Part III

Design

Filter Design Techniques

Discrete-time
filters filters
Discrete-time
Discrete

Discrete-time IIR filter

Specifications for DT filters

Specifications for DT filters in Log domain

A Design Example

Discrete-time IIR filter

Discrete-time IIR filter design is done using
analog filter techniques:
1. Analog IIR filter design methods have simple
closed form solutions;
2. Design examples have existed for years.
3. Direct design of IIR filters has traditionally
been avoided
4. Direct design of FIR filters is possible.

Discrete-time IIR filter Design Flow

Discrete-time IIR filter Design

1. Poles on the j axis in the s-plane correspond to

poles on the unit circle in the z-plane.
2. Poles in the left half of the s-plane correspond to
poles inside the unit circle in the z-plane.
Hence stable and causal continuous-time filters will
produce stable and causal discrete-time filters.

Traditional Analog Filter Design

Traditional Analog Filter Design

Butterworth Design

Butterworth Design

Chebyshev filters

Chebyshev filters

Chebyshev filters

Chebyshev filters

Elliptic filters

Example

Filter Design Techniques

Impulse Invariance
Bilinear Transformation


The design technique is as follows:

(1) Perform a partial fractions
expansion on H(s).

(2) Transform each pole into its transform equivalent.

(3) Combine the terms into a single
polynomial.

Impulse Invariance

Butterworth Design

To get a stable and causal filter,

choose Hc(s) to implement the poles in the left-hand plane.

Butterworth Filter

Butterworth Filter-Impulse Invariance

Butterworth Filter-Impulse Invariance

Example: Impulse Invariance

Take T = 1, value of T will not change the discrete-time

filter results.)

Bilinear Transformation

Bilinear Transform
To avoid aliasing, we need a one-to-one mapping
from the s-plane to the z-plane.

Bilinear Transform: Freq axis

Bilinear Transformation

Transformation is unaffected by
scaling. Consider inverse
transformation with scale factor
equal to unity
z = 1+ s
1 s

For s = o + j o
2
2

(1 + o ) + jo
(
1
+
)
+

2
o
o
z=
z =
2
2
(1 o ) jo
(1 o ) + o

and so

o = 0 z =1
o < 0 z <1
o > 0 z >1

Bilinear Transformation

Mapping of s-plane into the z-plane

Bilinear Transformation

Nonlinear mapping introduces a
distortion in the frequency axis called
frequency warping

Effect of warping shown below

Bilinear Transformation (Graphical Translation)

Bilinear Transform: Design Procedure

1. Perform frequency prewarp to obtain the

corresponding analog filter specs (pick any T)
2. Design the analog filter Hc(s) using any one of
the analog filter prototypes.
3. Transform Hc(s) to H(z).

Example

Bilinear Transform: Ex.

Bilinear Transform

FIR Filter Design

Windowing Principal

Windowing: Frequency Interpretation

Windowing Effects

Rectangular Window

Common Windows

Common window

Effect of Windowing

Windows Freq Domain

Other Windows in Feq Domain

Comparison

Kaiser Method

Kaiser

Kaiser

Kaiser

Marks McClellan Algo

Parks McClellan Algorithm

Butterworth Approx. in MATLAB

Butterworth Approximation

Chebyshev Approximation

Elliptic Approximation in MATLAB

Elliptic Approximation