Signal Processing and Linear Systems-B.P.Lathi copy

T he t otal i nput is therefore f k m6k puk 1

This preview shows page 1. Sign up to view the full content.

This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: odic s ignal I[k] b y c onsidering J[k] a s a l imiting c ase o f a p eriodic signal w ith t he p eriod a pproaching infinity. 10.1 Periodic Signal Representation by Discrete-Time Fourier Series (DTFS) A p eriodic s ignal o f p eriod No is called a n N o-periodic signal. F igure 8 .9 shows a n e xample o f a p eriodic s ignal o f p eriod 6. A c ontinuous-time p eriodic s ignaf o f p eriod To c an b e r epresented a s a t rigonometric F ourier series c onsisting o f a sinusoid o f t he f undamental f requency Wo = ~, a nd a ll i ts h armonics ( sinusoids o f f requencies t hat a re i ntegral m ultiples o f wo). T he e xponential f orm o f t he F ourier series consists o f e xponentials e jOt , e ±jwot, e ±j2wot, e ±j3w0 t, .... F or a p arallel d evelopment o f t he d iscrete t ime c ase, recall t hat t he f requency o f a s inusoid o f p eriod No is n o = 21r/No. H ence, a n N o-periodic d iscrete-time s ignal I[k] c an b e represented b y a d iscrete-time F ourier series w ith f undamental f requency n o = 'flo a nd i ts h armonics. A s i n t he c ontinuous-time case, we m ay u se a t rigonometric o r a n e xponential f orm o f t he F ourier series. B ecause o f i ts c ompactness a nd e ase o f m athematical m anipulations, t he e xponential f orm is preferable t o t he t rigonometric. F or t his r eason we s hall b ypass t he t rigonometric form a nd go d irectly t o t he e xponential f orm o f t he d iscrete-time F ourier series. T he e xponential F ourier series consists o f t he e xponentials e jOk e ±jflok e ±j2flo k e ±jnflok, . .. , a nd s o o n. T here w ould b e a n i nfinite numb~r o f h~rmonics: 617 10 Fourier Analysis o f D iscrete-Time Signals 10.1 P eriodic Signal R epresentation b y Discrete-Time Fourier Series ( DTFS) 619 e xcept for t he p roperty proved in Sec. 8.2: t hat d iscrete-time exponentials whose frequencies are s eparated b y 21r ( or integral multiples o f 21r) a re identical because We now have a discrete-time Fourier series ( DTFS) r epresentation of a n N o-periodic signal f[k] a s 6 18 N o-I J[k] = (10.1) L v rejrrlok (10.8) r =O T he consequence of this result is t hat t he r th h armonic is identical t o t he ( r+No)th n armonic. To d emonstrate t his, let 9n d enote t he n th h armonic ejnrlok. T hen w here 1 Vr No (10.2) a nd 9r = 9r+No = 9r+2No = . .. = 9r+mNo m , i nteger (10.3) T hus, t he first harmonic is identical t o t he (No + 1) st h armonic, t he s econd harmonic i s i dentical t o t he ( No+2)nd h armonic, a nd so on. I n o ther words, t here a re o nly No i ndependent h armonics, a nd t hey r ange over a n i nterval 21r ( because t he h armonics a re s eparated by n o = We m ay choose t hese No i ndependent h armonics as e jrrlok over 0 ::; r ::; No - 1, or over - 1 ::; r ::; No - 2, or over 1 ::; r ::; No, o r over a ny o ther s uitable choice for t hat m atter. E veryone o f t hese s ets will have t he s ame narmonics, a lthough i n different order. L et us t ake t he first choice (0 ::; r ::; No - 1). T his choice corresponds t o e xponential...
View Full Document

Ask a homework question - tutors are online