A survey report for performance analysis of finite

IJRET: International Journal of Research in Engineering and Technology eISSN: 2319-1163 | pISSN: 2321-7308
_______________________________________________________________________________________
Volume: 02 Issue: 12 | Dec-2013, Available @ http://www.ijret.org 265
A SURVEY REPORT FOR PERFORMANCE ANALYSIS OF FINITE
IMPULSE RESPONSE DIGITAL FILTER BY USING DIFFERENT
WINDOW TECHNIQUES
Jannatul Ferdous1
, Md. Sujan Ali 2
1,2
Assistant Professor, Department of Computer Science and Engineering, Jatiya Kabi Kazi Nazrul Islam University,
Trishal, Mymensingh-2220, Bangladesh
mjannatul@gmail.com, msujanali@gmail.com
Abstract
In the field of signal processing and communication, digital filter plays pivotal role. Digital FIR filter designed by different window
techniques perform better for reducing noise from signal. In this paper, we take an overview of various window methods such as
Rectangular, Triangular, Hanning, Hamming, Blackman, Kaiser and some modified windows. The magnitude responses are
demonstrated for different design methods at particular cut off frequency and different filter order. All these technique have their
merits and demerits. In this paper, we studied various techniques proposed earlier in literature for noise reduction from signal. This
paper also provides comparative study of various filters using different window. It has been seen that Kaiser window is better for
noise free signal.
Keywords: FIR filter, Rectangular window, Bartlett window, Hanning window, Hamming window, Blackman window
and Kaiser window etc.
--------------------------------------------------------------------***------------------------------------------------------------------
1. INTRODUCTION
In Linear Time Invariant (LTI) DSP systems, digital filters are
significant term to modify the frequency characteristics of the
input signal x(n) to meet certain specific design requirements.
Digital filters are broadly used for its certain advantages over
Analog filters. Digital filters is better than analog filters to
accomplish much improved signal to noise ratios. Finite
Impulse Response (FIR) and Infinite Impulse Response (IIR)
filters are the classification of digital filter. Depends on the
nature of the difficulty and requirement of the desired
frequency response, FIR or IIR digital filter is preferred. Flat
frequency response, magnitude in pass band is frequently
attractive for designing a digital filter. Linear phase is also a
vital enviable characteristic of the filter. A linear phase with
integer slope pilot to a simple delay in the time domain, and it
decreases smallest phase distortion in the frequency domain.
The FIR filter of length M is expressed by the convolution of
unit sample response h(n) of the system and is represented by
the following equation.
( ) ( )knxkhny
M
k
−= ∑
−
=
.)(
1
0
Thus the coefficient of FIR filter is indicated by the impulse
response of the filter. FIR filters specification include the
maximum tolerable pass band ripple, maximum tolerable stop
band ripple, pass band edge frequency and stop band edge
frequency [1].
It is a simple and quick technique to design a FIR filter using
windowing. In the window method, the desired frequency
response specification Hd(w), corresponding unit sample
response hd(n) is determined using the following relation:
( ) ( ) dwewHnh jwn
dd ∫−
=
π
ππ2
1
( ) jwn
n
dd enhwH −
∞
−∞=
∑= .)(re,whe
In general, unit sample response hd(n) obtained from the
above relation is infinite in duration, so it must be truncated at
some point say n = M-1 to yield an FIR filter of length M (i.e.
0 to M-1). Then hd(n) the impulse response of a preferred FIR
filter is given by:

_______________________________________________________________________________________
( )
( )
10
0
−≤≤





=
Mn
otherwise
n
d
h
nh
The Frequency response of the desired FIR filter is
( ) jwn
M
n
d
jw
d enheH −
−
=
∑= .)(
1
0
Direct truncation of hd(n) to M terms to obtain h(n) leads to
the Gibbs phenomenon effect which manifests itself as a fixed
percentage overshoot and ripple before and after an
approximated discontinuity in the frequency response due to
the non-uniform convergence of the Fourier series at a
discontinuity.
Thus the frequency response obtained by using above equation
restrains ripples in the frequency domain. In order to lessen
the ripples, hd(n) is multiplied by a window function ( )nw
whose duration is finite. And it satisfies the desired magnitude
response which is following.
The outline of filter specifications is depend upon the desired
normalized frequencies (wc, wc1, wc2), transition width and
stopband attenuation. Both the window function and filter
order are précised according to these constraints [2]. Most
frequently used windows are:
1. Rectangular window:
( ) { 101
0
−≤≤
=
Mn
otherwise
R nw
2. Bartlett (Triangular) window:
( )





=
−≤≤
−
−
−
− 10
1
2
1
2
1
0
Mn
M
M
n
otherwise
B nw
3. Hanning window:
Casual:
( )



=
−≤≤





−
− 10
1
2
cos5.05.0
0
Mn
M
n
otherwise
Hann nw
π
Non-Casual:
( )



=
−
≤





−
+
2
1
0
1
2
cos5.05.0
0
M
nfor
M
n
otherwise
Hann nw
π
4. Hamming window:
Casual:
( )



=
−≤≤





−
− 10
1
2
cos46.054.0
0
Mn
M
n
otherwise
H nw
π
Non-Casual:
( )



=
−
≤





−
+
2
1
0
1
2
cos46.054.0
0
M
nfor
M
n
otherwise
H nw
π
5. Blackman window:
Casual:
( )



=
−≤≤





−
+





−
− 10
1
4
cos08.0
1
2
cos5.042.0
0
Mn
M
n
M
n
otherwise
nw
ππ
β
Non-Casual:
( )



=
−
≤
−
+
−
+ 











2
1
0
1
4
cos08.0
1
2
cos5.042.0
0
M
nfor
M
n
M
n
otherwise
nw
ππ
β
( ) ( ) ( )nwnhnh d=

_______________________________________________________________________________________
6. Kaiser window:
( )







=
−≤≤













 −













 −
−−




 −
10
1
2
1
2
1
0
0
22
0
Mn
M
I
M
n
M
I
otherwise
K nw α
α
These window methods are available for the design of filters
[3][4]. In this paper we study the merits and demerits of these
different approaches for performance evaluation of FIR filter.
2. LITERATURE REVIEW
For the performance of FIR digital filters, several of window
techniques have been introduced. The DSP application
demands high speed and low power digital filters. In order to
meet these requirements, the order of the digital filter must be
kept as small as possible. There are various sophisticated
Computer Aided Design tools are available to make the digital
filter fast and power efficient. Filter design and analysis tool
(FDA) is one of the Computer Aided Design tool available
with MATLAB which enables design of the digital filter
blocks faster and more accurate. Prof.Gopal et.al. analyzed the
performance of low pass FIR filter design methods for
achieving minimum order [1]. In their work, they described
the diverse technique for FIR filter in shortly. Among various
technique, window technique showed advantageous for its
simplicity. This method has been preferable by well defined
equations for calculating window coefficient. For low pass
filter design, Kaiser window offers very low order to meet
given specification than others (Hannig, Hamming and
Blackman). According to them, the best digital filter design
results comes from the windowing design technique (Kaiser
window) which has parameter α that allows adjustment
between the overshoot reduction and transition region width
spreading.
Era Singhal worked on design technique of low pass FIR filter
using various window methods actually raised cosine
windows. The stability, filter order and the filter coefficients
for different window methods (Rectangular, Hamming,
Hanning and Blackman) were demonstrated. To understand
this technique she used window based filter equation. In this
study researcher also provide the relation between frequency
and magnitude for magnitude response and phase response. In
her paper for simulation, she was used filter order=20,
sampling frequency=48000Hz, Cut-off frequency = 10800Hz.
Along with divergent window methods it has been seen that
filter using hamming window method is better than others.
And the researcher finally proved that hamming window
provide better result because its less ripples in passband,
narrowest mainlobe and sidelobes, more stability and has a
linear phase as compared to other window techniques [2].
Sonika Gupta, et. al. provided the information that FIR filter
design by using hamming is stable as compare to rectangular
and hanning windows. In their paper, they have used filter
order =61.They have shown the magnitude and phase response
of the three windows. Ripples in pass band are less in
hamming as compare to other two techniques and it has high
stability linear phase as compare to rectangular and hanning
windows [5].
Akhilesh Chandra Bhatnagar, et. al. prepared a report is based
upon the performance of different windows. The Kaiser
window has been used to design a better FIR filter in terms of
ripple ratio, side-lobe roll off ratio and main-lobe width. They
compared the spectral characteristic of Hamming window and
Kaiser Window. A modification is also introduced in
Hamming window to perform better in terms of ripple ratio as
compared to Kaiser using advanced Peak windowing method.
They proposed an advanced peak windowing method (APW),
which effectively suppresses the peak signals to the desired
threshold level in case those successive peaks occur within a
half of the window length. In their work, they used computer
simulations. According to the simulation, they have seen that
better performance can be achieved by the Hamming window
over the Kaiser Window spectral characteristics. But
Hamming window gives higher ripple ratio as compared to
Kaiser window. They also overcome this drawback using cosh
expression with hamming window. For reducing the
complexity, they introduced a modified window for Hamming
and Kaiser window which have a new adjusting parameter to
increases the main lobe width and reduces the side lobe ratio
and ripple ratio [6].
Saurabh Singh Rajput, et. al. implemented a low-pass FIR
filter using Blackman window on the basis of an efficient
adjustable window function. Regarding their work, the
amplitude of main lobe and side lobe for a fixed bandwidth
varied by changing the value of α of this function. They
developed an adjustable window function and this window is
referred to as Blackman window for α = 0.16 and Hanning
window for α= 0.
( )



=
−≤≤
−
+
−
−
−


















10
1
4
cos
21
2
cos5.0
2
1
0
Mn
M
n
M
n
otherwise
nw
παπα
By selecting the value of α in the frequency spectrum, this
window method provides higher side lobe attenuation compare
to Hanning and Hamming window and the width of the main-
lobe is slightly greater than Hamming window function. In
some applications such as FFT, signal processing and

_______________________________________________________________________________________
measurement where higher side lobe attenuation is required
compared to Hamming window this type of filters can play
very significant role. Although the Blackman window function
has a wider main lobe width but it used for these types of
applications [7].
Suhaib Ahmed analyzed a design technique of high pass FIR
filters using Hanning, Bartlett and Kaiser Windows. He
demonstrated the magnitude and phase responses for different
methods at particular cut off frequency (0.4 rad/s) and filter
order (61). He has been shown that degree of flatness varies
with the length of the filter. From the paper, the researcher
concluded that for a high pass FIR filter, Kaiser Window is
suitable if the transition band is of utmost importance but if he
measured other parameters then Hanning window enhanced
pass band attenuation. It is commonly found that Hanning
window considered as a stable window as it provides good
attenuation and a better transition from stop band to pass band
than Bartlett and Kaiser windows [8].
Yadwinder Kumar, et. al. compared the performance of high
pass FIR filters using various windowing methods
(Rectangular, Hamming, Blackman and Kaiser Windows).
They established only the magnitude responses of different
windows at particular cut off frequency and a variety of filter
order. This paper briefly describes the window and optimal
techniques for design FIR filter. From the discussion, the
availability of well defined equations for calculating window
coefficient has made window method preferable but it has a
few design flexibility. Among various windows it has been
seen Kaiser window perform good magnitude response than
others at different filter order [9].
Ankan Bhattacharya in his paper presented a modified
window function for Finite Impulse Response (FIR) filter
design with an improved frequency response. He generated a
window function as like as w(n) = a – b cos(2π(n+0.1)/N) + c
cos(4π(n+0.3)/N) – d cos(6π(n+0.5)/N) + e cos(8π(n+0.7)/N)
– f cos(10π(n+0.9)/N) where, 0 ≤ n ≤N . And it has been
used to compute the frequency response of various types of
FIR filter i.e. high pass, low pass, band pass and band rejected
types. In perspective of stop band attenuation and presence of
ripples he also compared the frequency responses of newly
generated window function with other existing windows. The
compared result was justified by relevant simulations and
found improved performance. According to quantized and
generalized experimental results, the researcher also contrasts
the frequency response of the modified window with
Hamming window. For low pass filter response, modified
window function comparatively much more effective
attenuation than that using Hamming window. It is observe
that it has resulted in much more reduced side lobes in low
pass filter response. From the paper it is clear that the presence
of stop band ripples is excessively high in case of Hamming
window response which can be considered as a major
drawback. For high pass filter, band pass and band stop filter
response, presence of ripple effect much less in the modified
window response than Hamming window [10].
Ankan Bhattacharya worked on another paper and presented a
modified window function for FIR filter. According to
quantized and generalized experimental results, the researcher
also contrasts the frequency response of the modified window
with Bartlett-Hann window. For low pass filter response,
modified window function has much more effective
attenuation and reduced side lobes than Bartlett-Hann
window. But it has a major drawback which is low stop band
attenuation. For high pass, band pass and band stop filter
response, presence of ripple effect much less in the modified
window response than Bartlett-Hann window [11].
Subhadeep Chakraborty worked on a comparative study
between Blackman and hamming window. From the output of
the FIR using this two window, it is clear that Blackman
window technique is more powerful, perfect and useful than
Hamming window. Throughout the paper it was also shown
Blackman window has more terms compared to the Hamming
window. More term in calculation indicates more accuracy in
result. There are extra cosine terms in Blackman function.
This extra term reduces the side lobes and hence increases less
losing power. For FIR lowpass filter design, there are four
sidelobes exists in Hamming window but for same
specification and filter order, in Blackman window, there
exists only two sidelobes. Similar thing found in other filters
(high pass filter, band pass and band stop filter) and for this
reason, the result and simulation both are more accurate in
Blackman window [12].
3. FUTURE ENHANCEMENT AND CONCLUSION
There are various window techniques generated for analysis
the performance of FIR filters. Every method has its own
merits and demerits. For example, rectangular window, as the
window is made longer the main lobe becomes narrower and
higher, and the side lobe become more concentrated around
w=0. The main lobe width for the rectangular window is equal
to 4π/M. The width of the transition region depends on width
of the main lobe. As the filter length M increases, the main
lobe becomes narrower decreasing the width of the transition
region. The convolution of the desire response and side lobes
give rise to the ripples in both passband and stopband. The
amplitude of the ripples is dictated by the amplitude of the
sidelobes. For the rectangular window, the amplitude of the
sidelobes is unaffected by the length of the window. So
increase the length M will not reduce the ripples, but increase
its frequency. According to Gibbs phenomenon, maximum
ripples occurs just before and after the transition band. So, the
Gibbs phenomenon can be reduce by using a sudden and
unexpected truncation of filter coefficients.

_______________________________________________________________________________________
If we use Bartlett window then we reduce this problem.
Because the main lobe width is 8π/M or twice that of the
rectangular window. The triangular window has two drawback
over rectangular window, one is the transition region is high
and another attenuation in stopband is less. Because of these
characteristics , this window is not usually a good choice and
it is not popular for FIR filter design.
Consider the Hanning window which transition region is twice
in the lowpass filter. It produces small ripples in both
passband and stopband. But at higher frequencies it generates
greater stopband attenuation. However, at higher frequencies
the Hamming window bring into being low stopband
attenuation compared to of Hanning window. Because the
Hamming window generates less oscillation in the sidelobes
than the Hanning window, for the same main lobe width the
Hamming window is generally preferred.
For the Blackmann and some modified window, the additional
cosine term (compared with the Hamming and Hanning
windows ) reduces the sidelobes, but incresses the main lobe
width to 12π/M. So, the cost increased in transition width.
On the other hand, the Kaiser window is superior to other
windows, because, for given specifications its transition width
is always small. By varying the parameter α the desire
sidelobe level and main lobe peak can be achived. Further the
main lobe width can be varied by varying the length M. That
is why Kaiser window is the favorite window for designing
FIR digital filter.
The future work is develop a modified window technique
suitable for designing of filters with a given magnitude
response and reduce the noise of signal. It has been observed
that the modified window function indeed produces much
more effective type of frequency responses compared to the
other window. The future enhancement is to design a filter
which provide accurate type of responses with reduced side
lobes and comparatively less pass-band and stop-band ripples
and frequency responses of the other common window types
will be compared with the modified one.
REFERENCES
[1]. Prof.GopalS.Gawande,Dr.K.B.Khanchandani,T.P.Mar
ode “performance analysis of FIR digital filter design
techniques” International Journal of computing &
Corporate research volume 2 issue 1 January 2012.
[2]. Era Singhal “Performance Analysis of Finite Impulse
Response(FIR) Filter Design Using Various Window
Methods”; International Journal of Scientific
Research Engineering & Technology (IJSRET);
Volume 1 Issue 5 pp 018-021 August 2012; ISSN
2278 – 0882.
[3]. J.G. Proakis and D.G. Manolakis; “Digital Signal
Processing: Principles, Algorithms and Applications,”
Pearson Prentice-Hall, fourth edition, 2011.
[4]. “Digital Signal Processing”;P.Ramesh Babu,
“Scitech, third edition, 2007.
[5]. Sonika Gupta, Aman Panghal Performance
“PerformanceAnalysis of FIR Filter Design by Using
Rectangular, Hanning and Hamming Windows
Methods” International Journal of Advanced Research
in Computer Science and Software Engineering
Volume 2, Issue 6, June 2012.
[6]. Akhilesh Chandra Bhatnagar, R. L. Sharma, Rajesh
Kumar, “Analysis of Hammaing window using
Advance Peak Windowing” International Journal of
Scientific Research Engineering & Technology
(IJSRET); Volume 1 Issue 4 pp 015-020 July 2012;
ISSN 2278 – 0882.
[7]. Saurabh Singh Rajput, Dr S.S.Bhadauria,
“Implementation of FIR filters using efficient window
function and its application in filtering a speech
signal” International Journal of Electrical, Electronics
and Mechanical Controls (IJEEMC),vol.1,1November
2012.
[8]. Suhaib Ahmed, “Design Analysis of High Pass FIR
Filters Using Hanning, Bartlett and Kaiser Windows”
International Journal of Advanced Research in
Computer Science and Software Engineering; Volume
2, Issue 11, November 2012.
[9]. Yadwinder Kumar, Er.Gurpreet Singh Walia
“Review: Performance Analysis of FIR Digital High
Pass Filters” International Journal of Engineering
Research and Applications (IJERA) Vol. 3, Issue 2, ,
pp.1012-1015, March -April 2013; ISSN: 2248-9622
[10]. Ankan Bhattacharya; “A modified window function for
FIR filter design with an improved frequency response
and its comparison with the Hamming window”;
ISSN: 2278 – 7798; International Journal of Science,
Engineering and Technology Research (IJSETR);
Volume 2,Issue 5, May 2013.
[11]. Ankan Bhattacharya; “A modified window function for
FIR filter design with an improvedfrequency response
and its comparison with the Bartlett-Hann window”;
ISSN: 2278 – 7798; International Journal of Science,
Engineering and Technology Research (IJSETR);

_______________________________________________________________________________________
Volume 2,Issue6, June 2013.
[12]. Subhadeep Chakraborty. “Advantages of Blackman
Window over Hamming Window Method for designing
FIR Filter” International Journal of Computer Science
& Engineering Technology (IJCSET); Vol. 4 No. 08
Aug 2013; ISSN : 2229-3345
BIOGRAPHIES
Jannatul Ferdous: She achieved her
B.Sc.(Hons.) and M.Sc. degree in Applied
Physics, Electronics & Communication
Engineering from Islamic University,
Kushtia, Bangladesh. At present, she is an
PhD researcher in the Department of
Computer Science & Engineering,
University of Rajshahi, Rajshahi, Bangladesh. She has nearly
eight years teaching experience in electronics and
communication related subjects. She has published two
research paper in International Conference Proceedings and
eight journal articles in leading national and International
Journals. Her research area in the field of Signal Processing.
Md. Sujan Ali: He received his
B.Sc.(Hons.) and M.Sc. degree in Applied
Physics, Electronics & Communication
Engineering from Islamic University,
Kushtia, Bangladesh. He is working as an
Assistant Professor in the Department of
Computer Science and Engineering, Jatiya
Kabi Kazi Nazrul Islam University, Mymensingh, Bangladesh.
Currently he is persuing PhD degree on Brain Signal
Processing under the faculty of Engineering, Department of
Computer Science & Engineering, University of Rajshahi,
Rajshahi, Bangladesh.

A survey report for performance analysis of finite

More Related Content

What's hot (18)

Viewers also liked (20)

Similar to A survey report for performance analysis of finite (20)

More from eSAT Publishing House (20)

Recently uploaded (20)

A survey report for performance analysis of finite