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

# 1015 because as seen in sec 82 t he t wo exponentials

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: 0) k=-oo 10.2-1 As No - + 0, !1o becomes infinitesimal (!1 o - + 0). For t his reason i t will be appropriate t o replace !1o with a n infinitesimal notation L'l.!1: L'l.!1= 271" No L F(!1) = f[kJ (1O.25b) f[kJ. Therefore [F(r!1 o)!1o ] 271" J27r 271" F(!1) = F {f[k]} r=<No> In the limit as No r F(!1)e jkfl d!1 ~ where J27r indicates integration over any continuous interval of 271". T he s pectrum F(!1) is given by [Eq. (1O.22)J Using Eqs. (10.23) a nd (10.21), we c an express Eq. (10.20) as (1O.25a) 627 Aperiodic Signal Representation by Fourier Integral F(!1 + 271") = k=-oo L 00 f [kJe- j (fl+27r)k = f[kJe-jflke-j27rk = F(!1) (10.32) k=-oo Clearly, t he s pectrum F(!1) is a continuous a nd periodic function o f!1 w ith period 271". We must remember, however, t hat t o synthesize J[k], we need t o use the spectrum over a frequency interval o f only 271", s tarting a t any value of!1 [see Eq. (1O.30)J. 628 10 Fourier Analysis of Discrete-Time Signals 629 10.2 Aperiodic Signal R epresentation by Fourier Integral As a m atter o f convenience, we shall choose this interval t o be t he f undamental frequency range (-11", 11"). I t is, therefore, not necessary t o show discrete-time-signal spectra beyond t he f undamental range, although we often do so. f [kJ C onjugate S ymmetry o f F{!1) for real j [k] From Eq. (1O.31), we o btain (a) 00 F {-!1) = L j [k]e jflk 10 5 k-- k =-oo T he r ight-hand side of this equation is t he conjugate o f t he r ight-hand side of Eq. (1O.31) for real j [k]. Therefore, for real j [k], F{!1) a nd F { - !1) a re conjugates; t hat is, F {-!1) = F*{!1). Since F{!1) is generally complex, we have b oth a mplitude a nd angle (or phase) spectra F{!1) = IF{!1)le jLF (fl) (10.33) -3" - 2" 0 - It " 21t 3" Q __ Because o f c onjugate symmetry of F{!1), i t follows t hat IF{!1)1 = I F{ -!1)1 LF{!1) = - LF{ - !1) L F(Q) (1O.34a) (1O.34b) Therefore, t he a mplitude spectrum IF{!1)1 is a n even function of !1 a nd t he p hase spectrum LF{!1) is a n o dd function o f!1 for real j [k]. Linearity o f t he D TFT According t o Eq. (10.31), it follows t hat if F ig. 1 0.4 Exponential ' lurk] and its frequency spectra. and then (1O.35) E xistence o f t he D TFT Because l e-jkfll = 1, from Eq. (1O.31), i t follows t hat t he existence of F{!1) is g uaranteed if j [k] is absolutely summable; t hat is, exponentials (or sinusoids). T he Fourier s pectrum of a signal indicates t he relative amplitudes a nd p hases o f t he e xponentials (or sinusoids) required t o synthesize j [k]. I f J[k] is periodic, t hen i ts Fourier s pectrum has finite amplitudes a nd exists a t discrete frequencies (!1 0 a nd i ts multiples). Such a s pectrum is easy t o visualize, b ut t he s pectrum of a n a periodic signal is n ot e asy t o visualize because i t is continuous. T he physical meaning of t he continuous s pectrum is fully explained in Sec. 4.1-1. • E xample 1 0.3 Find the DTFT of j[k] = -yku[k]. 00 L Ij[k]1 < 0...
View Full Document

## This note was uploaded on 04/14/2013 for the course ENG 350 taught by Professor Bayliss during the Spring '13 term at Northwestern.

Ask a homework question - tutors are online