Download as pdf or txt
Download as pdf or txt
You are on page 1of 4

9/22/2022

Frequency domain
D.T. signals and systems

DIGITAL SIGNAL PROCESSING


SEM-1, 2022-23

ECE/EEE F434 Lecture 9


Dr. Sarang C. Dhongdi, Dept of EEE, BITS Goa

Sinusoidal/Complex exp sequences Sinusoidal/Complex exp sequences


 Fundamental property of LTI system is that the steady
state response to sinusoidal i/p is sinusoidal of the same 
frequency as that of i/p, with amplitude and phase
determined by the system.
If we define H ( e j )   h ( k )e
k  
 jk

Consider x ( n )  e j n

Then, y ( n )  H ( e j ) e j n
 h( k )e
By the def of j ( n  k )
convolution y (n) 
k   Frequency Response -Represents change
in complex amplitude of the complex
 exp as function of w.

y ( n )  e jn  h ( k )e
k  
 j k

Frequency Response Sinusoidal/Complex exp sequences


 A sinusoid can be expressed as linear combination
H (e j )  H R (e j )  jH I (e j ) of complex exponentials.

A j j 0 n A  j  j 0 n
x ( n )  A cos(  0 n   )  e e  e e
jw
2 2
H (e j )  H ( e j ) e j arg( H ( e )) Then

y (n) 
A
2

H ( e j  0 ) e j  e j 0 n  H ( e  j 0 ) e  j  e  j  0 n 
y ( n )  A H (e j 0 ) cos(  0 n     )
Θ =Phase response at freq ωo

1
9/22/2022

Sinusoidal/Complex exp sequences Sinusoidal/Complex exp sequences



H ( e j )   h ( n )e
n  
 jn  x(n) i.e. input signal can be represented as the superposition
of complex exponentials

 y(n) i.e. output signal is then the superposition of responses to


1 
each complex exponential (i.e. multiply by H(ejw))
 H (e )e d
j j n
h(n) 
2 
 
1
X ( e j )   x ( n )e
n  
 j n
y (n) 
2  H (e
j
) X (e j )e j n d 


1
 X (e
j
x(n)  )e j n d Y ( e j )  H ( e j ) X ( e j )
2 

F.T. and the convergence F.T. and the convergence


 If x(n) is absolutely summable,
X ( e j )   x ( n )e
n  
 j n

1
  x(n)  
 X (e
j
x(n)  )e j n d n  
2 

Above equation of F.T. does not always converse. It is


X ( e j )   x ( n )e
n
 jn
  x ( n ) e  j n  
n
subject to variety of definitions and interpretations.
Converges uniformly to a continues function of ω

F.T. and the convergence F.T. and the convergence

 Absolutely summable sequences always has “finite energy”, But


Finite length sequences are absolutely summable. finite energy sequence is not necessarily absolutely summable.

So, F.T. converges uniformly and does exist.  Ex. x(n) = 1/n n≥1

 Here, one may employ a type of convergence in which the


series converges, to make the “Mean square error” as zero.
Infinite length sequences may or may not converge uniformly, i.e., F.T.
may or may not exist.

2
9/22/2022

Mean square error convergence Low pass filter

H LP ( e j )  1 0    c

2
 X (e
j
lim )  X k ( e j ) d   0 H LP ( e j )  0 c    
k  


1
H LP ( e j )

  c c 

Low pass filter L.P.F. and the convergence


hLP ( n )  ?
sin  c n
H ( e j )
hLP ( n ) 
1 n

Finite energy, non-summable sequence


 2   c c  2
sin  c n  jn
 ( 2   c ) ( 2   c )
hn
LP ( n ) e  jn  
n n
e

does not converge to HLP (ejw )for all values of w

Convergence in mean square sense


k
sin  c n  jn
H LP , k ( e j )  
n k n
e

 Convergence in mean square sense.

 Approximating HLP,k(ejw ) to HLP(ejw ) in a


mean square sense at point of discontinuity.
Ref. Textbook :S K Mitra

3
9/22/2022

LTI –FIR discrete time system LTI –IIR discrete time system
N2 N M
y (n)   h(k ) x(n  k )  d k y (n  k )   pk x(n  k )
k 0 k 0
k  N1

N M

 d Y ( e  )e   p k X ( e j ) e  j  k
N2 j  jk
Y ( e j )   h ( k ) X (e j )e  jk
k  N1
k 0
k
k 0
M
N2
Which is a rational function of (ejw )

Y (e ) j pe k
 j k

H (e )  j
 h( k )e  jk
H ( e j ) 
X ( e j )
 k 0
N

d
k  N1
k e  jk
Typically (but not necessarily) a polynomial in (ejw ) k 0

You might also like