Lab.6+7.

Tutorial on IIR Filters

Draft Handouts

Definition

As discussed earlier, FIR filters are relatively easy to design. The textbox below summarizes
the characteristics of FIR filters as well as the most popular design techniques

One of the drawbacks of FIR filters is that they require a large filter order to meet some
design specifications. By using feedback, it is possible to meet a set of design specifications
with a far smaller filter order than a comparable FIR filter. This is the idea behind IIR filter
design. IIR filters get their name because their impulse response extends for an infinite period
of time. This is because they are recursive, i.e., they utilize feedback. Although they can be
implemented with fewer computations than FIR filters, IIR filters do not match the
performance achievable with FIR filters, and do not have linear phase. Textbox below
summarizes the IIR filter characteristics

EE462-Digital Signal Processing Lab Electrical Engineering Department Prof. Hazem Al-Otum 1
 Design of digital IIR filters is heavily dependent on that of their analog counterparts
because there are plenty of resources, works and straightforward design methods concerning
analog feedback filter design while there are hardly any for digital IIR filters. As a result,
usually, when a digital IIR filter is going to be implemented, an analog filter (e.g. Chebyshev ,
Butterworth, Elliptic, etc) is first designed and then is converted to a digital filter by applying
discretization techniques such as: 1)Bilinear transform or 2) Impulse invariance.

Transfer function derivation

where:
 is the feedforward filter order

 are the feedforward filter coefficients

 is the feedback filter order

 is the input signal

 is the output signal.

A more condensed form of the difference equation is:

which, when rearranged, becomes:

To find the transfer function of the filter, we first take the Z-transform of each side of the
above equation, where we use the time-shift property to obtain:

EE462-Digital Signal Processing Lab Electrical Engineering Department Prof. Hazem Al-Otum 2
We define the transfer function to be:

Considering that in most IIR filter designs coefficient =1, the IIR filter transfer function
takes the more traditional form:

note that the positive sign indicates negative feedback

Description of block diagram

A typical block diagram of an IIR filter looks like the following. The block is a unit delay.
The coefficients and number of feedback/feedforward paths are implementation-dependent.

Fig. Simple IIR filter block diagram (Direct Form I)

EE462-Digital Signal Processing Lab Electrical Engineering Department Prof. Hazem Al-Otum 3
Stability

The transfer function allows us to judge whether or not a system is stable. To be specific, the
stability criteria requires that the ROC of the system includes the unit circle. For example, for a
causal system, all poles of the transfer function have to have an absolute value smaller than
one. In other words, all poles must be located within a unit circle in the -plane.

The poles are defined as the values of which make the denominator of equal to 0:

Clearly, if then the poles are not located at the origin of the z-plane. This is in
contrast to the FIR filter where all poles are located at the origin, and is therefore always stable.

Note: IIR filters are sometimes preferred over FIR filters because an IIR filter can achieve a
much sharper transition region roll-off than FIR filter of the same order.

Example
Let the transfer function of a filter H be

with ROC and

which has a pole at a, is stable and causal. The time-domain impulse response is

which is non-zero for .

IIR Common Analog Filters

A popular method for IIR filter design is to first design the analog-equivalent filter and then
mathematically transform the transfer function H(s) into the z-domain, H(z). Multiple pole
designs are implemented using cascaded biquad sections. The most popular analog filters are
the Butterworth, Chebyshev, Elliptical, and Bessel. There are many CAD programs available
to generate the Laplace transform, H(s), for these filters.

 Butterworth (also called maximally flat) has no ripple in the passband or stopband and
has monotonic response in both regions. Butterworth filters are maximally-flat IIR

EE462-Digital Signal Processing Lab Electrical Engineering Department Prof. Hazem Al-Otum 4
 filters. For this reason, the only design parameters are the cutoff frequency and the filter
order.

Example: Design a 7th order Butterworth filter with a 3 dB point at 0.3pi.

Hf = fdesign.lowpass('N,F3db',7,0.3);
Hb = design(Hf,'butter');

Fig. General Filter description

Fig. Example Result

 Chebyshev Type 1 filter has a faster rolloff than the Butterworth (for the same number
of poles) and has ripple in the passband. Chebyshev type I filters can attain a smaller
transition width than a Butterworth filter of the same order by allowing for ripples in
the passband of the filter. The stopband is, as with Butterworth filters, maximally flat.
For a given filter order, the trade-off is thus between passband ripple and transition
width.

Example: Compare the 7th-order Butterworth filter from previous examples with a
7th-order Chebyshev type I filter with 1 dB of peak-to-peak passband ripple.

EE462-Digital Signal Processing Lab Electrical Engineering Department Prof. Hazem Al-Otum 5
  The Chebyschev Type 2 filter is rarely used, but has ripple in the stopband rather than
the passband.

Example Design a 6th order filter with a 3-dB point of 0.45pi. The filter must
have an attenuation of at least 80 dB at frequencies above 0.75pi and the passband
ripple must not exceed 0.8 dB.

Hf1 = fdesign.lowpass('N,F3db',6,0.45);
Hf2 = fdesign.lowpass('N,F3db,Ap',6,0.45,0.8);
Hf3 = fdesign.lowpass('N,F3db,Ast',6,0.45,80);
Hb = design(Hf1,'butter');
Hc1 = design(Hf2,'cheby1');
Hc2 = design(Hf3,'cheby2');

EE462-Digital Signal Processing Lab Electrical Engineering Department Prof. Hazem Al-Otum 6
The three designs are shown in Figure. Only the Chebyshev type II filter reaches the required
attenuation of 80 dB by 0.75pi. Also, Chebyshev type II, the latter’s group-delay1 is smaller
than the former’s for most of the passband. Even though all three designs in the previous
example are of 6th order, the Chebyshev type II implementation actually requires more
multipliers

Methods for IIR design

Traditionally, IIR filter design is based on the concept of transforming a continuous- time, or
analog, design into the discrete - time domain. In impulse invariance method, the impulse
response of the digital filter is the samples of the impulse response of the continuous-time
filter:

h[n]=T h[nT]

where T represents the sampling interval.

This method is based on the concept of mapping each s-plane pole of the continuous
time filter to a corresponding z-plane pole using the substitution (1-epktsz-1) for ( s + pk ) in
H(s). This can be achieved by several different means. Partial fraction expansion of H (s ) and
substitution of (1-e ktsz-1) for ( s + pk ) is a direct method to do.

B. Bilinear Transform Method of Digital Filter Implementation.

The bilinear transform method of converting an analog filter design to discrete time is
relatively straightforward, often involving less algebraic manipulation than the impulse
invariant method. It is achieved by making the substitution:

1
 Group delay Characterizes the phase distorsion (waveform distorsion) introduced by the filter.

EE462-Digital Signal Processing Lab Electrical Engineering Department Prof. Hazem Al-Otum 7
In H (s), where T is the sampling period of the digital filter; that is,

The concept behind the bilinear transform is that of compressing the frequency response of an
analog filters design such that its response over the entire range of frequencies from zero to
infinity is mapped into the frequency range zero to half the sampling frequency of the digital
filter. This may be represented by:

As a result of the frequency warping inherent in the bilinear transform, the cutoff
frequency of the discrete-time filter obtained is not equal to the cutoff frequency of the analog
filter. A technique called pre-warping the prototype analog design (used by default in the
MATLAB filter design and analysis tool fdatool) can be used in such a way that the bilinear
transform maps an analog frequency wA=wc, in the range 0 to ws/2 , to exactly the same
digital frequency the same digital frequency wD=wc. This technique is based on the selection
of T according to T= 2 tan(wc/ ws) / wc.

EE462-Digital Signal Processing Lab Electrical Engineering Department Prof. Hazem Al-Otum 8