06/10/2016 DIGITAL SIGNAL PROCESSING Frequency-Domain Representation of Discrete-Time Signals and Systems 1 Fourier Transform (FT) Company LOGO • Analysis Analysis F x(n) X (e j ) xn e X e j jn Fourier Transform (FT) X (e j ) F [ x(n)] n Synthesis x ( n) 1 2 X e e d j 2 j Inverse Fourier Transform (IFT) x(n) F -1[ X (e j )] www.company.com 1 06/10/2016 Fourier Transform (FT) Company LOGO 3 j X (e ) n x[n]e j n is a complex-valued function n X (e j ) X R (e j ) jX I (e j ) magnitude X (e j ) X I ( e j ) phase | X ( e j ) | X ( e j ) X (e j ) | X (e j ) | e jX (e j ) X R ( e j ) j X (ewww.company.com ) X ( e j ) e j ( ) Company LOGO 4 • Example:Find fourier transform of x(n): x ( n) ( n) x ( n) ( n no ) x ( n) u ( n) n 1 x ( n) u ( n) 2 n x ( n) 2 u ( n) x(n) rect N (n) www.company.com 2 06/10/2016 Key Issue Company LOGO We need that |X(ej)| < for all Analysis j X (e ) x ( n )e j n n Does X(ej) exist for all ? Synthesis 1 X (e j )e jn d 2 x ( n) www.company.com Fourier Transform (FT) Sufficient Condition for Convergence Company LOGO Sufficient Condition for Convergence | x(n) | | X (e j ) | for all n | X ( e j ) | x ( n ) e j n n | x(n) || e j n | x ( n )e j n | n | n | x ( n) | www.company.comn 3 06/10/2016 Company LOGO Conjugate-Symmetric and Conjugate-Antiymmetric Sequences • Conjugate-Symmetric Sequence xe (n) xe* (n) Called an even sequence if it is real. • Conjugate-Antisymmetric Sequence xo (n) xo* (n) Called an odd sequence if it is real. www.company.com Company LOGO • Fourier Transform (FT) Sequence Decomposition Any sequence can be expressed as the sum of a conjugate-symmetric one and a conjugate-antisymmetric one, i.e., x(n) xe (n) xo (n) Conjugate Symmetric Conjugate Antiymmetric xe (n) 12 [ x(n) x * (n)] xo (n) 12 [ x(n) x * (n)] www.company.com 4 06/10/2016 Company LOGO • Fourier Transform (FT) Sequence Decomposition Any function can be expressed as the sum of a conjugatesymmetric one and a conjugate-antisymmetric one, i.e., X (e j ) X e (e j ) X o (e j ) Conjugate Symmetric Conjugate Antiymmetric X e (e j ) 12 [ X (e j ) X * (e j )] X o (e j ) 12 [ X (e j ) X * (e j )] www.company.com Company LOGO Conjugate-Symmetric and Conjugate-Antiymmetric Functions • Conjugate-Symmetric Function X e (e j ) X e* (e j ) Called an even function if it is real. • Conjugate-Antisymmetric Function X o (e j ) X o* (e j ) Called an odd function if it is real. www.company.com 5 06/10/2016 Fourier Transform (FT) Symmetric Properties Company LOGO • Symmetric Properties F x(n) X (e j ) F x(n) X (e j ) x(n)e jn n magnitude x ( n )e jn X ( e j ) n magnitude phase phase www.company.com Fourier Transform (FT) Symmetric Properties Company LOGO • Symmetric Properties F x * (n) X * (e j ) F x(n) X (e j ) x * (n)e jn n x(n)e jn * n magnitude phase www.company.com * x(n)e jn X * (e j ) n magnitude phase 6 06/10/2016 Fourier Transform (FT) Symmetric Properties Company LOGO • Symmetric Properties F x(n) X (e j ) F x * (n) X * (e j ) magnitude magnitude phase www.company.com Company LOGO phase Fourier Transform (FT) Symmetric Properties • Symmetric Properties F x(n) X (e j ) F Re{x(n)} X e (e j ) Re{x(n)} 12 [ x(n) x * (n)] 1 2 F j * j 1 [ x(n) x * (n)] )] 2 [ X (e ) X (e F j Im{x(n)} X o (e j ) j Im{x(n)} 12 [ x(n) x * (n)] www.company.com 2 [ x ( n) 1 F j * j 1 x * (n)] )] 2 [ X (e ) X (e 7 06/10/2016 Company LOGO Fourier Transform (FT) Symmetric Properties • Symmetric Properties F x(n) X (e j ) F xe (n) X R (e j ) xe (n) 12 [ x(n) x * (n)] F j * j 1 1 2 [ x(n) x * (n)] 2 [ X (e ) X (e )] F xo (n) jX I (e j ) xo (n) 12 [ x(n) x * (n)] www.company.com 2 [ x ( n) 1 Company LOGO F j 1 x * ( n)] ) X * (e j )] 2 [ X (e Fourier Transform (FT) Symmetric Properties • Symmetric Properties F x(n) X (e j ) F x * (n) X * (e j ) magnitude Facts: phase 1. real part is even X R (e j ) X R (e j ) 2. Img. part is odd X I (e j ) X I (e j ) j j 3. Magnitude is even | X I (e ) || X I (e ) | 4. Phase is odd X (e j ) X (e j ) www.company.com 8 06/10/2016 Fourier Transform Theorems Company LOGO • Linearity FT FT x1 (n) X 1 (e j ), x2 (n) X 2 (e j ) F ax1 (n) bx2 (n) aX 1 (e j ) bX 2 (e j ) • Example:Find fourier transform of n x( n) 2 x1 (n) 3 x2 (n) n 1 x1 (n) u (n) 2 1 x2 ( n ) u ( n ) 3 www.company.com Company LOGO Fourier Transform Theorems F • Time Shifting x(n nd ) e jn X (e j ) d F [ x(n nd )] x(n n d n x(n)e )e jn j ( n n d ) n e jnd x(n)e jn n e jnd X (e j ) Phase Change www.company.com 9 06/10/2016 Fourier Transform Theorems Company LOGO FT • Frequency Shifting x (n) X (e j ) F e j0n x(n) X (e j (0 ) ) F [e j 0 n x(n)] e j0 n n x(n)e jn x ( n )e j ( 0 ) n n X (e j (0 ) ) www.company.com Company LOGO Fourier Transform Theorems • Time Reversal F x (n) X ( e j ) F x(n) X (e j ) F [ x(n)] x ( n )e n x ( n )e jn j ( ) n n X (e j ) www.company.com 10 06/10/2016 Fourier Transform Theorems Company LOGO FT X (e j ) • Differentiation in Frequency x(n) F nx(n) j F [nx(n)] nx(n)e d X (e j ) d jn n 1 de jn x ( n ) j n d d d j x(n)e jn j X (e jn ) d n d www.company.com Fourier Transform Theorems Company LOGO • The Convolution Theorem y ( n) x(k )h(n k ) Y (e F j ) X ( e j ) H ( e j ) k F [ y (n)] y ( n)e jn n k x(k )h(n k ) e jn n k x(k ) h(n k )e k n jn x(k ) h(n)e n j ( n k ) j k x ( k ) e h ( n ) e j n k n X (e j ) H (e j ) www.company.com 11 06/10/2016 Fourier Transform Theorems Company LOGO • The Modulation or Window Theorem F y (n) x(n) w(n) Y (e j ) Y ( e j ) w(n) x(n)e jn n www.company.com 1 X (e j )W (e j ( ) )d 2 1 w(n) X (e j )e jn d e jn 2 n 1 w(n) X (e j )e j ( ) n d 2 n 1 2 1 2 j X ( e ) w(n)e j ( ) n d n X (e j )W (e j ( ) )d Fourier Transform Theorems Company LOGO • Parseval’s Theorem 1 x ( n) y * ( n) 2 n Facts: X (e j )Y * (e j )d F x(n) X ( e j ) F y * (n) Y * ( e j ) F y (n) x(n) w(n) Y (e j ) x(n) y * (n)e j n www.company.com 1 X (e j )W (e j ( ) )d 2 1 X (e j )Y (e j ( ) )d 2 n Letting =0, then proven. 12 06/10/2016 Fourier Transform Theorems Company LOGO • Parseval’s Theorem- Energy Preserving 1 | x ( n) | 2 2 n | x ( n) | 2 n | X (e j ) |2 d x ( n) x * ( n) n 1 X (e j )X * (e j )d 2 Energy spectral density www.company.com 1 | X (e j ) |2 d 2 j s xx X (e ) 2 Fourier Transform Sinusoidal and Complex Exponential Sequences Play an important role in DSP Company LOGO h( k ) x ( n k ) y ( n) h( k ) x ( n k ) y ( n) x(n) e jn LTI k k h(n) h( k )e j ( nk ) k h(k )e jk e jn k H (e j )e jn www.company.com 13 06/10/2016 Company LOGO x ( n ) e j o n LTI y ( n) h( k ) x ( n k ) k h(n) H (e j0 )e jn0 y ( n ) H ( e j 0 ) x ( n ) www.company.com Fourier Transform Frequency Response Company LOGO Frequency Response X ( e j ) Y (e j ) X (e j ).H (e j ) H ( e j ) Y ( e j ) H (e ) X ( e j ) j www.company.com 14 06/10/2016 Fourier Transform Frequency Response Company LOGO Frequency Response j H (e ) h ( n )e jn n H (e j ) H R (e j ) jH (e j ) H (e j ) | H (e j ) | eH ( e j ) phase magnitude www.company.com Fourier Transform Frequency Response Company LOGO Example: Ideal Lowpass Filter H ( e j ) c c 1 | | c H ( e j ) 0 c www.company.com 1 c H (e j )e jn d c 2 1 c jn e d 2 c 1 c jn e d ( jn) 2 jn c h( n) 1 e jn 2 jn c c sin c n n 15 06/10/2016 Fourier Transform Frequency Response Company LOGO Example: Ideal Lowpass Filter h( n) sin c n n n 0,1,2, The ideal lowpass fileter Is noncausal. 0.6 0.4 0.2 0 -0.2 www.company.com -60 -40 -20 0 20 40 60 Fourier Transform Frequency Response Company LOGO Example: Ideal Lowpass Filter h( n) sin c n n n 0,1,2, TheTo ideal lowpass fileter approximate the Is noncausal. ideal lowpass filter using a window. 0.6 j sin c n jn e 0.2 n n M 0.4 M H (e ) 0 -0.2 -60 -40 www.company.com -20 0 20 40 60 16 06/10/2016 Fourier Transform Theorems Company LOGO Example: Ideal Lowpass Filter 2 M=3 1 0 -1 -4 -3 -2 -1 0 1 2 3 4 3 4 3 4 2 M=5 H ( e j ) sin c n jn e n n M 1 M 0 -1 -4 -3 -2 -1 0 1 2 2 M=19 1 0 www.company.com -1 -4 -3 -2 -1 0 1 2 17