Lecture 9

FIR and IIR Filter design using Matlab

2007/11/16

Prof. C.M Kyung

 FIR and IIR Filter

GOAL
 Linear-Time-Invariant (LTI) system and Impulse response

 z-Transform

 Characteristics of FIR and IIR filters

 Design procedure of FIR and IIR filters

 FIR and IIR Filter

LTI System
 Input (x(t) or x[n]) and Output (y(t) or y[n]) is defined first
 For a given system, it can be either LTI or non-LTI depending on how we
define the input and output.

 Linearity
 For arbitrary x1[n] and x2 [n] , the output of the system for x[n]  x1[n]  x2 [ n]
is the sum the output for x1[n] and x2 [n] . (superposition)

i.e. output{x1[n]  x2 [n]}  output{x1[n]}  output{x2 [n]}

 Time-invariance
 Time-shift in input results in time-shift in output by same amount for time-
invariant systems.
i.e. y[n]  output{x[n]}  y[n  m]  output{x[n  m]}
 FIR and IIR Filter

Impulse Response
 Definition
 Impulse response is the output of the system when impulse signal
 (t ) or  [n] is applied as the input of the system.

 Importance
 Impulse response “fully describes” the system if the system is LTI.
– Why and how?
– Why the impulse response CANNOT fully describe a non-LTI system?

 Fourier transform of the impulse response shows the characteristic of the

system is frequency domain.
 FIR and IIR Filter

z-Transform
 Definition 
 The z-transform of a sequence x[nis]defined as X ( z )  {x[n]}   x[n]z
n  
n

 Example
 For x[n]  a n u[n,]
  
1
X ( z )  ,{x[n]}   x[n]z
n  
n
  a u[n]z
n  
n n
  (az 1 ) n 
n 0 1  az 1
| z || a |

 Note
 z-transform is reduced to discrete-time Fourier transform (DTFT) if is
substituted by . z
j
 This means that ez-transform on the unit-circle on the complex plain is same as
DTFT.
 Laplace Transform CTFT ~ z-Transform  DTFT
 FIR and IIR Filter

Ideal frequency-selective filter
 A filter whose frequency response is unity over a certain frequency
range and zero for other frequencies.

 Frequency response of an ideal low-pass filter

Ideal Low pass filter

0.8
Frequency Response

0.6

0.4

0.2

-0.2
-3 -2 -1 0 1 2 3
frequency

 However, an ideal low-pass filter is noncausal.

 FIR and IIR Filter

FIR / IIR filter
 Definition
 If the length of the impulse response is finite, the filter is an FIR (finite impulse
response) filter. Otherwise, the filter is an IIR (infinite impulse response) filter.

 FIR
 Inherently BIBO (bounded-input, bounded-output) stable
 Nonzero pole does not exist in its transfer function
 Easy to implement
 Can be designed to have linear phase property

 IIR
 Sometimes unstable
 Nonzero pole exists in its transfer function
 Lower filter order than a corresponding FIR filter
 Usually have nonlinear phase property
 FIR and IIR Filter

Filter Design Procedure
 Design continuous-time IIR filter
 Obtain desired using Butterworth, Chebyshev methods
H (s )
 Convert it to discrete-time IIR filter using impulse invariance
 Impulse invariance :  , 
j
h[n]  Td hc (nTd ) H (e )  H c ( j ) |  | 
if , Td
H c ( j)  0 |  |  / Td
 Obtain discrete-time FIR filter by windowing the IIR filter
 Windowing :
Commonly used hwindows : rectangular, Bartlett, Hanning, Hamming,
FIR [ n ]  hIIR [ n ]w[ n]


Blackman, Kaiser, …
 However, windowing does not give the optimum solution and other approaches can be
used.
 FIR and IIR Filter

Frequency response of various filters
 FIR and IIR Filter

Problem Statements

 Design several types of FIR and IIR filters

 IIR – butterworth, chebyshev type1, chebyshev type 2, …
 FIR – using different windows ( Hamming, Hanning, Bartlett, … )

 Remove the noise in acoustic signal using the filters

 What are the differences between the filters ?

 Understand the effect of sampling frequency on the sampled

signal distortion (aliasing)
 FIR and IIR Filter

Experiment Requirements

 PC
 Matlab software (with signal processing toolbox)
 FIR and IIR Filter
 References

 Fundamentals of Signal & System using the web and matlab

- Edward W. Kamen, Bonnie S.
Heck

 Discrete-Time Signal Processing

- Alan V. Oppenheim, Ronald W.
Schafer

 http://www.mathworks.com/access/helpdesk/help/toolbox/signal/f
ilterde.html