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

In practice t he s ampling interval t c annot be zero

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

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: soid cos Ok is periodic only if O/27r is a r ational number. Second, discrete-time sinusoids whose frequencies 0 differ by a n integral multiple of 27r a re identical. Consequently, a discrete-time sinusoid of a ny frequency 0 is identical t o some discrete-time sinusoid whose frequency lies in t he interval - 7r t o 7r (called t he f undamental frequency range). Further, because c os ( - Ok + {I) = cos (Ok - 0), a sinusoid of a frequency in t he range from - 7r t o 0 can b e expressed as a sinusoid of frequency in the range 0 t o 7r. T hus, a discrete-time sinusoid of any frequency can be expressed as a sinusoid of frequency in t he range 0 t o 7r. T hus, in practice, a discrete-time sinusoid frequency is a t most 7r. T he highest r ate o f oscillation in a discrete-time sinusoid occurs when its frequency is 7r. In a given time, a sinusoid of frequency other t han 7r will have a fewer number of cycles ( or oscillations) t han t he sinusoid o f frequency 7r. T his peculiarity of nonuniqueness o f waveforms in discrete-time sinusoids of different frequencies has a far reaching consequences in signal processing by discrete-time systems. O ne useful measure of t he size of a discrete-time signal is i ts energy defined by t he s um L:k If[k]i2, if it is finite. I f t he signal energy is infinite, t he p roper measure i s its power, if it exists. T he signal power is t he t ime average of its energy (averaged o ver t he entire time interval from k = - 00 t o 00). For periodic signals, t he time averaging need be performed only over one period in view o f t he periodic repetition o f t he signal. Signal power is also equal to t he m ean squared value o f t he signal (averaged over t he entire time interval from k = - 00 t o 00). Sampling a continuous-time sinusoid cos (wt + 0) a t uniform intervals of T s econds results in a discrete-time sinusoid cos (Ok +0), where 0 = wT. A continuous 569 Problems time sinusoid of frequency F Hz must be sampled a t a r ate no less t han 2 F Hz. Otherwise, t he r esulting sinusoid is aliased; t hat is, it appears as a sampled version of a sinusoid of lower frequency. Discrete-time signals classification is identical t o t hat of continuous-time signals, discussed in chapter l . A signal j [k] delayed by m t ime units (right-shifted) is given by f [k - m]. O n t he o ther hand, f [k] advanced (left-shifted) by m t ime units is given by f [k +m]. A signal f[k], when time inverted, is given by f [-k]. T hese operations are t he s ame as those for t he continuous-time case. T he case of time scaling, however, is somewhat different because of t he discrete n ature o f variable k. Unlike t he c ontinuous-time case, where time compression results in t he s ame d ata a t a higher speed, time compression in t he discrete-time case eliminates p art of t he d ata. Consequently, this operation is called decimation o r downsampling. T ime expansion operation of discrete-time signals results in t ime e xpanding t he signal, t hus c...
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