Hossein Sameti Department of Computer Engineering Sharif University of Technology • Eigen vector of matrix A: AX X • In other words, once matrix A is multiplied by vector X, the direction of X is preserved. • Eigen function of a system: Ф(n) System αФ(n) Hossein Sameti, CE, SUT, Fall 1992 2 cosn j sin n x(n) LTI System e jn y(n) H ( )e jn k k y ( n) x ( k ) h( n k ) h( k ) x ( n k ) H ( ) h(k )e jk k y ( n ) h ( k )e k j ( n k ) ( h(k )e jk )e jn k Frequency response y (n) H ( )e jn jn H ( ) magnifies the input e based on freq ω. Clarification: Some textbooks use H (e j ) instead of H(w ). Hossein Sameti, CE, SUT, Fall 1992 3 H ( ) h(k )e jk 1 k H ( 2 ) h(k )e j ( 2 ) k h(k )e jk e j 2k H ( ) k k • Frequency response is periodic with the period of 2π. Implication? Hossein Sameti, CE, SUT, Fall 1992 4 cos 0 n : Hossein Sameti, CE, SUT, Fall 1992 5 H ( ) h(k )e jk k H ( ) h ( k ) e j k h ( k ) k k h( k ) k The same condition as the stability condition Hossein Sameti, CE, SUT, Fall 1992 6 DTFT{x(n)} X ( ) x(n)e Same mathematical representation as the freq. response jn n • Existence of DTFT: x ( n) n • Inverse DTFT: x(n) is absolutely summable. x(n) 1 2 jn X ( ) e d Fourier analysis considers signals to be constructed from a sum of complex exponentials with appropriate frequencies, amplitudes and phase. Frequency components are the complex exponentials which, when added together, make up the signal. Hossein Sameti, CE, SUT, Fall 1992 7 IDTFT of the ideal low-pass filter: c 1 X ( ) 0 c c c x(n) 1 X ( )e jn d 1 e jn d 2 2 1 e jc n e jc n jn c x(n) [e ]c j 2n j 2n x ( n) sin c n n Hossein Sameti, CE, SUT, Fall 1992 8 sin(𝜔𝑐 𝑛) 1, ⟺𝑋 𝜔 = 0, 𝑛 |𝜔| ≤ 𝜔𝑐 𝜔𝑐 < 𝜔 ≤ 𝜋 Hossein Sameti, CE, SUT, Fall 1992 9 x ( n) a n u ( n ) n jn X ( ) a e n 0 DTFT? (ae j ) n X ( ) n 0 1 1 ae j a 1 What happens if a>1? 1 a cos a sin j 1 a 2 2a cos 1 a 2 2a cos X R () X I () Hossein Sameti, CE, SUT, Fall 1992 10 x ( n) a n u ( n ) X ( ) 1 1 ae j X () X () e jX ( ) 2 X ( ) X ( ) X * ( ) 1 1 a 2 2a cos X ( ) arctan( a sin ) 1 a cos Hossein Sameti, CE, SUT, Fall 1992 11 Hossein Sameti, CE, SUT, Fall 1992 12 Hossein Sameti, CE, SUT, Fall 1992 13 x1(n) X1() x2 (n) X 2 () • Linearity: a1x1(n) a2 x2 (n) a1X1() a2 X 2 () • Time-shifting: • Time-reversal: x(n k ) e jk X () x(n) X ( ) • Convolution : x(n) LTI System h(n) y ( n) x ( n ) * h( n) y(n) Y ( ) X ( ) H ( ) Hossein Sameti, CE, SUT, Fall 1992 14 • Cross-correlation: y(n) x1(n) * x2 (n) Y () X1() X 2 () •Frequency Shifting: e j0 n x(n) X ( 0 ) •Parseval’s Theorem: * * 1 x ( n ) y ( n ) X ( ) Y ( )d 2 2 2 y ( n) x ( n) x(n) 21 X ( ) Hossein Sameti, CE, SUT, Fall 1992 15 • Modulation: x(n) cos 0n 1 2 X ( 0 ) 12 X ( 0 ) • Multiplication: x1(n) x2 (n) 1 2 X1 ( )X 2 ( )d • Differentiation in the freq. domain: nx(n) j dX ( ) d • Conjugation: x* ( n ) X * ( ) Hossein Sameti, CE, SUT, Fall 1992 16 • Conjugate Symmetric: x(n) x* (n) • Conjugate Anti-Symmetric: x(n) x* (n) • Why are these properties important? Conjugate Symmetric xe (n) 12 [ x(n) x* (n)] x(n) xe (n) xo (n) Conjugate Anti-symmetric xo (n) 12 [ x(n) x* (n)] X e () 12 [ X () X * ()] X () X e () X o () X o () 12 [ X () X * ()] Hossein Sameti, CE, SUT, Fall 1992 17 x (n ) X ( ) x* ( n ) X * ( ) x* (n) X * ( ) Re{ x(n)} X e () j Im{x(n)} X o () xe (n) xo (n) X R () Re{ X ()} jX I () j Im{X ()} Hossein Sameti, CE, SUT, Fall 1992 18 x (n ) : real x(n) Re{ x(n)} X () 12 [ X () X * ()] X () X e () X ( ) X * ( ) • If a sequence is real, then its DTFT is conjugate symmetric. Hossein Sameti, CE, SUT, Fall 1992 19 x (n ) X ( ) x (n ) : real X ( ) X * ( ) x (n ) : real X R () X R () x (n ) : real X I () X I () x (n ) : real X ( ) X ( ) x (n ) : real X ( ) X ( ) Hossein Sameti, CE, SUT, Fall 1992 20 Proakis, et.al Hossein Sameti, CE, SUT, Fall 1992 21 Determining an inverse fourier transform X (e X (e j j 1 ) (1 ae j )(1 be j ) a /(a b ) b /(a b ) ) j 1 ae 1 be j b a n x [n ] ( )a ( )b n u [n ] a b a b Hossein Sameti, CE, SUT, Fall 1992 22 Determining the Impulse response from the frequency response Hossein Sameti, CE, SUT, Fall 1992 23 Determining the Impulse response for a Difference 1 1 Equation y [n ] y [n 1] x [n ] x [n 1] 2 4 To find the impulse response h[n], we set x [n ] [n ] 1 1 h [n ] h [n 1] [n ] [n 1] 2 4 Applying the DTFT to both sides of equation. We obtain 1 1 H (e j ) e j H (e j ) 1 e j 2 4 1 1 e j 4 H (e j ) 1 1 e j 2 Hossein Sameti, CE, SUT, Fall 1992 24 Example: x[n] a nu[n 5] 1 a u[n] 1 ae j j 5 e a n5u[n 5] 1 ae j a 5e j 5 n a u[n 5] 1 ae j n Hossein Sameti, CE, SUT, Fall 1992 25 Reviewed Discrete-time Fourier Transform, some of its properties and FT pairs Next: the Z-transform Hossein Sameti, CE, SUT, Fall 1992 26