Signal Processing: An International Journal (SPIJ)

---

SIGNAL PROCESSING: AN INTERNATIONAL
JOURNAL (SPIJ)







VOLUME 6, ISSUE 1, 2012

EDITED BY
DR. NABEEL TAHIR








ISSN (Online): 1985-2339
International Journal of Computer Science and Security is published both in traditional paper form
and in Internet. This journal is published at the website http://www.cscjournals.org, maintained by
Computer Science Journals (CSC Journals), Malaysia.


SPIJ Journal is a part of CSC Publishers
Computer Science Journals
http://www.cscjournals.org

---

SIGNAL PROCESSING: AN INTERNATIONAL JOURNAL (SPIJ)
Book: Volume 6, Issue 1, February 2012
Publishing Date: 21-02-2012
ISSN (Online): 1985-2339

This work is subjected to copyright. All rights are reserved whether the whole or
part of the material is concerned, specifically the rights of translation, reprinting,
re-use of illusions, recitation, broadcasting, reproduction on microfilms or in any
other way, and storage in data banks. Duplication of this publication of parts
thereof is permitted only under the provision of the copyright law 1965, in its
current version, and permission of use must always be obtained from CSC
Publishers.



SPIJ Journal is a part of CSC Publishers
http://www.cscjournals.org

(c) SPIJ Journal
Published in Malaysia

Typesetting: Camera-ready by author, data conversation by CSC Publishing Services - CSC Journals,
Malaysia



CSC Publishers, 2012

---

EDITORIAL PREFACE

This is first issue of volume six of the Signal Processing: An International Journal (SPIJ). SPIJ is
an International refereed journal for publication of current research in signal processing
technologies. SPIJ publishes research papers dealing primarily with the technological aspects of
signal processing (analogue and digital) in new and emerging technologies. Publications of SPIJ
are beneficial for researchers, academics, scholars, advanced students, practitioners, and those
seeking an update on current experience, state of the art research theories and future prospects
in relation to computer science in general but specific to computer security studies. Some
important topics covers by SPIJ are Signal Filtering, Signal Processing Systems, Signal
Processing Technology and Signal Theory etc.

The initial efforts helped to shape the editorial policy and to sharpen the focus of the journal.
Starting with volume 6, 2012, SPIJ appears in more focused issues. Besides normal publications,
SPIJ intend to organized special issues on more focused topics. Each special issue will have a
designated editor (editors) - either member of the editorial board or another recognized specialist
in the respective field.

This journal publishes new dissertations and state of the art research to target its readership that
not only includes researchers, industrialists and scientist but also advanced students and
practitioners. The aim of SPIJ is to publish research which is not only technically proficient, but
contains innovation or information for our international readers. In order to position SPIJ as one of
the top International journal in signal processing, a group of highly valuable and senior
International scholars are serving its Editorial Board who ensures that each issue must publish
qualitative research articles from International research communities relevant to signal processing
fields.

SPIJ editors understand that how much it is important for authors and researchers to have their
work published with a minimum delay after submission of their papers. They also strongly believe
that the direct communication between the editors and authors are important for the welfare,
quality and wellbeing of the Journal and its readers. Therefore, all activities from paper
submission to paper publication are controlled through electronic systems that include electronic
submission, editorial panel and review system that ensures rapid decision with least delays in the
publication processes.

To build its international reputation, we are disseminating the publication information through
Google Books, Google Scholar, Directory of Open Access Journals (DOAJ), Open J Gate,
ScientificCommons, Docstoc and many more. Our International Editors are working on
establishing ISI listing and a good impact factor for SPIJ. We would like to remind you that the
success of our journal depends directly on the number of quality articles submitted for review.
Accordingly, we would like to request your participation by submitting quality manuscripts for
review and encouraging your colleagues to submit quality manuscripts for review. One of the
great benefits we can provide to our prospective authors is the mentoring nature of our review
process. SPIJ provides authors with high quality, helpful reviews that are shaped to assist authors
in improving their manuscripts.

Editorial Board Members
Signal Processing: An International Journal (SPIJ)

---

EDITORIAL BOARD

EDITOR-in-CHIEF (EiC)

Dr Saif alZahir
University of N. British Columbia (Canada)



ASSOCIATE EDITORS (AEiCs)


Professor. Wilmar Hernandez
Universidad Politecnica de Madrid
Spain

Dr Tao WANG
Universite Catholique de Louvain
Belgium

Dr Francis F. Li
The University of Salford
United Kingdom


EDITORIAL BOARD MEMBERS (EBMs)


Dr Thomas Yang
Embry-Riddle Aeronautical University
United States of America

Dr Jan Jurjens
University Dortmund
Germany

Dr Jyoti Singhai
Maulana Azad National institute of Technology
India

Assistant Professor Weimin Huang
Memorial University
Canada

Dr Lihong Zhang
Memorial University
Canada

Dr Bing-Zhao Li
Beijing Institute of Technology
China

---

TABLE OF CONTENTS




Volume 6, Issue 1, February 2012



Pages
1 - 21
Efficient Spectral Analysis Through Effective Reduction of Adjacent Channel Interference

in Multirate Processing

Ganekanti Hemanja, K. Satya Prasad, P. Venkata Subbaiah












Signal Processing: An International Journal (SPIJ), Volume (6), Issue (1) : 2012

---

Mr.G. Hemanja,Dr. K. Satya Prasad & Dr. P. Venkata Subbaiah
Efficient Spectral Analysis Through Effective Reduction of
Adjacent Channel Interference in Multirate Processing


Mr. G. Hemanja
ganekantihemanj@yahoo.com
Research Scholar/ECE Department
J.N.T.University
Kakinada,533001,India

Dr. K. Satya Prasad


Prasad_kodati@yahoo.co.in
Professor in ECE Dept. and Rector
J.N.T.University
Kakinada,533001,India

Dr.P. Venkata Subbaiah


pvs_ece2000@yahoo.co.in
Professor in ECE Dept. and Principal
ASIST, J.N.T.University
Kakinada, 533001, India

Abstract

Spectral analysis is considered to be important area of consideration in which the volume
reconstruction and visualization takes place, besides amounting to increase in computational
efficiency. The reception of selected frequency and amplitude level takes an important role for
further processing at the succeeding stages of Digital Signal Processing. The number of
developments are undertaken through research work for effective reception of signals. In this
paper, the proposed method is based on Multirate technique, specified by authors [1][2], in Finite
Impulse Response digital filter bank through Modified Kaiser window, subsequently followed by
Fast Fourier transform. A remarkable spectral output is achieved by way of increase in
magnitude, linear phase response, constant width and sharp rise of response, less adjacent
channel interference and better stopband attenuation as compared to authors of [3][4][5][6]
besides significant improvement of specific parameters listed among the respective methods are
elicited. In addition, a better reduction in computational complexity is achieved. This method of
spectral analysis is suitable in most of applications, especially in digital hearing aid applications
as it is compatible with low frequency, low delay and low power requirements. The simulation
results are added on account of satisfactory performance and comparison is drawn to enlighten
the advantages in the proposed method.

Keywords: Mutirate Processing, Bandpass Filtering, Decimation, Modified Kaiser Window, Net
Input Samples, Spectral output.


1. INTRODUCTION
Many applications need effective detection of signals mainly in terms of amplitude and frequency
components, which become easier in processing of signals and implementation of hardware at
the succeeding stages. In addition, the received signals become effective if they are received at
maximum peak response, free from adjacent frequency interference and better stopband
attenuation. More over, the reception becomes significant if the computational complexity is less.
In order to meet the specified requirements, this paper proposes a method in Multirate Digital
filter bank followed by Fast Fourier transform (FFT). Thus, this paper is emerged with a novel
technique in spectral analysis to best suit among various applications. There is a possibility for
the enhancement in selectivity and less adjacent channel interference as mentioned in the
results, depicted by authors [3][4][7]. In addition to these characteristics, further there is a
possibility of improvement in characteristics like peak response and stopband attenuation as per
Signal Processing: An International Journal (SPIJ), Volume (6) : Issue (1) : 2012
1

---

Mr.G. Hemanja,Dr. K. Satya Prasad & Dr. P. Venkata Subbaiah
the results shown in [5][6][8]. For these reasons, the proposed method of spectral analysis is
suggested to overcome such disadvantages among these methods, depicted for various
applications. The Finite Impulse Response (FIR) band pass filter is used as it has a great number
of advantages including precise control over the phase response which greatly facilitates signal
reconstruction. The filter is designed based on `Modified Kaiser window' approach which
improves sharp rise and constant width response characteristic. Among others, particularly in
digital hearing aid applications, the filter bank provides a natural decomposition of the input signal
into frequency bands which may be processed independently to best compensate for the hearing
loss and meet prescriptive targets, as hearing loss is a function of both frequency and input level.
A reduction in computational complexity and hardware complexity is achieved through
heterodyning operation and arrival of response using the techniques in Fast Fourier transform.
The heterodyning operation shifts the original band of input frequencies to the lower frequencies
starting from `0' Hz. In addition, the techniques in FFT algorithm presents further minimizing of
computations rather than specified one in the reference [8]. These techniques in the proposed
method have led to specific advantages in respect of linear phase response, constant width and
sharp rise of response, minimum adjacent channel interference and better stopband attenuation
as compared to the existing methods apart from great reduction in computational complexity. A
novel technique is appended to the actual part of work resulting to further enhancement in
stopband attenuation and peak output response. This paper is a description of an extremely
flexible frame work for separating the input into frequency bands as it forms the basis for a
Multichannel compression in hearing aid development as demonstrated by authors [9][10], in
addition to analysis of spectral components. This method is well described in this paper with the
help of analytic equations. The simulation results and performance characteristics are drawn to
verify with the facts. This method is particularly verified at lower frequencies because of quality
assurance, implemented even at higher frequencies and so chosen with the input signal as
`Cosine function' as this function represents the real time signals in most of applications.

2. METHODOLOGY AND RELATED EQUATIONS

2.1
Design of Bandpass Filter
This method of approach is arrived through partial modification of designated parameters in the
existing Kaiser window method [8].
The frequency response of bandpass filter is given as

.
1 ........ ..... for L f
f f
(
h z )
p1
p2
=
----------------------------(2.1.1)

.
0 ........ .... forL f f Land L f f
p1
p2
Where fp1 is lower pass band frequency and fp2 is upper passband frequency .
The desired impulse response is given by hd(n).
1 2 nf


2 nf



sin
c 2 - sin
c1
L
L
for Ln0



n

F

F

h (n) =
---------------(2.1.2)
d

2
( f - f ) L
L
forLn = 0

c 2
c1
F

Where fc1 = fp1-(df/2), fc2 = fp2+(df/2)
df2 = fp1-fs1, df3 = fs2-fp2
where df = min{df2,df3},
F = Sampling frequency
fs1 = Lower stop band frequency
fs2 = Upper stop band frequency
n = integer
N' = Order of the filter
The frequency response of the hd(n) is given by h(z).
Signal Processing: An International Journal (SPIJ), Volume (6) : Issue (1) : 2012
2

---

Mr.G. Hemanja,Dr. K. Satya Prasad & Dr. P. Venkata Subbaiah
(N '-1)
(N '- )
1 / 2
-


h(z) =
2
z
h (0) +
h n
f
nT
--------------------------(2.1.3)
d
2 ( )

cos(2
)
d


n=1


( N '- )
1 / 2
The equation, h (0) +
h n
f
nT represents the magnitude response equation,
d
2 ( )cos(2
)
d
n=1
where T = 1/F.

2.2
Design of Window
I (B)
(N -
'
)
1
o
- - - - - forLn
The basic Kaiser Window function, a (n) =
2
-------- (2.2.1)
k
I (A)
o

0 - - - - - - - -otherwise
Where `A' is an independent variable empirically determined by Kaiser
0.5

2


2n


B = A 1

-




(N '-1



2
2
2

(0.5x )
( .
0 5x )

I (x) = 1 +
+
+ ......... .
. ...
o
!
1
( )2
( !
2 ) 2


This series converges rapidly and can be computed up to 25 terms to attain the desired accuracy.
Considering the bandpass specifications such as the passband ripple (Ap) and minimum
stopband attenuation (As) in decibels are given by
1
( + dp)
Ap = 20 log

10
1
( - dp)
As = 2
- 0 log (ds)
10

The `ds' and `dp' can be determined as

( .
0
As)
ds
05
10 -
=

1
( 0 Ap - )
1
dp =

1
( 00 0.5Ap + )
1
da = min( dp, ds)

The actual stopband attenuation is arrived to be As = 2
- 0 log (da)
10
The value of `A' is determined from the empirical design equation

0 - - - - - - for L
L
As 21

A = 0.5942(As - 2 )
1 0.4 + 0 7
. 986( As - 20) - - - - forL21 As 50

1102(As - 8.7) - - - - - - - for L
L
As50

A parameter `D' is determined from the empirical design equation,
Signal Processing: An International Journal (SPIJ), Volume (6) : Issue (1) : 2012
3

---

Mr.G. Hemanja,Dr. K. Satya Prasad & Dr. P. Venkata Subbaiah
.
0 9222 - - - - - for
L
L
As 21

D = (As - .
7 95)


- - - for
L
L
As21
14 3
. 6
The filter order for the lowest odd value of `N' ', calculated to be
(F )(D)
N '
+ 1
df
The order of the filter increases with increase in the sampling frequency or decrease in the
transition width. Finally the impulse response of the filter is determined as h(n) = ak(n) hd(n).
Where ak(n) and hd(n) are window coefficients and desired impulse response respectively. The
frequency response and the output response, for an input signal band of frequencies from 480 Hz
to 530 Hz, of the filter bank are plotted according to Modified Kaiser window approach, as shown
in Figure-1 and Figure-2 respectively. It is observed that a better spectral response with
improvement in passband ripple and quality of output are attained as compared to the methods of
reference [5][6][8]. Accordingly, this novel approach of filter bank design can bring attractive
advantages in the proposed method of filter bank design, which follows.
1.2
First stage response
1
Second stage response
Third stage response
Fourth stage response
0.8
0.6
E
D
U
I
T
L
P
M
A 0.4
0.2
0
-0.20
200
400
600
800
1000
1200
1400
FREQUENCY
FIGURE 1: Modified Kaiser Window Based Frequency Response of Bandpass Filter Bank.






Signal Processing: An International Journal (SPIJ), Volume (6) : Issue (1) : 2012
4

---