Digital Signal Processing(BEC-42)

Unit-3

Lecture-1
(FIR Filter Design)

Date:12/10/2020

Digital Signal Processing (BEC-42)

➢ In the FIR system, the impulse sequence is of finite duration,
i.e. it has a finite numbers of non –zero terms.
For example, the system with the impulse response

➢ Usually implemented using structures with no feedback(non

recursive structures- all poles)
➢ Suppose a system has the following difference equation with
x(n) and y(n) as input and output.

Digital Signal Processing (BEC-42)

Response of FIR filter depends upon only present and past input samples.

Digital Signal Processing (BEC-42)

Design Technique

Digital Signal Processing(BEC-42)

If h(nT) is an even sequence, then will be zero and if h(Nt) is
an odd sequence, then the will be zero.

Digital Signal Processing(BEC-42)

Digital Signal Processing(BEC-42)
Digital Signal Processing(BEC-42)
Digital Signal Processing(BEC-42)
Design Process:
➢ Decide whether is to be set equal to zero.
➢ Expand in a Fourier series.
➢ The unit pulse response is determined from Fourier
coefficient.

Digital Signal Processing(BEC-42)

Problem: Use the Fourier series method to design a low-pass
digital filter to approximate the ideal specifications given by

𝟏 𝒇𝒐𝒓 𝒇 ≤ 𝒇𝒑
𝑯 𝒆𝒋𝝎 = ൞ 𝑭
𝟎 𝒇𝒑 < 𝒇 ≤
𝟐
Where 𝒇𝒑 is pass-band frequency and F is sampling frequency.

Digital Signal Processing (BEC-42)

Digital Signal Processing (BEC-42)
Digital Signal Processing (BEC-42)
 Frequency Sampling Method

➢ In this method, a set of samples is determined from desired

frequency response
➢ IDFT is applied to get filter coefficients

𝑯𝒅 (𝒆𝒋𝝎 ) 𝑯𝒅 (𝒌) h(𝒏)

Frequency Sampling Inverse DFT

Digital Signal Processing (BEC-42)

Design Techniques:
➢ Type-I Design ( includes sample at frequency 𝜔 =0)
➢ Type-II Design ( exclude sample at frequency 𝜔 =0)

Type-I Design

Digital Signal Processing (BEC-42)

Then apply IDFT to get filter coefficient

For real valued h(n)

Digital Signal Processing(BEC-42)

and

Digital Signal Processing(BEC-42)

Type-II Design: Sample at frequency 𝜔 =0 is omitted

Digital Signal Processing(BEC-42)

Problem: A low-pass filter has the desired response as given
below

−𝒋𝟑𝝎
𝝅
𝒆 𝟎≤𝝎≤
𝑯𝒅 𝒆𝒋𝝎 = 𝟐
𝝅
𝟎 ≤𝝎≤𝝅
𝟐

Determine the filter coefficients h(n) for M=7 using type-I

frequency sampling method.

Digital Signal Processing (BEC-42)

Digital Signal Processing(BEC-42)
Digital Signal Processing(BEC-42)
Problem: Design a band pass filter with the following
specifications:
Sampling frequency, F= 8kHz
Cut off frequencies, 𝑓𝑐1 = 1 𝑘𝐻𝑧 𝑎𝑛𝑑 𝑓𝑐2 = 3 𝑘𝐻𝑧
Use frequency sampling method and determine the filter
coefficient for M=7.

Digital Signal Processing (BEC-42)

Digital Signal Processing (BEC-42)
Digital Signal Processing (BEC-42)
Problem: Design a Finite Impulse Response low-pass filter
with a cut-off frequency of 1 kHz and sampling rate of 4kHz
with 11 samples using Fourier series.

Digital Signal Processing (BEC-42)

Digital Signal Processing (BEC-42)
Digital Signal Processing(BEC-42)
Digital Signal Processing(BEC-42)
Digital Signal Processing(BEC-42)
Problem: Design an ideal filter with a frequency response

𝝅 𝝅
𝟏 − ≤𝝎≤
𝑯𝒅 𝒆𝒋𝝎 = 𝟐 𝟐
𝝅
𝟎 ≤ 𝝎 ≤𝝅
𝟐

Determine h(n) and H(z) for M=11 and plot the magnitude
response.

Digital Signal Processing (BEC-42)

Digital Signal Processing(BEC-42)
Digital Signal Processing(BEC-42)
Digital Signal Processing(BEC-42)
Digital Signal Processing(BEC-42)
Digital Signal Processing(BEC-42)
Problem: Determine the filter coefficients of a linear phase
FIR filter of length M=15 which has a symmetric unit
sample response and a frequency response that satisfies the
conditions

𝟐𝝅𝒌 𝟏 𝒌 = 𝟎, 𝟏, 𝟐, 𝟑
𝑯 =ቊ
𝟏𝟓 𝟎 𝒌 = 𝟒, 𝟓, 𝟔, 𝟕

Digital Signal Processing (BEC-42)

Solution: For odd M

Digital Signal Processing (BEC-42)

Digital Signal Processing (BEC-42)
Problem: (a) Use frequency sampling method to design an
𝜋
FIR low-pass filter 𝜔𝑐 = for M=15, (b) Repeat part (a) by
4
selecting an additional sample 𝐻(𝑘) =0.5 in the transitions
band.

Digital Signal Processing (BEC-42)

Solution: For odd M

Digital Signal Processing (BEC-42)

Digital Signal Processing (BEC-42)
(b) By introducing an additional sample in the transition band, the
magnitude of frequency response is given by

Digital Signal Processing (BEC-42)

Digital Signal Processing (BEC-42)
Digital Signal Processing (BEC-42)
 Windowing Techniques

➢ The desired frequency response of digital filter is given as

where impulse response( Fourier coefficients) of filter is given

as

➢ Impulse response is of infinite duration.

➢ Filter is non-causal and unrealizable.

Digital Signal Processing (BEC-42)

Digital Signal Processing (BEC-42)
Digital Signal Processing (BEC-42)
Desirable characteristics of window functions:

Digital Signal Processing (BEC-42)

 1

Digital Signal Processing(BEC-42)

Digital Signal Processing(BEC-42)
 2

From equation 1 and 2, the linear phase response of causal filter is

given by and the non-causal impulse response
has a zero phase shift.

Digital Signal Processing(BEC-42)

Problem: A low-pass filter is to be designed with the
following desired frequency response

𝝅 𝝅
𝒆−𝒋𝟐𝝎 − ≤𝝎≤
𝑯𝒅 𝒆𝒋𝝎 = 𝟒 𝟒
𝝅
𝟎, ≤ 𝝎 ≤𝝅
𝟒

Determine the filter coefficients ℎ𝑑 (n) if the window

1, 0≤𝑛≤4
function is defined as 𝑤 𝑛 = ቊ
0, 𝑜𝑡ℎ𝑒𝑟𝑤𝑖𝑠𝑒
Also, determine the frequency response H( 𝑒 𝑖𝜔 ) of the
designed filter.

Digital Signal Processing (BEC-42)

Digital Signal Processing(BEC-42)
Digital Signal Processing(BEC-42)
Digital Signal Processing (BEC-42)
 Windowing Technique

➢ Hamming Window Function:

Digital Signal Processing (BEC-42)

Digital Signal Processing (BEC-42)
Digital Signal Processing (BEC-42)
➢ Hanning Window Function:

Digital Signal Processing (BEC-42)

Problem: Design a filter with
−𝒋𝟑𝝎
𝝅 𝝅
𝒆 , − ≤𝝎≤
𝑯𝒅 𝒆−𝒋𝝎 = 𝟒 𝟒
𝝅
𝟎, ≤ 𝝎 ≤𝝅
𝟒

using a Hamming window with M=7.

Digital Signal Processing (BEC-42)

Digital Signal Processing(BEC-42)
Digital Signal Processing(BEC-42)
Digital Signal Processing(BEC-42)
Digital Signal Processing(BEC-42)
Digital Signal Processing(BEC-42)
Problem: Design an ideal high-pass filter with a frequency
response

𝝅
𝟏, ≤ 𝝎 ≤𝝅
𝑯𝒅 𝒆𝒋𝝎 = 𝟒
𝝅
𝟎, 𝝎 ≤
𝟒

Find h(n) and H(z) for M=11 using Hanning Window.

Digital Signal Processing (BEC-42)

Solution:

Digital Signal Processing(BEC-42)

Digital Signal Processing(BEC-42)
Digital Signal Processing(BEC-42)
Problem: Design a high pass filter using Hamming
window with a cut-off frequency of 1.2 rad and M=9.

Digital Signal Processing (BEC-42)

Digital Signal Processing (BEC-42)
Digital Signal Processing (BEC-42)
Digital Signal Processing (BEC-42)
Digital Signal Processing (BEC-42)
Digital Signal Processing (BEC-42)
Problem: Design a band-pass filter to pass frequencies in
the range 1-2 rad using Hanning window M=5.

Digital Signal Processing (BEC-42)

Digital Signal Processing (BEC-42)
Digital Signal Processing (BEC-42)
Digital Signal Processing (BEC-42)
Digital Signal Processing (BEC-42)
Blackman Window Function

Digital Signal Processing (BEC-42)

 Bartlett Window Function

Frequency-domain characteristics of window functions

Digital Signal Processing (BEC-42)

Digital Signal Processing (BEC-42)
Digital Signal Processing (BEC-42)
Magnitude Response

Digital Signal Processing (BEC-42)

Phase Response Digital Signal Processing (BEC-42)
Digital Signal Processing (BEC-42)
Digital Signal Processing (BEC-42)
Frequency-domain characteristics of window functions

➢ Main lobe width is inversely proportional to M.

➢ Increase in window length decreases the transition band.
➢ Minimum stopband attenuation is independent of M and function
of window.
➢ Min stopband attenuation has the max. main lobe width.
➢ For desired transition band, length of filter must be increased for
reducing the main lobe width.
Digital Signal Processing (BEC-42)
➢ A simple approximation to these functions have been developed
by Kaiser in terms of zeroth order modified Bessel functions of the
first kind.
Kaiser window starts with realistic demands:
• Allow ripples
• Allow certain transition widths
• As a results we get number of taps and window function.
➢ The Kaiser window function is given by

➢ The side lobe level can be controlled with respect to the main lobe
peak by varying a parameter, .
➢ The width of the main lobe can be varied by adjusting the length of
the filter.
Digital Signal Processing (BEC-42)
Digital Signal Processing (BEC-42)
where α is an independent variable determined by Kaiser. The
parameter β is expressed by

The modified Bessel function of the first kind, 𝐼0 (x), can be computed
from its power series expansion given by

The actual passband ripple (𝐴𝑃 ) and minimum stopband attenuation

(𝐴𝑆 ) are given by

Digital Signal Processing (BEC-42)

Digital Signal Processing (BEC-42)
Digital Signal Processing (BEC-42)
Digital Signal Processing (BEC-42)
Digital Signal Processing (BEC-42)
Digital Signal Processing (BEC-42)
Digital Signal Processing (BEC-42)
Digital Signal Processing (BEC-42)
Digital Signal Processing (BEC-42)
Problem: Design an FIR digital filter to approximate an ideal low-
pass filter with passband gain of unity, cut-off frequency of 850 Hz
and working at a sampling frequency of 𝑓𝑠 =5000 Hz. The length of
the impulse response should be 5. Use a rectangular window

Digital Signal Processing (BEC-42)

Digital Signal Processing (BEC-42)
Digital Signal Processing (BEC-42)
Problem: Design an ideal high-pass filter with a frequency
response

𝝅
𝟏, ≤ 𝝎 ≤𝝅
𝑯𝒅 𝒆𝒋𝝎 = 𝟒
𝝅
𝟎, 𝝎 ≤
𝟒

Find h(n) and H(z) for M=11 using (a) Hanning window (b)
Hanning window and plot the magnitude response.

Digital Signal Processing (BEC-42)

Digital Signal Processing(BEC-42)
Digital Signal Processing(BEC-42)
Digital Signal Processing(BEC-42)
Digital Signal Processing(BEC-42)
Digital Signal Processing(BEC-42)
Digital Signal Processing(BEC-42)
Digital Signal Processing(BEC-42)
Digital Signal Processing(BEC-42)
Digital Signal Processing(BEC-42)
Digital Signal Processing (BEC-42)
Digital Signal Processing (BEC-42)
Problem: Design a low-pass digital FIR filter using Kaiser window
satisfying the specifications given below.
Passband cut-off frequency, 𝑓𝑝 =150 Hz, stopband cut-off frequency,
𝑓𝑠 = 250 Hz, passband ripple, 𝐴𝑝 = 0.1 dB, stopband attenuation, 𝐴𝑠
= 40 dB and sampling frequency, F =1000 Hz.

Digital Signal Processing (BEC-42)