(4) = - =o Let m=2, x(a acl to + (+9) wy? + Co) vy + 2-9]) wy ® | = y to + (-2423) (-1) + @2) (4) 42-95) (4) | xb) = “a = 3 Let m=%, (8) = i [" to 4 bats a vy +2) Wy (2 }) Wo VI _ ‘ y - Se see = f4,2, 3,4 | ZI METHOD ¢— Cos “kn wet, um) = + & X(kc) Wa ‘ Neo Abwe equatien carn also be written aa, = * —> Row "= NN Xn ww Xn Xn Vectors, Given N=4, % Xy = 7 Xy Wi 1 ‘ P ht 4 4 Xq =f to G242j) -2 C2-2)) : ‘ 4 j ---J 4-4 4 -4 4 -j -1t J = LT 1 -2+9j-2-2-2} 4 4 -oj-2 +942)-2] = 4 3 \o -2) J — (0 t2-2)-2 42 424 4 Io WO 4o)42 42-2) 42 l6 NOTES ANSWER do) a(n) = 1,2,3,4) Ts IN Row, REPRESENTED As COLOMN ONLY FoR CONVENIENCE, © Find the DET of xn) = Sem), Sol i- From dafin cs DFT, Nol DET a(n) } = xX) = 2 xm) we” 2. N=0 NzI Hence, DFT Sén) | = z S¢n) rae N=0 WAT, S(m) = 4, ob N=0 0; tor n¢0 2 DET L som} = StolWy = ty rs or $m) ems + @ find the PFT ob, x(m) = S(m-M). solm:— From =the defin of DFT, net kn DET j tnd} = xX) = Z xl) Wr N=0 Noy kn Herre, pet | stn-mo)} = SF S(m-m) Wy M=0 4, at N=No reall -Mo) = ! \W.K-T, (nm) Vo eae ko Go DET | Slm-m)) = SO) Wy? = Wyrm $(n- No) pales ae jot © Fimd N-point DET x(n) = ewe". Solan — From de gienition 4 DFT Nel oer j cxler) | = xh) = & x(n) we 7: Hence, pet | oe aa aw N Noe n —j2 ¢y- ° 2 nl -[ oe tI] Z I XC) - Po i -J2T (e- ko? = ___— _ fa I- a ak Keke, x(e) = & ; Using} L-Hospifals Method , (-" eo POs) 6 - = - cist Ce Ke) - _pm elt Ce ts) 0 (-§ " keto Ae KaKop © = jam. € = N =r 8 Jol.e N & Xe) = Nat Ke ko ak kK#Kko x(k) = N at k=ko [xGe = WN sCe-ra} m p= | 0 fr Eto @ Find the DFT oh x(n) = et kom Solo'- From the defn of DFT, N pet ham} = xe) = SF xen) wh” =o . N-1 vot -] = -j27 J2M ben Hence, DET fe 1 koro} =e Teoma! N=0 \\ [ ox Cerra n=O M - \ c—_ o ea WE fe a iN & I x NN ®3 m0 Nt] y = Il-a@ - 21 Cette) ee a- eR rl oe - O ok k= ko, XC) = —- 0 Using L- Hospital Rode, 7S (-ie") el TICict to) XQ) = \ — PT Ck +b) —)20 0. mac at) & yy keke atk=-te, X(kh)=e YU = N A at kK t-ko, Xk) = 0 oo Not keko ce [xce) = N Se +o) | oy |x(e) = Oot ov X(k) = N SCK-N tke) ; oa ‘4 Find the N-point DFT of x(m) = Cos 2 kom N Solmi— From depo of DFT, N-I « DET {x(n } = Xk)= 4 xm we” =0 per | taal (om Teom){ = Z Eun) vin N N-t eA iat | othe —)2M jeen e N 20 Ce Not 2M (kK Ko)n 2 dere? = gen —~ 1-0 Nt _32T (h+ ko) + ee 1 (ett i M=0 eal | NSCe- bo) + NS Ce+ta)h “ bos ( 2.x) pe & { Slee) + § Cettal @ Simi los problerm , a(n) = sim 20 kon ans: 30 ( en) SM [ SCe- vo) Stet to)| >] 20 (@ Fimd the 5- pam DFT Hh x(n) = } i, 1,4] SelM j From the dep of DFT, ower y N-I Xe) = atm) wh M=0 Q a4 a R 3 a) I Ke Ot.) NOI thiven N=S, A X(k) = = xn) we jp Ke 0,4, 2,3,4 X(c) = tle) + CWE + 26) tale + b2) We # x(a) Wel yle) = EWE We | ke aya | Twiddle Fadyst N=S We = 4 We = oF cos (22\-3 sim (227) 0.2090 — 0.96 We = eV? woe (a) | vin (10) Ocoq jose es We = ¢ iN (ws ( eu j an) o.noq 4 JOSST a1 —j 20.4 st) _ | sin fst Wie = @ 5 = tos ( $2) J sin (2m) = 0.309 + J 0.98 Evawoat x(k), at k=0 xf) = ttttt = (lo t W at kK=4, X(4) t two + we an os - jis3ge at K22, XG) = t+ We + Wel = os + jo.3633 3 6 . at Kea, K(s) = + + Ws + Ws = 0.8 ~j 0.3685 at coh, XG) = D4 Wel + We = oS +] Lees X(c) = | 3, 0S —jl.538g, 0.5 +) 0-363, 0.502633, 0S + J 15388 5 @ Hed IFT fpr the following sequence Xte) = | 1, 9, Wi), 0, 4, 0, WHi),0f Sol*n:— Lemath 4 the Sequence is 8, eo N=8 From the difin Of IDET, sot} xtc} = xem) = 4 Z Xe) Wy a2 For N=8, E eck E X(t) W, K=0 x(n) = a g x(m) = + RCO) eerx Cl) eae) lea sm + xO) we?” + XG) We" + XC5) We - -4 + X(6) Wen” + X(T) We "| axl) = t [ eS) Ome (|) | 0 peti +0 + (i ‘Due + o| wnare n= 0/1,2,...,F xm) = 4 [ See awe?” i Wey (is) a or TWLDDLE FACTORS 5 Nes 2 cI -4 IMs we = 0.404 +) 0F0F ; oA OF0F +JOF0F OY pieeelay Ws Ve oO 1 We — o “A Wet 8 Wa 23 For m=0, X(0) = | Shtiejytat ij tL g -=SB st g = For m= 4, a(t) = 4 J s + (-J)wg” + We! + (i) wet S s +0-j)j + C+ (aie M " + 4p stinr-t alg & ar) q ” For m22, 20) = t Js + (A)) We + We ° +) wg? } = 4 | s4O-Q) +r ++ GHD] = ¢ [5-145 +1-1-J] a a 12 4 \s + (-J) We + We (4}) we? } fir mz3, (2) -4 | SHC) +O) +H} 24 wo) 2g bs jrP-tti re) . 7a | ad fs-j-r-1+j-1} ° $7 formen, or) = b | st Cine? + wey a+j) wg | a + 5 s +Ui-j)+i +G4j)} => Sit _—_ dav nes, © —\0 -20 x(S) = + , Ss +(I-j) Wg + We HO) 03%? | k fs eG-s Hey + Gagpenl kK | s+i--! -3-7} $ \l “ fa m=b, x6) = + }s +0-j) wy +We u We \s eO-a) +1 +G4i)C0} a\- i ont tat 2s formet, x9) = ths +(- De" a + (itp) We “yp a | s +(0-J)(-j) +G) +aga} + 4 ul s -j+p-itj+r} a % co x(n) = frre ty test, () Find ver of the sequence moe oem eo. 0, etherwise ~£\- dy ee, Plob |XCe)| 2 ZX(e). Selmi From the dain of DFT, N-I ke per ham} = xe) = A lm) Wy N=0 for N=8, ZL tm) an . De) aaa emer N=0 4, 4,4} Grivenn 9C(m) 26 ci k oe bo X(c) = 4 + We + We F0+ 0404040 ee aa 3 | xt = Lt Wet Wed |, R= Ol 2 t Twiddle Fadors fev N=% We = 4 eer We” = 0-404 - jot We = 70-407 + jo FoF We =I we = J We = — 030+ —Jo-t04 Wet = 0-704 + 0.404 (Locode, points em the umit circle im complex plame direction is Clockwise fpr Det) sata Oe ame CO) imate tallest lee oe Poy cea, XO) = b+ We + We = 1 + (0x04 jot) + CS) = Lt —Jitot fv cea, XG) = bE WY + Wg! 21+ +1eo =a for k=3, — x(3) Ny 1 + We + Wee \l 4 + (-0404~Jjo.404) +] 0-293 +f 0.293 —————— (l 2 for K=H, X(y) = 1 + Wel + we t+) ¢it = 4 fy tes, Xs) = 2 + WE + 0° l 4 4 (-0.10++ jot04) +Cj) = 0.293 — J 0-293 dy ce, Xl6) = 4 + We® + We =f£4+j+G) =J 1+ we 4 wel a tov k= 4, (4) = 4 40.404 + J oto tj it “XCk) = 1 3, Ct -Jhaop, —j, (993 + jo.299), £, (0.299 -j 0.293) ,f, (1.40% +4 4.404) } Por Magltad Plo b KGa) = 43, 24m, 2, 04H, 2, O4lY, 4, 2.414 \ Phone Ploty -Tt -7 po fh EE EES aa 23 29 PROPERTIES OF DET 1) LINEARITY §- STATEMENT EN ager) PVs xX, () x(n)

You might also like