International Journal of Engineering Applied Sciences and Technology, 2019

Vol. 4, Issue 3, ISSN No. 2455-2143, Pages 302-304

Published Online July 2019 in IJEAST (http://www.ijeast.com)

A COMPARATIVE STUDY OF WINDOWS FOR

SPECTRAL ANALYSIS
D.V.L.N.Sastry, B. Anil Kumar, P.Kameswar Rao, V.Laxmi, A.Jayalaxmi, J.Swathi
Assistant Professors ECE Department
Aditya institute of technology and Management, Tekkali

ABSTRACT - In this paper we presented a spectral II. WHY NEED WINDOWS IN SIGNAL
analysis of different windows designed based on cosine PROCESSING
functions. this paper includes comparative study of
different windows based on their spectral characteristics When converting the continuous signals into discrete signals
like side lobe attenuation (SLA) half bandwidth (HBW) using sampling before transforming into frequency domain
and side lobe fall of ratio (SLFOR).since windowing and transmitting there is a present of frequencies other than
method one of the best method in the design of linear sampled instants referred as spectral leakage.
phase FIR filters using in digital signal processing Windows are mathematical functions applied to signals
applications like audio processing, video processing, either continuous or sampled to reduce the spectral leakage.
speech processing and bio medical signal processing etc.
Windows plays major role in design of Finite Length III. BASIC WINDOW FUNCTIONS
digital filters for removing pikes in stop band and
improves pass band attenuation characteristics. And lot In this section we are presenting different presenting
of research also done on windows. Many windows are mathematical functions of windows used in the design of
developed based on their side lobe attenuation linear phase FIR filters.
characteristics. Optimisation techniques also applied in
the filed of window based FIR design, which shows III.I Bartlett window
better results compare to conventional methods. The mathematical expression for Bartlett window is
Here we proposed basic mathematical kernel functions to
develop window functions. ( )

Key words: Bartlett window, Hamming window,

Hanning window, proposed window functions. ----------------- (1)

I. INTRODUCTION III.II Hamming window

The mathematical expression for Hamming window is
Windowing technique is used in to process the signals
( ) ( )
described by Fourier Transform modelled[1].FIR filters are
designed by windowing method to remove the noises in = 0 otherwisw ------------- (2)
ECG signals[2]. Windows are mathematical functions
applied to both periodic signals and non periodic signals[3]. III.III Hanning window
Windowing based FIR filters are used in communication The mathematical expression for Hanning window is
systems to design of channel equalizers[4]. The magnitude
and phase responses of FIR filters are important for ( ) ( )
selection of windows while designing.[5] the side lobe = 0 otherwise ------------------------- (3)
attenuation of spectral characteristics is key factor in the
design of windows[6]. The sharper magnitude and liner IV. PROPOSED WINDOWS
phase response of FIR filters are obtained by windowing
technique[7]. The stability and filter coefficients of FIR The window functions are generally continuous cosine
filters are varied for different window functions[8]. Window functions implemented as a powers of integers of cosine
is a mathematical function allows the signal between the functions which are referred as cosine windows given as
interval and zero value outside the interval[9]. Hybrid
windows are designed which gives better results compare to
basic window functions[10]. ( ) [ ]
=0 otherwise ---------- (4)

302
 International Journal of Engineering Applied Sciences and Technology, 2019
Vol. 4, Issue 3, ISSN No. 2455-2143, Pages 302-304
Published Online July 2019 in IJEAST (http://www.ijeast.com)

But it is also confirmed that windows are not only Time domain Frequency domain
40
continuous but also developed from their second derivatives 1
20
called as Dirchlet Kernels. The Dirchlet kernels are located
0.8 0
about the central kernel which are zeros of central kernel. So

Magnitude (dB)
Amplitude
-20
the summation of kernels side lobes results the cancellation 0.6

-40
of side lobes because of opposite phases. so this partial 0.4
-60
structure of side lobes with three or more kernel functions 0.2
-80
results to construction of new windows which are described
-100
in mathematical function is 0
5 10 15 20 25 0 0.2 0.4 0.6 0.8
Samples Normalized Frequency ( rad/sample)

( ) Fig1: Spectral response of Bartlett window

Time domain Frequency domain

40
1
20

------(5) 0.8 0

Magnitude (dB)
-20
Where x=0,1,2,3----N-1

Amplitude
0.6
-40

The coefficients x determines n-tem window where n is the 0.4 -60

number of coefficients. So here we proposed the windows 0.2

-80

with 3 terms and 4 terms and observed their spectral -100

0 -120
responses with basic conventional windows. 5 10 15
Samples
20 25 0 0.2 0.4 0.6 0.8
Normalized Frequency ( rad/sample)

IV.I Proposed window1: It is a 3 term window and the Fig2: Spectral response of Hanning window
mathematical function is
Time domain Frequency domain
40

( ) 1
20

0.8 0

Magnitude (dB)
Amplitude

-20
0.6

-40
0.4
-60
------- (6) 0.2
-80

0 -100
5 10 15 20 25 0 0.2 0.4 0.6 0.8
IV.I Proposed window2: It is a 4 term window and the Samples Normalized Frequency ( rad/sample)

mathematical function is
Fig3: Spectral response of Hamming window

( )
Time domain Frequency domain
40
1
20

0.8 0
Magnitude (dB)
Amplitude

-20
0.6

-40
0.4
-60
---------------- (7) 0.2
-80
V. SIMULATION & RESULTS
0 -100
5 10 15 20 25 0 0.2 0.4 0.6 0.8
Samples Normalized Frequency ( rad/sample)

In this sections we developed simulation results for spectral

responses of Bartlett window, Hamming window, Hanning Fig4: Spectral response of proposed window1
window and two proposed windows. The simulation results
are obtained from MATLAB.

303
 International Journal of Engineering Applied Sciences and Technology, 2019
Vol. 4, Issue 3, ISSN No. 2455-2143, Pages 302-304
Published Online July 2019 in IJEAST (http://www.ijeast.com)

Time domain Frequency domain VII. REFERENCES

20
1

[1] Fredric J.harris, ‘’On the use of windows for Harmonic

0

0.8
Analysis with the discrete Fourier Transform’’, Proceedings
-20

Magnitude (dB)
Amplitude

-40
0.6
of the IEE vol:66, no:1,1978
-60
0.4 [2] Dvln Sastry,K.Srinadh gupta, V.mohan krushana, DSVS
0.2
-80
Nagendra reddy, ‘’ IMPROVED SNR OF ECG SIGNAL
-100
WITH NEW WINDOW- FIR DIGITAL FILTERS ‘’,
-120
0
5 10 15 20 25 0 International Journal of Advanced Research in Electronics
0.2 0.4 0.6 0.8
Samples Normalized Frequency ( rad/sample)
and Communication Engineering (IJARECE) Volume 1,
Issue 3, September 2012
Fig5: Spectral response of proposed window2
[3] Mitun Shil, Hrishi Rakshit, Hadaate Ullah,’’ An
Adjustable Window Function To Design An FIR Filter’’,
40
Frequency domain 978-1-5090-6004-7/@2017IEEE
20
Bartlett
Hanning
[4] Manira Khatun,’’ IMPLEMENT A NEW WINDOW
0
Hamming
proposed1
FUNCTION AND DESIGN FIR FILTERS BY USING
THIS NEW WINDOW’’, International Journal Of
Magnitude (dB)

-20 proposed2

-40 Engineering And Computer Science ISSN:2319-7242

-60
Volume- 3, Issue -3 , March, 2014 Page No. 4087-4090
-80
[5] Suhaib Ahmed,’’ Design Analysis of High Pass FIR
Filters Using Hanning, Bartlett and Kaiser Windows’’,
-100

-120
0 0.1 0.2 0.3 0.4 0.5 0.6
Normalized Frequency ( rad/sample)
0.7 0.8 0.9
International Journal of Advanced Research in Computer
Science and Software Engineering, Volume 2, Issue 11,
Fig6: Spectral response of all windows November 2012 ISSN: 2277 128X
[6] Suhaib Ahmed, Mudasir Bashir, Ashish Suri,’’ ‘’Low
Pass FIR Filter Design and Analysis Using Hamming,
Table1: Spectral characteristics of windows Blackman and Kaiser Windows’’, International Journal of
s.no Window Side lobe Half bandwidth Side Advanced Research in Electrical, Electronics and
attenuation HBW(db) lobe fall Instrumentation Engineering (An ISO 3297: 2007 Certified
SLA(dB) of ratio Organization) Vol. 3, Issue 4, April 2014
1 Bartlett -26.1 0.10156 17.28 [7] Girdhar chandra, Hemlal sahu, Pranay Kumar Rahi,’’
2 Hanning -31.5 0.10938 56.36 DESIGN OF LOWPASS FIR FILTER USING NUTTALL,
TAYLOR AND TUKEY WINDOW TECHNIQUES’’,
3 Hamming -41.2 0.10156 7.39
International Journal of Scientific Research Engineering &
4 Proposed1 -65 0.125 1.45 Technology (IJSRET), ISSN 2278 – 0882 Volume 6, Issue
5 Proposed2 -91.6 0.14844 12.03 6, June 2017
[8] Sonika Aggarwal, Aashish Gagneja, Aman Panghal
VI. CONCLUSION ‘’International Journal of Advanced Research in Computer
Science and Software Engineering’’, Volume 2, Issue 7,
We have examined some cosine function based windows July 2012 ISSN: 2277 128X
and proposed windows which satisfies their performance [9] Qeethara Al-Shayea , Muzhir Al-Ani,’’Efficient Window
towards spectral leakage and harmonic analysis. for multiple Approach of FIR Filter Design’’, IJCSNS International Journal of
speech detection through DFT the windows should exhibit Computer Science and Network Security, VOL.16 No.2,2016
concentrated central lobe width and narrow side lobes which [10]EPPILI JAYA, K.CHITAMBARA RAO, JAYA
results sharp attenuation in stop band of the filter. We have LAXMI. ANEM,’’ FIR FILTER DESIGN USING NEW
demonstrated the proposed windows which exhibits higher HYBRID WINDOW FUNCTIONS’’, International Journal
side lobe attenuation i.e -65dB for 3 term window (proposed of Science, Engineering and Technology Research
window1) and -91.6 dB for 4 term window (proposed (IJSETR), Volume 4, Issue 6, June 2015
window2). See Table1 for reference. And we also presented
spectral time and frequency domain analysis of all windows
shown from Fig1 to Fig6.

304