FACULTY OF SCIENCE AND TECHNOLOGY

APPLIED PHYSICS PROGRAMME

MATHEMATICAL PHYSICS
(SFS 2073)
(CONVOLUTION THEOREM)

NAME: NUR SHAHIRAH BINTI GHAZALI

MATRIC NUMBER: 1140692
LECTURERS NAME: PROF DR. KARSONO BIN AHMAD DASUKI

ACKNOWLEDGMENT

Alhamdulillah, in the name of Allah The Most Gracious and The Most Merciful.
First of all, Im very thankful to God because give me a wonderful life as a student in
USIM. I got to learn many things here from my teacher and my friend and those are
among the biggest chances. He gave me to be a good person as well as a student.
Secondly, I would like to express my appreciation and respect to my Mathematical
Physics lecturer, Prof Dr. Karsono bin Ahmad Dasuki for guiding and help me in order
to complete this assignment.
I take this opportunity to show my gratitude to all of the members of Applied
Physics Programme for helping me to find the sources for this assignment and help me in
completing this task. Next, I would like to thanks my beloved parent, Ghazali bin Mamat
and Zainab binti Abd Wahab for the spiritual support and encouragement to me.
Last but not least, I also place my gratitude to any person who directly or indirectly
helping me to finish this assignment.

ABSTRACT
The main purpose of assignment is to extract the knowledge of physics and
mathematics in this world based on Islamic and scientific perspective. In this assignment
I want to enlighten the important theorem in Mathematics and Physics which called
Convolution Theorem.
Mathematically, a convolution is defined as the integral over all space of one
function at x times another function at u-x. The convolution is an important construct
because of the Convolution Theorem which gives the inverse Laplace transform of a
product of two transformed functions
Convolution have its own special and specific properties which are commutative,
associative, distributive over addition, identity element or 1, bilinear, time shifting and
time scaling.
We use the convolution in daily life. For example, we use it in electrical
engineering for circuit analysis when it is difficult or impossible to find a Laplace
transform for our circuit equations. Convolution is used all the time in image processing
(i.e. blur filters / Gaussian smoothing). It is also used to find the response of any linear
system to a given input. Other than that, it is used in digital signal processing, filtering,
controls.
Last but not least, in this assignment, I also discuss about the mathematical
theorem in term of Islamic perspective. I choose a few Islamic scholars that related to
mathematical Physics subject.

TABLE OF CONTENT
1. HISTORY OF CONVOLUTION THEORY ..7
1.1. HISTORICAL DEVELOPMENTS ....10
2. WHAT IS MEANT BY CONVOLUTION? ..............................................................11
2.1. CONVOLUTIONS ARE COMMUTATIVE ..........16
3. THE CONVOLUTION THEOREM .....17
4. PROOF OF THE CONVOLUTION THEOREM ......17
4.1. PROOF OF SECOND STATEMENT OF CONVOLUTION THEOREM ........17
4.2. PROOF OF SECOND STATEMENT OF CONVOLUTION THEOREM ........18
5. CONVOLUTION THEOREM FOR INVERSE FOURIER TRANSFORM .........19
6. EXAMPLE OF CONVOLUTION .....21
7. CONVOLUTION PROPERTIES....22
7.1. ALGEBRAIC PROPERTIES FOR CONVOLUTION ...........23
7.1.1. COMMUTATIVITY ....23
7.1.2. ASSOCIATIVITY ....23
7.1.3. DISTRIBUTIVITY ..23
7.1.4. ASSOCIATIVITY WITH SCALAR MULTIPLICATION .....................23
7.1.5. DISTRIBUTIVE OVER ADDITION ......23
7.1.6. IDENTITY ELEMENT OR 1 .......24
7.1.7. BILINEAR ....24
7.1.8. TIME SHIFTING .....24
7.1.9. TIME SCALING ......24
7.2. THE CORRELATION AND CONVOLUTION ALGEBRAS ......24
7.2.1. MULTIPLICATIVE IDENTITY .....26
7.2.2. INVERSE COMPONENT ...27
7.2.3. COMPLEX CONJUGATION ......27
7.2.4. INTEGRATION ...27
7.2.5. DIFFERENTATION ....27
7.3. CONVOLUTION INTEGRAL ...29
8. THE CORRELATION THEOREM .......30
8.1. WHAT IS A CORRELATION FUNCTION? ........30
3

9. CONVOLUTION IN PARSEVAL'S THEOREM .....31

9.1. CONVOLUTION WITH A GAUSSIAN ...31
10. CONVOLUTION WITH A DELTA FUNCTION .......32
11. APPLICATIONS OF THE CONVOLUTION THEOREM ......................................34
11.1.

IN IMAGE PROCESSING ......34

11.2.

IN COMPUTERIZED INFORMATION PREPARING ......34

11.3.

IN ACOUSTICS ...34

11.4.

ELECTRICAL DESIGNING .......35

11.5.

MATERIAL SCIENCE ....35

11.6.

IN LIKELIHOOD THEOREM ........36

11.7.

RADIOTHERAPHY TREATMENT .......36

11.8.

CONVOLUTIONAL NEURAL SYSTEM ......36

11.9.

ATOMIC SCATTERING FACTORS ......36

11.10.

B-FACTORS ....36

11.11.

DIFFRACTION FROM A LATTICE ..........38

11.12.

DIFFRACTION FROM A CRYSTAL ........39

11.12.1.

DETERMINATION TRUNCATION ......39

11.12.2.

MISSING INFORMATION .....40

11.12.3.

DENSITY MODIFICATION .......41

11.12.3.1. SOLVENT FLATTENING ..........41

11.12.3.2. SAYRE'S EQUATION ....41

12. APPLICATIONS OF THE CORRELATION THEOREM ...........41

12.1.

THE PATTERSON FUNCTION .........41

13. TYPE OF CONVOLUTION...42

13.1.

CIRCULAR CONVOLUTION ....42

13.2.

DISCRETE CONVOLUTION .....43

13.3.

CIRCULAR DISCRETE CONVOLUTION ....43

13.4.

FAST CONVOLUTION ALGORITHMS .......44

14. CONVOLUTION OF TWO COMPLEX-VALUED FUNCTIONS ..........45

15. CONVOLUTION FUNCTION ......48
16. CONVOLUTIONS ON GROUPS .....49
17. CONVOLUTION OF MEASURES ...51
4

18. BIALGEBRAS ...52

19. TITCHMARSH CONVOLUTION THEOREM ........52
20. CONVOLUTION AS A METHOD TO CALCULATE ZERO STATE RESPONSE
.........................54
21. THE CONVOLUTION IS A SUM OF IMPULSE RESPONSES ............ 55
22. VISUALIZING THE CONVOLUTION INTEGRAL .......61
23. MATHEMATICS IN THE ISLAMIC PERSPECTIVE
23.1.

AL-KHAWARIZMI..................................................................................64

23.2.

AL-BIRUNI...............................................................................................66

24. CONCLUSION..........................................................................................................67
25. REFERENCES............................................................................................................68

LIST OF FIGURES
Figure 1: Graphs of the superposition of several weighted of g and the integral...........13
Figure 2: Correlation function...........................................................................................31
Figure 3: Set of closely-spaced horizontal lines to a widely-spaced vertical row of points
Figure 4: Impulse Respond ...............................................................................................54
Figure 5: Impulse per unit time .........................................................................................55
Figure 6: Discrete approximation of impulse responses ...................................................56
Figure 7: The reaction of the system to the individual impulse ........................................57
Figure 8: The impulse responses when T=0.3 seconds ..................................................59
Figure 9: The impulse responses when T=0.3 seconds and T=0.8 ..............................59
Figure 10: Exact impulse response and respond using convolution .................................60
Figure 11: Impulse response .............................................................................................61
Figure 12: The function h(t-) and its product .................................................................62
Figure 13: Result of the integration for all time, with a black dot at t=1..........................63

1. HISTORY OF CONVOLUTION THEORY

It was the 1760s when the Swiss numerical virtuoso Leonhard Euler (17071783)
endured complete visual deficiency. However this ailment did not keep him from adding
to science. Truth has been told, in those years, he composed a huge book about necessary
math and the arrangement of differential comparisons (DEs) by certain unmistakable
integrals. In Chapter X of this book, and particularly in issue 131, he concentrated on the
arrangements of second-request halfway DEs of the structure
= [() + ()]()

(1.1)

A first and basic instance of this arrangement is K(u) = u and Q(x) = x and a second case
is K(u) = - u and Q(x) = x. Euler concentrated on the primary case in, which is the main
source in which the relationship operation is utilized thoroughly as a part of measurable
applications and identified with convolution. Concerning the second case, (1.1) results in
a specific type of the numerical operation demonstrate the purpose behind this paper.
= ( ) ()

(1.2)

In any case, it appears that Euler did not concentrate on (1.2) specifically, but rather he
studied the situation when Q(x) = x and K(u) is some other function.
A more broad type of (1.2) shows up "normally" in the arrangement of DEs by
fundamental transform (like Fourier or Laplace transform) identified with computational
liquid progression, electrical building, heat conduction, flag and picture preparing, direct
acoustics, mechanical designing, optics, likelihood and measurements, radiotherapy,
spectroscopy, and viscoelasticity. In particular, for the given two complex-valued
appropriate function f and g of a genuine or complex variable contention, the
accompanying expression is more broad than (1.2):
()( )

(1.3)

Here, l might be either an interim of joining (a, b) if is a genuine variable or a bend if

is a mind boggling variable. For the main occurrence, the three most normal cases are the
accompanying, where

 = < < <

From a graphical perspective and being the primary variable, this operation comprises
of the scientific procedure that can be deciphered graphically by collapsing the
contention of working, interpreting the contention of function g, increasing the
subsequent function g by function f, and coordinating that duplication. Also, this
understanding and its graphical representation conceivably showed up without precedent
for the same book in its first version in 1942.
In this way, the operation is in some sense the "spinning" of one of the data function with
the other one. Especially in arithmetic, material science, and related territories, the verb
generally used to assign such a rotating is "to convolve." This verb originates from the
Latin words con and volve, which signify "together" and "move up," individually; along
these lines, convolve, signifies "move up together." Accordingly, the activity of
convolving is called convolution. This is the reason it is typically referred to today as the
CCO, or just convolution, in spite of the fact that it has been given numerous different
names previously, as talked about in the area "Names of the CCO."
The undertaking of following back the historical backdrop of the CCO is not a simple
one since it requires investigating and confirming data contained in various books and
papers. At times, the data gave is not correct, while in others, it involves just some
remarks with no references. Regardless, following back these events, genuine or
gathered, is the principle objective of this area. Luckily, there is no less than a beginning
stage: the book by Euler.
A case of a gathered event is appended to the French mathematician Pierre-Simon
Laplace (17491827). Truth be told, Miller, in the section "Convolution," puts forth the
accompanying expression:
Expressions that would now be depicted as "convolutions" show up in Laplace's most
punctual work on entireties of autonomous irregular variables, "Mmoire sur l'inclinaison
moyenne des orbites des comtes, sur la figure de la terre et sur les fonctions," Mmoires
de l'Acadmie Royale des Sciences de Paris, Savants trangers, vol. 7, pp. 503540,
1773.
Nevertheless, subsequent to surveying this diary, no hint of the CCO is found. Mill
operator noticed that:
8

A progression of French and Russian mathematicians took after Laplace and without
utilized convolutions; it appears, advancing a name for them. See A. Hald History of
Mathematical Statistics from 1750 to 1930 (1998).
Notwithstanding, Hald's book expresses that Laplace in reality utilized the CCO for
taking care of the issue of finding the dispersion of the mean, and this happened on p.
478 of a journal written in 1778 and distributed in Mmoires de l'Acadmie Royale des
Sciences de Paris in 1781. In this way, the reference given by Miller about Laplace's
work is by all accounts wrong.
At the beginning of the 19th century, a young mathematician interested in the application
of mathematics to physical phenomena started to write a memoir called Thorie de
Mouvement de la Chaleur Dans les Corps Solides, which was later part of his celebrated
book first published in 18211822, Thorie Analytique de la Chaleur. This
mathematician is, of course, Jean Baptiste Joseph Fourier (17681830) [9]. In this book,
the CCO appeared in section 265 [9, pp. 233234]:
This journal by Laplace has been the motivation of numerous connected mathematicians
and specialists for creating numerous different theorem and results. This is the situation
of the work about contrasts and arrangement distributed in a book in 1800 by the French
mathematician Sylvestre Franois Lacroix (17651843). In that, the CCO showed up in
an application with respect to the summation of arrangement by unmistakable integrals. It
merits saying this is maybe the first run through the CCO showed up in its more broad
structure, as in (3). Incidentally, the German mathematician Heinrich Friedrich Karl
Ludwig Burkhardt (18611914) referred to the work by Lacroix in a footnote that
showed up in [8, p. 400].

() = ()

{ +
= ( )}

(1.4)

(1.5)

Fourier's book is the first in which the CCO showed up various times (see, e.g., Chapter
IX).

Two peers of Fourier were the French mathematicians Augustin Louis Cauchy (1789
1857) and Simon Denis Poisson (17811840). Cauchy made utilization of various CCO
expressions in a long journal on the hypothesis of engendering of waves introduced to
l'Acadmie des Sciences de l'Institut de France and submitted to the 1815 rivalry for the
recompense of scientific examination. Concerning, the CCO initially showed up in his
diary on the hypothesis of waves distributed in 1816.
Poisson utilized the CCO in his consequent papers. For instance, in a 1826 journal about
the attraction in development, the CCO showed up various times. Concerning this diary
and the utilization of the CCO, Gardner and Barnes compose [2, p. 364]: "The treatment
of this paper prompts the supposition that these thoughts were at that point known."
Indeed, in the 1820s, the possibility of CCOs was entirely basic in the connected science
environment. For instance, a declaration of the sort given by (2) showed up in the
arrangement of certain necessary mathematical statements got from clear issues in
material science. This is the situation of the necessary comparison that showed up in the
issue of tautochrone bends, which was settled by the Norwegian mathematician Niels
Henrik Abel (18021829) and distributed in two papers in 1823 and 1826.
Toward the end of and after the 1820s, the utilization of the CCO turned out to be more
normal in papers identified with immaculate and connected arithmetic (Alejandro
Dominguez.2015).

1.1. Historical developments

One of the earliest uses of the convolution integral appeared in DAlemberts derivation
of Taylors theorem in Recherches sur diffrents points importants du systme du monde,
published in 1754.
Also, an expression of the type:

(). ( )

(1.6)

is used by Sylvestre Francois Lacroix on page 505 of his book entitled Treatise on
differences and series, which is the last of 3 volumes of the encyclopedic series: Trait
du calcul diffrentiel et du calcul intgral, Chez Courcier, Paris, 1797-1800.Soon
10

thereafter, convolution operations appear in the works of Pierre Simon Laplace, Jean
Baptiste Joseph Fourier, Simeon Denus Poisson, and others. The term itself did not come
into wide use until the 1950s or 60s. Prior to that, it was sometimes known as faltung
(which means folding in German), composition product, superposition integral, and
Carson's integral. Yet it appears as early as 1903, though the definition is rather
unfamiliar in older uses.
The operation:

()( )

<

(1.7)

is a particular case of composition products considered by the Italian mathematician Vito

Volterra in 1913.

2. WHAT IS MEANT BY CONVOLUTION?

A standout amongst the most critical concept in Fourier theory, and in crystallography, is
that of a convolution. Convolutions emerge in numerous appearances, as will be
demonstrated as follows. In view of a scientific property of the Fourier transform, alluded
to as the convolution theorem, it is advantageous to do counts including convolutions.
Be that as it may, first we ought to characterize what a convolution is. Understanding the
idea of a convolution operation is more critical than comprehension a proof of the
convolution theorem.
Scientifically, a convolution is characterized as the fundamental over all space of one
capacity at x times another capacity at u-x. The joining is assumed control over the
variable x (which might be a 1D or 3D variable), normally from less vastness to
limitlessness over all the measurements. So the convolution is a component of another
variable u, as appeared in the accompanying mathematical statements. The cross around
is utilized to demonstrate the convolution operation. The convolution is now and again
likewise known by its German name, faltung ("collapsing").

11

() = () () = ()( )

(2.1)

() = () () = ()( )

(2.2)

Note that it doesn't make a difference which work you take to start with, i.e. the
convolution operation is commutative. We'll demonstrate that beneath, however you
ought to consider this as far as the representation underneath. This outline demonstrates
how you can consider the convolution, as giving a weighted aggregate of moved
duplicates of one function: the weights are given by the function estimation of the second
function at the movement vector. The top pair of diagrams demonstrates the first
function. The following three sets of diagrams appear (on the left) the function g moved
by different estimations of x and, on the right, that moved function g increased by f at the
estimation of x.
Conceptually, a convolution is characterized as a result of functions and that is items in
the polynomial math of Schwartz functions in Rn. Convolution of two functions and over
a limited reach is given by

[ ]() ()( )

(2.3)

where the image signifies convolution of and g.

Convolution is all the more regularly assumed control over an endless reach,

()( )

= ()( )

12

(2.4)
(2.5)

Figure1: Graphs of the superposition of several weighted of g and the integral.

The base pair of diagrams appears, on the left, the superposition of a few weighted and
moved duplicates of g and, on the privilege, the necessary (i.e. the whole of all the
weighted, moved duplicates of g). You can see that the greatest commitment originates
from the duplicate moved by 3, i.e. the position of the top of f.
On the off chance that one of the capacities is unimodal (has one crest), as in this outline,
the other function will be moved by a vector equal to the position of the crest, and spread
out by a sum that relies on upon how sharp the crest is. However, on the other hand we
could switch the parts of the two capacities, and we would see that the bimodal function
g has multiplied the tops of the unimodal function f.
13

In arithmetic, the convolution theorem expresses that under reasonable conditions the
Fourier transform of a convolution is the point wise result of Fourier transform. As such,
convolution in one space (e.g., time area) measures up to point-wise augmentation in the
other space (e.g., recurrence area). Variants of the convolution theorem are valid for
different Fourier-related transform. Give f and g a chance to be two functions with
convolution f * g. (Note that the reference mark signifies convolution in this setting, and
not increase. The tensor item image is once in a while utilized rather.) Let F mean the
Fourier change administrator, so F{f} and F{g} are the Fourier transform of f and g,
separately. At that point
{ } = {} {}

(2.6)

where denotes point-wise multiplication. It also works the other method:

{ } = {} {}

(2.7)

{ } = {{} {}}

(2.8)

and

Note that the connections above are substantial for the type of the Fourier transform
appeared in the Proof area underneath. The change might be standardized in different
courses, in which case steady scaling elements (regularly or) will show up in the
connections above.

This theorem additionally holds for the Laplace transform, the two-sided Laplace change
and, when reasonably adjusted, for the Mellin transform and Hartley transform (see
Mellin reversal hypothesis). It can be stretched out to the Fourier transform of theoretical
symphonies investigation characterized over locally conservative abelian bunches.

This plan is particularly helpful for actualizing a numerical convolution on a PC: The
standard convolution calculation has quadratic computational many-sided quality. With
the assistance of the convolution hypothesis and the quick Fourier change, the many14

sided quality of the convolution can be decreased to O(n log n). This can be misused to
develop quick duplication calculations.

Now we present the convolution of two functions f(t), g(t) which we mean by (f g)(t).
The convolution is a vital build on account of the Convolution Theorem which gives the
opposite Laplace transform of a result of two changed function:

{()()} = ( )()

(2.9)

Give f(t) and g(t) a chance to be two elements of t. The convolution of f(t) and g(t) is
likewise a component of t, meant by (f g)(t) and is characterized by the relation

( )() = ( )()

(2.10)

Notwithstanding if f and g are both causal functions then (entirely) f(t), g(t) are
composed f(t)u(t) and g(t)u(t) individually, so that

( )() = ( )( )()() = ( )()

(2.11)

because of the properties of the step functions (u(t x) = 0 if x > t and u(x) = 0 if x < 0).
If f(t) and g(t) are causal functions then their convolution is characterized by:

( )() = ( )()

(2.12)

This is an odd looking definition yet it ends up having impressive use both in Laplace
transform theory and in the demonstrating of straight designing frameworks. The reader
ought to note that the variable of incorporation is x, to the extent the mix procedure is
concerned the t-variable is (briefly) viewed as a constant.

15

2.1. CONVOLUTIONS ARE COMMUTATIVE

() () = () ()

(2.13)

We expressed that the convolution vital is commutative. Here, in the event that you are
intrigued, is a speedy confirmation of that. In the first place, we begin with the
convolution necessary kept in touch with restricted. For accommodation we will manage
the 1D case, yet the 3D case is precisely comparable to.

() = ()( )

(2.14)

Now we substitute variables, replacing u-x with a new x'.

= , =

() = ( )( )

(2.15)
(2.16)

Note that, in light of the fact that the indication of the variable of mix transformed, we
need to change the indications of the points of confinement of incorporation. Since these
breaking points are interminable, the movement of the birthplace (by the vector u) doesn't
change the greatness of the cut off points.

Presently we switch the request of the breaking points, which changes the indication of
the mathematical statement, and swap the request of the functions g and f.

() = ( )( )

() = ()( )

16

(2.17)
(2.18)

3. THE CONVOLUTION THEOREM

Since there will be such a variety of Fourier transform in whatever remains of this
presentation, it is helpful to present a shorthand documentation. T will be utilized to
show a forward Fourier transform, and its reverse to demonstrate the opposite Fourier
transform.
(()) = (. )

(3.1)

There are two methods for communicating the convolution theorem:

The Fourier transform of a convolution is the product of the Fourier transforms.

The Fourier transform of a product is the convolution of the Fourier transforms.

() = ()()

(3.2)

() = () ()

(3.3)

The convolution hypothesis is helpful, to some degree, since it gives us an approach to

rearrange numerous counts. Convolutions can be exceptionally hard to compute
specifically, yet are regularly much less demanding to figure utilizing Fourier transform
and

augmentation

(Convolution

Theorem

in

http://mathworld.wolfram.com/ConvolutionTheorem.html).
4. PROOF OF THE CONVOLUTION THEOREM
4.1. Proof of first statement of convolution theorem

(()()) = ( ()( ))

= ()( ) ()

(4.1)
(4.2)

Presently we substitute another variable w for u-x. As over, the unending reconciliation
limits don't change. At that point we extend the exponential of a total into the result of
exponentials and revise to unite expressions in x and expressions in w.

17

(()()) = ()() [( + )]

(4.3)

= () ()

(4.4)

Expressions in x can be taken out of the indispensable over w so we have two separate
integrals, one over x with no terms containing w and one over w with no terms
containing x.

(()()) = () () () ()

(4.5)

(()()) = () () () ()

(4.6)

= (())(())

(4.7)

On the off chance that you look through the determination above, you will see that we
could have utilized a short sign as a part of the exponential when taking the first Fourier
transform, and afterward the two Fourier transform toward the end would likewise
contain less signs in the exponentials. At the end of the day, the convolution theorem
applies to both the forward and invert Fourier transforms. This is not astounding,
subsequent to the two headings of Fourier transform are basically indistinguishable.

4.2. Proof of second statement of convolution theorem

To demonstrate the second proclamation of the convolution hypothesis, we begin with

the form we have officially demonstrated, i.e. that the Fourier transform of a convolution
is the result of the individual Fourier transform.
() = ()()

(4.8)

To start with we'll characterize some shorthand, where capital letters show the Fourier
transform mates of lower case letters.
= () = ()

(4.9)

= () = ()

(4.10)

18

We utilize these connections to recast the announcement above as far as the Fourier
transform mates of the first functions. At that point we take a reverse Fourier Transform
on every side of the mathematical statement, to get (basically) the second explanation of
the convolution theorem. The main distinction is that it is communicated regarding the
backwards Fourier Transform.
[ () ()] =

(4.11)

() () = ()

(4.12)

But, as we mentioned above, we could have proved the convolution theorem for the
inverse transform in the same way, so we can express this result in terms of the forward
transform (Randy J Read.2009).
() = ()()

(4.13)

5. CONVOLUTION THEOREM FOR INVERSE FOURIER TRANSFORM

With similar argument as the above proof, we have the convolution theorem for the
inverse

Fourier

transforms

(Convolution

Theorem

in

http://mathworld.wolfram.com/ConvolutionTheorem.html)

{{ } = {} {}

(5.1)

{ } = { {} {}}

(5.2)

= { {} {}}

(5.3)

And

19

Functions of discrete variable sequences

By comparative contentions, it can be demonstrated that the discrete convolution of
arrangements and is given by:
= [{}. {}

(5.4)

where DTFT represents the discrete-time Fourier transform.

An essential exceptional case is the circular convolution of x and y characterized by x N*y
where xN is a periodic summation:
[]
= [ ]

(5.5)

It can then be shown that:

= [{ }. {}
= [{ }. { }

(5.6)
(5.7)

where DFT represents the discrete Fourier transform.

The proof follows from DTFT periodic data, which indicates that { }can be
written as:

{ }() =
=({ } []). ( /)

(5.8)

The product with {}().is thereby reduced to a discrete-frequency function:

{ }[]. {} () . ( )

(5.9)

The inverse DTFT is:

( )[] =

{ }[]. ( ) .
=

20

(5.10)

{ }[]. { }[]. ( ) .
=

(5.11)

(/)
=
= { }[]. { }[].

(5.12)

= [{ }. { }]

(5.13)

6. EXAMPLE OF CONVOLUTION

Find the convolution of f and g if f(t) = tu(t) and g(t) = t^2u(t).

Solution
( ) = ( )( ) () = ()
Therefore

( )() = ( ) = [ ]
=
2) Find the convolution of

() = () and () = ()

Solution
Here ( ) = ( )( ) and () = () and so

( )() = ( )

We will need to integrate by parts. We discover, recalling again that t is a constant in the
integration process,

(
[(
) =
) ] ()()

21

= [ + ]

= []

So that,
(f g)(t) = t sin t or, equally, for this situation (t sin t) (t) = t sin t.

7. CONVOLUTION PROPERTIES

() = ()() () = () ()

(7.1)

() = () () () = () ()

(7.2)

Convolution in the time domain is equivalent to multiplication in the frequency domain

and

vice

versa

("Books

about

http://www.ericweisstein.com/encyclopedias/books/Convolution.html )
Proof of second line:
Given u(t), v(t) and w(t) satisfying w(t) = u(t)v(t) W(f) = U(f) V (f)
Define dual waveforms x(t), y(t) and z(t) as follows:
x(t) = U(t) X(f) = u(f)
y(t) = V (t) Y (f) = v(f) z(t) = W(t) Z(f) = w(f)
Now the convolution property becomes:
w(f) = u(f)v(f) W(t) = U(t) V (t)
Z(f) = X(f)Y (f) z(t) = x(t) y(t)
Convolution: w(t) = u(t) v(t) , R u( )v(t )d

22

Convolution."

7.1. ALGEBRAIC PROPERTIES FOR CONVOLUTION

The convolution characterizes an item on the direct space of integrable functions. This
item fulfils the accompanying arithmetical properties, which formally imply that the
space of integrable function with the item given by convolution is a commutative
variable based math without personality (Strichartz 1994, 3.3). Other direct spaces of
functions, for example, the space of persistent elements of reduced backing, are shut
under the convolution, thus likewise frame commutative algebras.

7.1.1. Commutativity
=

(7.3)

() () = () ()

(7.4)

7.1.2. Associativity
( ) = ( )
() () () = (() ()) () = () (() ())

(7.5)
(7.6)

7.1.3. Distributivity
( + ) = ( ) + ( )

(7.7)

7.1.4. Associativity with scalar multiplication

( ) = ()

(7.8)

for any genuine (or complex) number .

7.1.5. Distributive over addition
() (() + ()) = () () + () ()

23

(7.9)

7.1.6. Identity Element or 1:

() () = () () = ()

(7.10)

7.1.7. Bilinear:
(()) (()) = (() ())

(7.11)

7.1.8. Time Shifting:

( + ) ( + ) = ( + + )

(7.12)

7.1.9. Time Scaling:

() () = ||()

(7.13)

7.2. THE CORRELATION AND CONVOLUTION ALGEBRAS

Convolution and correlation operations are defined in the usual way on the linear space
of functions with a "well-behaved" Fourier transform (for example the Hilbert space L2 (
- 00, 00) or suitable restrictions of it).The convolution function of I(t) and g(t) is then

( ) = ()( )

(7.14)

and the correlation is

() = ()( )

(7.15)

If I and g are a periodic functions, definitions (7.14) and (7.15) take the forms

(With|()|, |()| finite)

( ) = ()( )

() = ()( )
24

(7.16)
(7.17)

The extension to functions of many variables is straightforward.

The Fourier isophormic relationships are (disregarding normalization factors)
( ) = ()()

(7.18)

And
() = ()()

(7.19)

where F and G are the Fourier transforms of I and g, and F is the complex conjugate of
F. Eq. (7.19) is usually referred to as the Wiener theorem. It can be seen from definitions
that under the apparent similarity an important difference exists: convolution is
commutative and associative, correlation is neither commutative nor associative. As a
matter of fact, in the Fourier isomorphic space, where multiplication is by definition
commutative and associative, the following relations hold




( )
( )
()
()
It is straightforward to check that, with the usual operations (addition and multiplication
with scalars) correlation and convolution induce, on the space L2, two algebras, denoted

25

respectively by U* and U.We may regard the algebras U* and Uas having the same
quantities but different laws for forming products.
In the following, we restrict our attention to discrete spaces. In particular, we consider the
linear algebras B* and B over the real field, which consist of all vectors f
f = (..., f-1, f0, f1....)

(7.20)

together with the following product laws

( ) =

(7.21)

() = +

(7.23)

As shown in Appendix II the algebras B* and B have a direct physical interpretation. In

fact the infinite vectors f can be assumed to represent in terms of the Whittaker basis the
class of band-limited real-valued signals.
7.2.1. Multiplicative identity
No variable based math of functions has a character for the convolution. The absences of
identity are ordinarily not a noteworthy impediment, since most accumulations of
function on which the convolution is performed can be convolved with a delta circulation
or, at any rate (similar to the instance of L1) concede approximations to the character.
The direct space of minimally upheld circulations does, be that as it may, concede a
character under the convolution. In particular,
=

(7.24)

where is the delta distribution.

26

7.2.2. Inverse component

A few circulations have an opposite component for the convolution, S(1), which is
characterized by
() =

(7.25)

The arrangement of invertible circulations shapes an abelian bunch under the

convolution.
7.2.3. Complex conjugation

(7.26)

7.2.4. Integration
In the event that f and g are integrable function, then the essential of their convolution all
in all space is just gotten as the result of their integrals:
( )() = ( ()) ( ())

(7.27)

This takes after from Fubini's theorem. The same result holds if f and g are just thought
to be nonnegative quantifiable capacities, by Tonelli's theorem.

7.2.5. Differentation

In the one-variable case,

( ) =

27

(7.28)

where d/dx is the subsidiary. Almost for the most part, on account of elements of a few
variables, a similar to recipe holds with the incomplete subordinate:

( ) =

(7.29)

A specific outcome of this is the convolution can be seen as a "smoothing" operation: the
convolution of f and g is differentiable the same number of times as f and g are
altogether.
These characters hold under the exact condition that f and g are totally integrable and no
less than one of them has a completely integrable (L1) frail subsidiary, as a result of
Young's disparity. For example, when f is constantly differentiable with minimal backing
and g is a discretionary locally integrable function,

( ) =

(7.30)

These characters additionally hold a great deal all the more comprehensively in the
feeling of tempered dispersions in the event that one of f or g is a minimally upheld
appropriation or a Schwartz function and the other is a tempered dissemination. Then
again, two positive integrable and vastly differentiable functions may have a no place
constant convolution.

In the discrete case, the distinction administrator D f(n) = f(n + 1) f(n) fulfils a
practically equivalent to relationship:
( ) = () = ()

28

(7.31)

7.3. CONVOLUTION INTEGRAL IN TERM OF MEMORY

An integral of the type

() = ( )()

Such as arose in the convolution theorem, is called a convolution integral.

Typically, f acts as some kind of cause such as a driving force or voltage, and
x(t) stand for certain effect produced.
Choose a time t for observation; then divide the interval = 0 to = into a
large number of equal time steps . We have

() = ( )()
0

( )()

=0

Now we choose any moment 1 between 0 and t: there was a force (1 ) applied
at this moment, and its contribution to x at time > is
( 1 )(1 )
The factor ( 1 ) takes into account the time elapse between the cause and its
effect in some problems is would be appropriate to call 1 the age of (1 )
at the moment t of observation, and g an ageing factor. Depending on the type of
problem, this factor might weaken or amplify the contributions of (1 )to the
integral as time t passes. The elapse time is increase if either we take an earlier
(0 ) or delay the time of observation by increasing t (D.W. Jordan & P. Smith,
2002).

29

8. THE CORRELATION THEOREM

The correlation theorem is firmly identified with the convolution theorem, and it likewise
ends up being helpful in numerous calculations. The correlation theorem is an outcome
that applies to the co, which is a necessary that has a definition reminiscent of the
convolution basic.

8.1. WHAT IS A CORRELATION FUNCTION?

In a correlation integral, rather than taking the estimation of one function at u-x, you take
the estimation of that function at x-u. Comparably, you take the estimation of the other
function at x+u. This is appeared in the accompanying mathematical statement, alongside

the variable substitution that permits the two expressions to be interconvert.

= ()( ) ;

(8.1)

Substitute = +

= ( )( )

= ( )()

(8.2)
(8.3)

The figure underneath delineates why the relationship function has the name that it does.
In the event that the two function f and g contain comparative components, however at an
alternate position, the connection function will have an expansive worth at a vector
relating to the movement in the position of the element.

30

Figure 2: Correlation function

The relationship hypothesis can be expressed in words as takes after: the Fourier
transform of a connection indispensable is equivalent to the result of the unpredictable
conjugate of the Fourier change of the principal transform and the Fourier transform of
the second function. The main distinction with the convolution theorem is within the
sight of an unpredictable conjugate, which turns around the stage and compares to the
reversal of the contention u-x.
= ( ) =

(8.4)

9. CONVOLUTION IN PARSEVAL'S THEOREM

9.1. Convolution with a Gaussian

To start with we have to characterize a Gaussian function. We will stick, for the
occasion, to 1D Gaussians. The Gaussian capacity is the recognizable chime moulded
bend, with a crest position (r0) and standard deviation.
() =

( )

(9.1)

We won't derive the Fourier transform of a Gaussian, but it is given by the following
equation.

31

(()) = ( )( )

(9.2)

Note that the Fourier transform of a Gaussian is another Gaussian (albeit without the
standardization steady). There is a stage term, relating to the position of the focal point of
the Gaussian, and afterward the negative squared term in an exponential. Likewise see
that the standard deviation has moved from the denominator to the numerator. This
implies, as a Gaussian in genuine space gets more extensive, the comparing Gaussian in
equal space gets smaller, and the other way around. This bodes well, things being what
they are: as the Gaussian in genuine space gets more extensive, commitments from
focuses inside that Gaussian begin to meddle with each other at lower and lower
resolutions.
Convolution with a Gaussian will move the birthplace of the function to the position of
the top of the Gaussian, and the function will be spread out, as outlined previously.
10. CONVOLUTION WITH A DELTA FUNCTION
Delta functions have an extraordinary part in Fourier hypothesis, so it merits investing
some energy getting to know them. A delta functions is characterized as being zero all
over the place however for a solitary point, where it has a weight of solidarity.
( ) = {

=
,

Saying that it has a weight of solidarity is that the indispensable of the delta function over
all space is 1. The delta function is given a contention of r-r0 with the goal that it can be
characterized as having its non-zero point at the beginning. At the point when r is
equivalent to r0, the contention of the delta function is zero.
( ) =
()( ) = ( )

(10.1)
(10.2)

A more broad property of the delta function is that the basic of a delta function times
some other function is equivalent to the estimation of that other function at the position
of the delta function.

32

In what manner can a solitary point, with no width, broadness or profundity, have a
weight of one? The estimation of the delta function by then should be an extraordinary
sort of boundlessness, and this implies it must be characterized as a point of confinement.
There are various approaches to characterize a delta function. One of them is to
characterize it as a boundlessly sharp Gaussian. The fundamental over all space of a
Gaussian is 1, which fulfils one of the properties required for the delta function, and on
the off chance that we take the utmost of a Gaussian as the standard deviation tends to
zero, it fulfils alternate properties. The accompanying mathematical statement
characterizes a 3D delta function as the farthest point of an isotropic 3D Gaussian.

( ) =
(( )/ (

| |

))

(10.3)

With this meaning of the delta function, we can utilize the Fourier transform of a
Gaussian to decide the Fourier transform of a delta function. As the standard deviation of
a Gaussian tends to zero, its Fourier transform has a tendency to have a steady extent of
1. All that is left is the stage shift term.
[( )] = (. )

(10.4)

So we see that the Fourier transform of a delta function is only a stage term. Consider the
photo we had of an electron at a point; it contributed only a stage term, with unit weight,
to the diffraction design. So now we see that we can consider an electron at a point to be
a delta capacity of electron thickness.

At last we can consider the significance of the convolution of a function with a delta
function. On the off chance that we record the mathematical statement for this
convolution, and remember the property of integrals including the delta function, we see
that convolution with a delta work essentially moves the inception of a function.
( )( ) = ( )
( )() = ( )

33

(10.5)
(10.6)

11. APPLICATIONS OF THE CONVOLUTION THEOREM

Convolution and related operations are found in many applications in science,

engineering and mathematics.
11.1.

In Image Processing

In advanced image processing, convolutional sifting assumes a critical part in numerous

essential calculations in edge discovery and related procedures.
In optics, an out-of-focus photo is a convolution of the sharp image with a lens function.
The photographic term for this is bokeh.
Usually, this theorem is used in picture preparing applications, for example, including
obscuring.
11.2.

In computerized information preparing

In analytical chemistry, SavitzkyGolay smoothing channels are utilized for the

investigation of spectroscopic information. They can enhance sign-to-noise ratio with
negligible twisting of the spectra.
In statistic, a weighted moving average is a convolution.

11.3.

In Acoustics

In acoustics, resonation is the convolution of the first solid with echoes from articles
encompassing the sound source.
In computerized signal preparing, convolution is utilized to delineate motivation reaction
of a genuine room on an advanced sound sign.
In electronic music convolution is the inconvenience of an unearthly or cadenced
structure on a sound. Regularly this envelope or structure is taken from another sound.
The convolution of two signs is the sifting of one through the other.
34

11.4.

Electrical Designing

In electrical designing, the convolution of one function (the information signal) with a
second function (the motivation reaction) gives the yield of a direct time-invariant
framework (LTI). At any given minute, the yield is an aggregated impact of all the earlier
estimations of the info capacity, with the latest values regularly having the most impact
(communicated as a multiplicative component). The drive reaction function gives that
component as an element of the past time subsequent to every information esteem
happened.

11.5.

Material Science

In material science, wherever there is a straight framework with a "superposition

standard", a convolution operation shows up. Case in point, in spectroscopy line
widening because of the Doppler impact all alone gives a Gaussian otherworldly line
shape and crash expanding alone gives a Lorentzian line shape. At the point when both
impacts are agent, the line shape is a convolution of Gaussian and Lorentzian, a Voigt
function.
In Time-determined fluorescence spectroscopy, the excitation sign can be dealt with as a
chain of delta heartbeats, and the deliberate fluorescence is an entirety of exponential rots
from every delta beat.
In computational liquid progression, the extensive swirl reproduction (LES) turbulence
model uses the convolution operation to bring down the scope of length scales essential
in calculation in this manner decreasing computational expense.

35

11.6.

In Likelihood Theorem

In likelihood hypothesis, the likelihood dispersion of the aggregate of two autonomous

irregular variables is the convolution of their individual conveyances.
In bit thickness estimation, dispersion is evaluated from test focuses by convolution with
a bit, for example, an isotropic Gaussian. (Diggle 1995).

11.7.

Radiotherapy Treatment

In radiotherapy treatment arranging frameworks, most piece of all current codes of

estimation applies a convolution-superposition algorithm.[clarification needed]

11.8.

Convolutional Neural System

Convolutional neural systems apply numerous fell convolution pieces with applications
in machine vision and counterfeit consciousness

11.9.

Atomic scattering factors

Before this, we have seen it essentially. We can tabulate atomic scattering factors by
working out the diffraction pattern of different atoms placed at the origin. Then we can
apply a phase shift to place the density at the position of the atom. Our new interpretation
of this is that we are convoluting the atomic density distribution with a delta function at
the position of the atom.
11.10.

B-factors

We can think of thermal motion as smearing out the position of an atom, i.e. convoluting
its density by some smearing function. The B-factors (or atomic displacement
parameters, to be precise) correspond to a Gaussian smearing function. At resolutions
typical of protein data, we are justified only in using a single parameter for thermal
36

motion, which means that we assume the motion is isotropic, or equivalent in all
directions. (In crystals that diffract to atomic resolution, more complicated models of
thermal motion can be constructed, but we won't deal with them here.)
Above, we worked out the Fourier transform of a 1D Gaussian.
(()) = ( ) ( )

(11.1)

In fact, all that matters is the displacement of the atom in the direction parallel to the
diffraction vector, so this equation is suitable for a 3D Gaussian. All we have to
remember is that the term corresponding to the standard deviation refers only to the
direction parallel to the diffraction vector. Since we are dealing with the isotropic case,
the standard deviation (or atomic displacement) is equal in all directions.
The B-factor is used in an equation in terms of sin/ instead of the diffraction vector,
because all that matters is the magnitude of the diffraction vector. We replace the
variance (standard deviation squared) by the mean-square displacement of the atom in
any particular direction. The B-factor can be defined in terms of the resulting equation.
|| =

= /

(()) = ( /
=

(11.2)
(11.3)
(11.4)

Note that there is a common source of misunderstanding here. The mean-square atomic
displacement refers to displacement in any particular direction. This will be equal along
orthogonal x, y and z axes. But often we think of the mean-square displacement as a
radial measure, i.e. total distance from the mean position. The mean-square radial
displacement will be the sum of the mean-square displacements along x, y and z; if these
are equal it will be three times the mean-square displacement in any single direction. So
the B-factor has a slightly different interpretation in terms of radial displacements.

37

 = + + =
=

11.11.

(11.5)

(11.6)

Diffraction from a lattice

The convolution theorem can be utilized to clarify why diffraction from a cross section
gives another grid specifically why diffraction from a cross section of unit cells in
genuine space gives a grid of structure variables in corresponding space. The Fourier
transform of an arrangement of parallel lines is an arrangement of focuses, opposite to
the lines and isolated by a separation conversely relative to the space between the lines.
(This is identified with the possibility that diffraction from an arrangement of Bragg
planes can be portrayed as far as a diffraction vector in corresponding space, opposite to
the arrangement of planes. Truth be told, we can expand "n" in Bragg's law (n=2 d sin)
to get a progression of diffraction vectors for that arrangement of planes.)
In the figure beneath, one arrangement of firmly dispersed flat lines offers ascend to a
broadly divided vertical column of focuses. A second arrangement of all the more
generally space corner to corner lines offers ascend to an all the more firmly divided
column of focuses opposite to these lines. On the off chance that we increase one
arrangement of lines by another, this will give a variety of focuses at the crossing points
of the lines in the base part of the figure. The Fourier transform of this cross section of
focuses, which was gotten by increasing two arrangements of lines, is the convolution of
the two individual transform (i.e.rows of focuses) , which produces an equal grid.

38

Figure 3: Set of closely-spaced horizontal lines gives rise to a widely-spaced vertical row
of points
Obviously, the same contention can be connected to a 3D cross section.

11.12.

Diffraction from a crystal

A crystal comprises of a 3D cluster of rehashing unit cells. Scientifically, this can be

created by taking the substance of one unit cell and convoluting it by a 3D grid of delta
functions. The diffraction example is consequently the result of the Fourier transform of
the substance of one unit cell and the Fourier transform of the 3D grid. Since the
transform of a cross section in genuine space is a corresponding grid, the diffraction
example of the gem tests the diffraction example of a solitary unit cell at the purposes of
the equal cross section.

11.12.1.

Determination truncation

Truncating the determination of the information utilized as a part of computing a

thickness guide is proportionate to taking the whole diffraction design and increasing the
structure components by a capacity which is one inside the constraining circle of
39

determination and zero outside the restricting circle. The impact on the thickness is
proportional to taking the thickness that would be gotten with every one of the
information and convoluting it by the Fourier transform of a circle. Presently, the Fourier
transform of a circle has a width contrarily corresponding to the sweep of the circle in
this way, the littler the circle (i.e. the lower the constraining determination), the more the
points of interest of the guide will be spread out (in this way decreasing the determination
of the elements it appears). Furthermore, the Fourier transform of a circle has swells
where it goes negative and after that positive once more, so a guide figured with
truncated information will likewise have Fourier ripples. These will be especially solid
around locales of high thickness, for example, substantial particles.

11.12.2.

Missing information

So also, forgetting any part of the information set (e.g. in the event that you are feeling
the loss of a wedge of information in proportional space on the grounds that the precious
stone kicked the bucket before the information gathering was done, or you didn't gather
the cusp information) will prompt bends of the electron thickness that can be
comprehended through the convolution hypothesis. With respect to determination
truncation, missing information can be comprehended as duplicating the full arrangement
of information by one where you have included information and zero where it is absent.
The convolution hypothesis lets us know that the electron thickness will be adjusted by
convoluting it by the Fourier transform of the ones-and-zeros weight function. The more
efficient the loss of information (e.g. a missing wedge versus arbitrarily missing
reflections), the more methodical the mutilations will be.

40

11.12.3.

DENSITY MODIFICATION

11.12.3.1. Solvent flattening

The presentation on density modification demonstrates that the impact of solvent

flattening can be comprehended as far as the convolution theorem. The impact of solvent
flattening is that every structure component after dissolvable straightening is gotten by
taking the weighted normal of its neighbours (in the corresponding cross section) before
solvent flattening, with the weighting terms being acquired as the Fourier transform of
the veil around the protein. So the littler the protein veil (i.e. the greater the dissolvable
part), the more extensive the components of its Fourier transform and in this manner the
all the more broadly the new stages counsel the old stages in the encompassing locale of
equal space. So also, if the cover is more definite, then its Fourier transform spreads all
the more generally in complementary space and there is all the more blending of stage
data.
11.12.3.2. Sayre's equation
The same presentation demonstrates that Sayre's comparison can be considered as a great
instance of dissolvable straightening, in which each atom gives its own veil.

12.APPLICATIONS OF THE CORRELATION THEOREM

12.1.

THE PATTERSON FUNCTION

The Patterson function is just the connection function of the electron thickness with
itself. By the correlation theorem, this relationship is registered by taking the Fourier
transform of F times its unpredictable conjugate F*, i.e. the Fourier transform of |F|2. At
positions in the Patterson map comparing to vectors between tops of thickness in the
electron thickness map, there will be crests in light of the fact that the relative

41

interpretation of the two maps in the correlation function will put one crest on top of the
other.

The staged translation function

The staged translation function is the connection of one thickness map with another, as
an element of an translation shift connected to one of the maps. This can be figured,
utilizing the relationship hypothesis, by taking the Fourier transform of the structure
component for one guide times the intricate conjugate of the structure variable for the
other guide, i.e. by subtracting the stages. Points of interest are given in the presentation
on atomic substitution.

13. TYPE OF CONVOLUTION

13.1.

Circular convolution

At the point when a function gT is occasional, with period T, then for functions, f, such
that fgT exists, the convolution is additionally intermittent and indistinguishable to:

( () +[
= ( + )] ( ),

(13.1)

where to is a subjective decision. The summation is known as an intermittent summation

of the function f.
At the point when gT is an occasional summation of another function, g, then fgT is
known as a circular or cyclic convolution of f and g.
What's more, if the intermittent summation above is supplanted by fT, the operation is
known as an occasional convolution of fT and gT.

42

13.2.

Discrete convolution

For complex-esteemed function f, g characterized on the set Z of numbers, the discrete

convolution of f and g is given by:
( )[]
= [][ ]

= [ ][]

(13.2)

(13.3)

The convolution of two limited successions is characterized by extending the groupings

to limitedly upheld capacities on the arrangement of whole numbers. At the point when
the arrangements are the coefficients of two polynomials, then the coefficients of the
common result of the two polynomials are the convolution of the first two successions.
This is known as the Cauchy product of the coefficients of the sequences.
Along these lines when g has limited backing in the set {-M,-M+1,.....,M-1,M} , a limited
summation might be used:
( )[] =
= [ ][]

(13.4)

For example, a finite impulse responses.

13.3.

Circular discrete convolution

At the point when a function gN is occasional, with period N, then for function, f, such
that fgN exists, the convolution is additionally intermittent and indistinguishable to:

( )[]
=(= [ + ]) [ ]

(13.5)

The summation on k is known as an intermittent summation of the function f.

In the event that gN is an intermittent summation of another function, g, then fgN is

known as a circular convolution of f and g.

43

At the point when the non-zero lengths of both f and g are constrained to the interim [0,
N 1], fgN decreases to these basic structures:

( )[] =
= [] [ ]
= = [] [ ] +
=+ [] [ + ]
=
= [] [ ] ( )[]

(13.6)
(13.7)
(13.8)

The documentation (f N g) for cyclic convolution signifies convolution over the cyclic
gathering of whole numbers.
Circular convolution emerges regularly with regards to quick convolution with a FFT
calculation.
13.4.

Fast convolution algorithms

Much of the time, discrete convolutions can be changed over to circular convolutions so
fast transform with a convolution property can be utilized to execute the calculation. For
instance, convolution of digit arrangements is the bit operation in augmentation of multidigit numbers, which can in this manner be proficiently actualized with change strategies.
Eq.1 requires N number-crunching operations per yield worth and N2 operations for N
yields. That can be fundamentally decreased with any of a few quick calculations.
Computerized signal preparing and different applications normally utilize quick
convolution calculations to decrease the expense of the convolution to O (N log N)
many-sided quality.
The most widely recognized quick convolution calculations utilize quick Fourier
transform (FFT) calculations through the circular convolution theorem. In particular, the
circular convolution of two limited length successions is found by taking a FFT of every
arrangement, increasing pointwise, and after that performing a backwards FFT.
Convolutions of the sort characterized above are then productively actualized utilizing
that method as a part of conjunction with zero-augmentation and/or disposing of bits of

44

the yield. Other quick convolution calculations, for example, the SchnhageStrassen
calculation or the Mersenne transform, utilize quick Fourier transform in different rings.

On the off chance that one grouping is any longer than the other, zero-expansion of the
shorter succession and fast circular convolution is not the most computationally
proficient strategy available. Instead, decaying the more drawn out arrangement into
pieces and convolving every square takes into account quicker calculations, for example,
the Overlapsave technique and Overlapadd method. A half breed convolution
technique that consolidates piece and FIR calculations takes into consideration a zero
information yield inactivity that is helpful for continuous convolution computations.

14. CONVOLUTION OF TWO COMPLEX-VALUED FUNCTIONS

The convolution of two complex-valued functions on Rd is itself a complex-valued
function on Rd, defined by:
( )() = ()( ) = ( )()

(14.1)

is all around characterized just if f and g rot adequately quickly at endlessness all
together for the necessary to exist. Conditions for the presence of the convolution might
be precarious, since an explode in g at interminability can be effectively balanced by
adequately quick rot in f. The subject of presence therefore may include diverse
conditions on f and g:
Compactly supported functions
In the event that f and g are compactly supported continuous functions, then their
convolution exists, and is additionally compactly upheld and continuous. In general, if
either work (say f) is compactly upheld and the other is locally integrable, then the
convolution fg is very much characterized and consistent.

45

Convolution of f and g is additionally all around characterized when both capacities are
locally square integrable on R and support on an interim of the structure [a, +) (or both
supported on [-, a]).
Integrable function
The convolution of f and g exists if f and g are both Lebesgue integrable function in
L1(Rd), and for this situation fg is likewise integrable (Stein and Weiss 1971, Theorem
1.3). This is a result of Tonelli's hypothesis. This is additionally valid for capacities in ,
under the discrete convolution, or all the more for the most part for the convolution on
any gathering(Jordan.D.W, Smith. P. 2002).
Likewise, if f L1(Rd) and g Lp(Rd) where 1 p , then fg Lp(Rd) and

(14.2)

In the specific case p = 1, this demonstrates L1 is a Banach variable based math under the
convolution (and equity of the two sides holds if f and g are non-negative all over).
All the more for the most part, Young's inequalities suggest that the convolution is a
constant bilinear guide between appropriate Lp spaces. In particular, if 1 p,q,r fulfil

+ = +

(14.3)

Then

so that the convolution is a continuous bilinear mapping from LpLq to Lr. The Young
imbalance for convolution is likewise valid in different connections (circle bunch,
convolution on Z). The previous disparity is not sharp on the genuine line: when 1 < p, q,
r < , there exists a steady Bp, q < 1 such that:
,
The optimal value of Bp, q was discovered in 1975.
46

, ,

A stronger approximation is true provided 1 < p, q, r < :

, ,
where is the feeble Lq standard. Convolution likewise characterizes a bilinear consistent
guide , , ,

1 < , , < attributable to the frail Young

disparity
, , , ,
Functions of rapid decay
Notwithstanding compactly upheld capacities and integrable function, works that have
adequately quick root at interminability can likewise be convolved. An imperative
component of the convolution is that if f and g both rapid decay, then fg additionally
rapid decay. Specifically, if f and g are rapid decay function, then so is the convolution
fg. joined with the way that convolution drives with separation (see Properties), it takes
after that the class of Schwartz function is shut under convolution (Stein and Weiss,
1971).

Distribution
Under a few circumstances, it is conceivable to characterize the convolution of a capacity
with a circulation, or of two disseminations. On the off chance that f is a minimalistically
upheld capacity and g is dissemination, then fg is a smooth capacity characterized by a
distributional recipe similar to
()( )

(14.4)

All the more by and large, it is conceivable to broaden the meaning of the convolution
uniquely so that the cooperative law
( ) = ( )

47

(14.5)

stays substantial for the situation where f is a dispersion, and g a minimally upheld
conveyance (V.I. Sobolev.2013).

15. CONVOLUTION FUNCTION

The convolution function expresses that

{ } = . {}. {}

(15.1)

where signifies the Fourier transform of , and is a steady that relies on upon the particular
standardization of the Fourier transform. Adaptations of this theorem likewise hold for
the Laplace transform, two-sided Laplace transform, Z-change and Mellin transform.
Translation invariance

The convolution drives with translation, implying that

( ) = ( ) = ( )

(15.2)

where is the interpretation of the function f by x characterized by

In the event that f is a Schwartz capacity, then xf is the convolution with a deciphered
Dirac delta function = . So interpretation invariance of the convolution of
Schwartz function is a result of the associativity of convolution.
Besides, under certain conditions, convolution is the most broad interpretation invariant
operation. Casually, the accompanying holds;
Suppose that S is a direct administrator following up on functions which drives with
interpretations:
( ) = ()
48

(15.3)

for all x. At that point S is given as convolution with a function (or dissemination) gS;
that is = .
Consequently any interpretation invariant operation can be spoken to as a convolution.
Convolutions assume an essential part in the investigation of time-invariant system, and
particularly LTI system theorem. The function gS is representing the motivation reaction
of the change of S.
A more exact adaptation of the theory cited above requires determining the class of
function on which the convolution is characterized, furthermore requires expecting also
that S must be a continuous linear operator as for the proper topology. It is known, for
occurrence, that each nonstop interpretation invariant constant direct administrator on L1
is the convolution with a limited Borel measure. All the more for the most part, every
consistent interpretation invariant persistent straight administrator on Lp for 1 p < is
the convolution with a tempered distribution whose Fourier transform is limited. To
mind, they are all given by limited Fourier multipliers.

16. CONVOLUTIONS ON GROUPS

On the off chance that G is an appropriate group enriched with a measure , and if f and g
are genuine or complex esteemed integrable function on G, then we can characterize their
convolution by
( )() = ()( )()

(16.1)

It is not commutative as a rule. In ordinary instances of interest G is a locally smaller

Hausdorff topological group and is a (left-) Haar measure. All things considered, unless
G is unimodular, the convolution characterized along these lines is not the same as
( 1 )()() . The inclination of one over the other is made so that convolution
with an altered capacity g drives with left interpretation in the gathering:
( ) = ( )

49

(16.2)

Moreover, the tradition is likewise required for consistency with the meaning of the
convolution of measures given underneath. Be that as it may, with a privilege rather than
a left Haar measure, the last vital is favoured over the previous.
On locally minimized abelian bunches, a variant of the convolution hypothesis holds: the
Fourier change of a convolution is the point wise result of the Fourier changes. The circle
bunch T with the Lebesgue measure is a quick illustration. For a settled g in L1(T), we
have the accompanying commonplace administrator following up on the Hilbert space
L2(T):

() = ()( )

(16.3)

The administrator T is conservative. An immediate figuring demonstrates that its adjoint

T* is convolution with
()

By the commutative property referred to above, T is ordinary: T*T = TT*. Likewise, T

drives with the interpretation administrators. Consider the family S of administrators
comprising of every single such convolution and the interpretation administrators. At that
point S is a driving group of ordinary administrators. As per unearthly hypothesis, there
exists an orthonormal premise {hk} that at the same time diagonalizes S. This describes
convolutions on the circle. In particular, we have
() = ,

(16.4)

which are exactly the characters of T. Every convolution is a minimal duplication

administrator in this premise. This can be seen as a form of the convolution hypothesis
talked about above.
A discrete case is a limited cyclic gathering of request n. Convolution administrators are
here spoken to by circulated networks, and can be diagonalized by the discrete Fourier
transform.
A comparable result holds for minimal gatherings (not as a matter of course abelian): the
grid coefficients of limited dimensional unitary representations frame an orthonormal
50

premise in L2 by the PeterWeyl hypothesis, and a simple of the convolution theorem

keeps on holding, alongside numerous different parts of symphonious examination that
rely on upon the Fourier transform.

17. CONVOLUTION OF MEASURES

Give G a chance to be a topological gathering. In the event that and are limited Borel
measures on G, then their convolution is characterized by
( )() = ()()()

(17.1)

for each quantifiable subset E of G. The convolution is likewise a limited measure,

whose aggregate variety fulfils the accompanying rendition of Young's disparity

where the standard is the aggregate variety of a measure. Since the space of measures of
limited variety is a Banach space, convolution of measures can be treated with standard
techniques for utilitarian examination that may not make a difference for the convolution
of disseminations.

For the situation when G is locally minimized with (left-) Haar measure , and and
are completely persistent concerning a , so that each has a thickness capacity, then the
convolution is likewise totally ceaseless, and its thickness capacity is only the
convolution of the two separate thickness capacities.
In the event that and are likelihood measures on the topological gathering (R,+), then
the convolution is the likelihood circulation of the aggregate X + Y of two
autonomous arbitrary variables X and Y whose individual conveyances are and .

51

The convolution of any two Borel measures and of limited variety is the measure
characterized by (Rudin 1962)

()() = ( + )()()

(17.1)

This concurs with the convolution characterized above when and are viewed as
appropriations, and additionally the convolution of L1 functions when and are totally
persistent as for the Lebesgue measure.

18. BIALGEBRAS
Let (X, , , , ) be a bialgebra with comultiplication , duplication , unit , and
counit . The convolution is an item characterized on the endomorphism polynomial
math End(X) as takes after. Let , End(X), that is, , : X X are functions that
regard all arithmetical structure of X, then the convolution is characterized as the
creation

(18.1)

The convolution shows up remarkably in the meaning of Hopf algebras (Kassel 1995,
III.3). A bialgebra is a Hopf polynomial math if and just in the event that it has an
antipode: an endomorphism S such that
=

(18.2)

19. TITCHMARSH CONVOLUTION THEOREM

The Titchmarsh convolution theorem is named after Edward Charles Titcmarsh Edward,
a British mathematician. The theorem describes the properties of the support of the
convolution of two functions.

52

E.C. Titchmarsh demonstrated the accompanying hypothesis in 1926:

If () and ()are integrable functions, such that

()( ) =

(19.1)

Almost everywhere in the interval of0 < < , then there exist 0 and 0
fulfilling such that + such that () = 0 almost all over in (0,) and () = 0
almost all around in (0,)
This outcome, known as the Titchmarsh convolution hypothesis, can be restated in the
accompanying structure:
Let, 1 (). Then, = inf + inf if the right-hand side
is limited.
Thus, if = sup + sup the right-hand side is limited.
This theorem basically expresses that the surely understood consideration
+
is sharp at the limit.
The higher-dimensional speculation as far as the arched structure of the backings was
demonstrated by J.- L. Lions in 1951:
If, ( ), then . . = . . + . .
Now . , signifies the raised frame of the set means the space of conveyances with
conservative backing.
The theorem does not have a basic confirmation. The first verification by Titchmarsh
depends on the PhragmnLindelf guideline, Jensen's disparity, the Theorem of
Carleman, and Theorem of Valiron (Raouf Doss, 1990).

53

20. CONVOLUTION AS A METHOD TO CALCULATE ZERO STATE

RESPONSE
Convolution is a capable procedure that can be utilized to figure the zero state
reaction (i.e., the reaction to an information when the framework has zero starting
conditions) of a framework to a subjective contribution by utilizing the
motivation reaction of a framework. It utilizes the force of linearity and
superposition. To comprehend the force that this strategy gives, consider the
framework portrayed by the differential mathematical statement:
= ()
() + ()

where y(t) is the output, and f(t) is the input. The impulse response of this
system, h(t), can be shown to be equal to

() = . ( . ) ,

>0

(From Laplace Transform Table)

The impulse response is shown graphically below.

Figure 4: Impulse Respond

You know how to discover the yield y(t) if the input f(t) is well characterized
input, like a stage, motivation or sinusoid. Convolution permits you to decide the
reaction to more mind boggling inputs like the one demonstrated as follows.

54

Actually, you can utilize convolution to discover the yield for any information, on
the off chance that you know the motivation reaction. This gives unfathomable
power

Figure 5: Impulse per unit time

There are a few approaches to see how convolution functions. Initially
convolution will be created in a surmised structure as the total of impulse
reactions. This presentation is helpful for an instinctive comprehension of the
convolution process. After the inexact structure is created, the accurate logical
type of convolution is given.
21. THE CONVOLUTION IS A SUM OF IMPULSE RESPONSES.

To see how convolution functions, we indicate the consistent function appeared

above by a discrete function, as demonstrated as follows, where we take an
example of the input at regular 0.8 seconds.

55

Figure 6: Discrete approximation of impulse responses

The estimate can be made a stride further by supplanting each rectangular piece
by an impulse as demonstrated as follows. The area of every impulse is the same
as the range of the relating rectangular square. (Since the width of the piece was
0.8, every impulse drop slightly lowers the function).

The superposition theorem expresses that the reaction of the output to the series
of system is only the total of the reaction to the individual impulse. The reaction
of the system to the individual impulse is demonstrated as follows.

56

Figure 7: The reaction of the system to the individual impulse

Lets take a look at this chart and note the accompanying:
The reaction of the system to a single impulse, (t), is h(t).
The reaction of the system to a single impulse deferred by time
, ( . ), ( . )
The reaction of the system to a impulse of territory A deferred by time
. , . ( . )
The stature of the function at once t=iT is f(iT).
The range of the impulse at t=iT is f(iT)T.
The deferred and moved motivation reaction is given by f(iT)Th(t-iT).
Therefore the reaction of the system to a progression of deferred impulse, as appeared
above is
() =
((. ))( . )
In this mathematical statement

y(t) is the yield

57

(21.1)

iT is the time deferral of every impulse

(f(iT)T) is the range of the its impulse

if you take the farthest point as T0, the summation yields the
convolution basic (with iT=, T=d)

() = ()( )

(21.2)

This is the Convolution Theorem.

The vital is regularly given points of confinement of positive and negative

boundlessness:

() = ()( )

(21.3)

For our motivations the two integrals are equal in light of the fact that f()=0 for
<0

The type of the indispensable we will regularly utilize is

() = ()( )

(21.4)

This structure is additionally proportional on the grounds that h(t-)=0 for

t-<0 (or h(t-)=0 for >t). This makes the breaking points of joining
successfully 0 and t.

The contentions in the necessary can likewise be changed to give two

identical types of the convolution vital

() = ( )() = ()( )

(21.5)

The convolution is frequently indicated by a reference bullet (*) Lets look

carefully at this graph and note the following:
() = () () = () ()

58

(21.6)

This comparison only expresses that the yield is equivalent to the total of the
reactions from the individual motivations. Another (more numerical) deduction of
the convolution vital is given here.

Clearly on the off chance that you diminish the exactness of the estimation ought to
progress. The chart beneath demonstrates the drive reactions on the off chance that
we let T=0.3 seconds (rather than 0.8 seconds as appeared previously). Note that
the individual reactions are littler, since the zone of every motivation is littler.

Figure 8: The impulse responses when T=0.3 seconds

Figure 9: The impulse responses when T=0.3 seconds and T=0.8

59

The chart underneath demonstrates our inexact reaction, for T=0.8 and T=0.3,
alongside the definite reaction (the "careful" reaction is really a numerical
estimate, as ascertained by Matlab).

The chart underneath demonstrates the "accurate" reaction alongside the reaction
computed utilizing Matlab "conv" (convolve) capacity. Sort "help conv" at the
Matlab brief to perceive how it functions. Analyze the script, convolution.m,
which created all these diagrams for a case of how to utilize the function.

Figure 10: Exact impulse response and respond using convolution

Ideally this foundation gives you some physical knowledge into how convolution
can be utilized to discover the reaction of the framework to subjective
information.
Convolution can be utilized to ascertain the zero state reaction (i.e., the reaction to
an info when the framework has zero introductory conditions) of a framework to a
subjective contribution by utilizing the drive reaction of a framework. Given a
framework drive reaction, h(t), and the information, f(t), the yield, y(t) is the
convolution of h(t) and f(t):
() = () () = () ()

( )() = ()( )

60

(21.7)
(21.8)

Frequently h(t)=0 for t<0 and f(t)=0 for t<0. All things considered:

() = ( )() = ()( )

(21.9)

22. VISUALIZING THE CONVOLUTION INTEGRAL

To imagine the convolution we begin from the definition:

() = ()( )

(22.1)

On the other hand identically, if h(t)=0, t<0 and f(t)=0, t<0.

Consider the system explained before with the motivation reaction depicted
by:

() = . ( . ) , > 0

as demonstrated as follows

Figure 11: impulse response

61

To discover the reaction of the system, y(t), to an input, f(t), at t=1 sec we have to
assess the integral

() = ()( )

(22.2)

with t=1. To picture this procedure consider the item f()h(t-). A plot of f()
yields the input. A plot of h(t-) yields a period turned around form of the impulse
with a period postponement of t (1 second for this situation). These are appeared
beneath alongside the result of the two functions (strong dark line). The function
h(t-) is plotted in margenta and doesn't look much like the function above - yet
take note of the distinctive vertical scales.

Figure 12: The function h(t-) and its product

The next graph shows the result of the integration for all time, with a black dot at
t=1.

62

Figure 13: Result of the integration for all time, with a black dot at t=1
As t is increased further the function h(t-) shifts to the right. The multiplication
and integration can be evaluated for each value of t to yield the system response
y(t). The graphs below are similar to the ones above, except that t=4.

Using convolution integral it is possible to calculate the output, y(t), of any linear
system given only the input, f(t), and the impulse response, h(t). However, this
integration is often difficult, so we won't often do it explicitly. Later you will
learn a technique that vastly simplifies the convolution process.

63

23. CONVOLUTION IN ISLAMIC PERSPECTIVE

23.1.

MUSLIM SCHOLAR IN MATHEMATICS

The Islamic Empire established across Persia, the Middle East, Central Asia, North
Africa, Iberia and parts of India from the 8th Century onwards made significant
contributions towards mathematics. They were able to draw on and fuse together the
mathematical developments of both Greek and India.
The Quran itself encouraged the accumulation of knowledge, and a Golden Age of
Islamic science and mathematics flourished throughout the medieval period from the 9th
to 15th Centuries. The House of Wisdom was set up in Baghdad around 810, and work
started almost immediately on translating the major Greek and Indian mathematical and
astronomy works into Arabic.

23.1.1. Al-Khawarizmi

One of the earliest and most distinguished of the Arabic mathematicians was the ninth
century scholar Abu Ja'far Mohammed ibn Musa Al-Khwarizmi, who was an astronomer
to the caliph at Baghdad. His name indicates that he was from the town of Khwarizm
(now Khiva), on the Amu Darya river, south of the Aral Sea in what is now Uzbekistan.
(Khwarizm was part of the Silk Route, a major trading pathway between Europe and the
East.) Al-Khwarizmi's full name can be translated as "Father of Ja'far, Mohammed, son
of Moses, native of the town of Al-Khwarizmi". Al-Khawarizmi wrote a mathematical
work called Al-Jabr wa Al-Muqabala (820 CE), from whose title is derived the name
algebra, this book may be considered the first book written on the topic of algebra.
Amongst the achievements that Al Khawarizmi left to posterity were: (1) Solutions to
64

first and second-degree equations with a single unknown, using both algebraic and
geometric methods. (2) A method of algebraic multiplication and division.
Al Khawarizmi defined three kinds of quantities: (1) Simple numbers, such as 5, 17 and
131. (2) The root which is the unknown quantity shay in Arabic meaning a thing
However, in translations made in Toledo, (the centre for translation of Arabic books), the
absence of a sh sound in the Spanish language meant that a suitable letter had to be
chosen. The choice fell upon x, which may well explain why Don Quixote is often
pronounced as Don Quishote. (3) Wealth (mal) the square of the root (x) (Zohor
Shanan Idrisi.2011).

Al-Khwarizmi wrote several books that were to be enormously influential. In particular,

his book describing how to write numbers and compute with them using the place-value
decimal system that came out of India would, when translated into Latin three hundred
years later, prove to be a major source for Europeans who wanted to learn the new
system.
In fact, Al-Khwarizmi's book on arithmetic with the Hindu-Arabic numbers was so
important, it appears to have been translated several times. Many translations began with
the phrase "dixit Algorismi" ("so says Al-Khwarizmi"), a practice that led to the adoption
in medieval times of the term algorismto refer to the process of computing with the
Hindu-Arabic numerals. Our modern word "algorithm" is an obvious derivation from that
term.
Another of Al-Khwarizmi's manuscripts was called Kitab al jabr w'al-muqabala, which
translates roughly as "restoration and compensation". The book is essentially an algebra
text. It starts off with a discussion of quadratic equations, then goes on to some practical
geometry, followed by simple linear equations, and ending with a long section on how to
apply mathematics to solve inheritance problems. The Englishman Robert of Chester
translated Al-Khwarizmi's algebra book from Arabic into Latin in 1145. The part dealing
with quadratic equations eventually became famous. Such was the influence of this work
that the Arabic phrase al jabr in the book's title gave rise to our modern word "algebra".
He developed sine, cosine and trigonometrically tables, which were later translated to the
65

West. He also described six basic types of equations: nx = m, x^2= nx, x^2=m, m+x^2 =
nx,

m + nx +x^2 and x^2 = m+nx (Muslim Founder in Mathematics, 2007 from

http://www.muslimheritage.com/article/muslim-founders-mathematics).
After Al-Khwarizmi, algebra became an important part of Arabic mathematics. Arabic
mathematicians learned to manipulate polynomials, to solve certain algebraic equations,
and more. For modern readers, used to thinking of algebra as the manipulation of
symbols, it is important to realize that the Arabic mathematicians did not use symbols at
all. Everything was done in words.
23.1.2. AL- BIRUNI

AL-BIRUNI

Al-Biruni was among those who laid the foundation for modern trigonometry. He was a
philosopher, geographer, astronomer, physicist and mathematician. Six hundred years
before Galileo, Al-Biruni discussed the theory of the earth rotating about its own axis.
Al-Biruni carried out geodesic measurements and determined the earth's circumference in
a most ingenious way. With the aid of mathematics, he enabled the direction of the Qibla
to be determined from anywhere in the world.
In the domain of trigonometry, the theory of the functions; sine, cosine, and tangent was
developed by Muslim scholars of the tenth century. Muslim scholars worked diligently in
the development of plane and spherical trigonometry. The, trigonometry of Muslims is
based on Ptolemy's theorem but is superior in two important respects: it employs the sine
where Ptolemy used the chord and is in algebraic instead of geometric form.
66

He defined the ratios between the densities of gold, mercury, lead, silver, bronze, copper,
brass, iron, and tin. Al-Biruni displayed the results as combinations of integers and
numbers of the form 1/n, n = 2, 3, 4... 10.
Al-Brn's production exceeds 146 titles in more than twenty different disciplines,
ranging from astronomy to mathematics, mathematical geography, chronology,
mechanics, pharmacology, mineralogy, history, literature, religion, and philosophy. Only
22 works have survived the ravages of time; and only a dozen or so of these have been
published .
Though al-Brn dedicated himself to astronomy, he excelled also in mathematics in an
age when mathematics consisted of arithmetic, geometry, physics and music . Other than
being the master of geodesy, that is mathematical geography, al-Brn made simplified
stereographic projection similar to that first published by G. B. Nicolosi di Paterno in
1660 . However, as Max Meyerhof observed earlier this century, most of al-Brn 's
mathematical works and many other writings have not been published yet.

24.CONCLUSIONS
Convolution is a vital theorem in mathematics and physics. It is the theorem that
helps us in order to solve complex mathematical problem. There are a few type
of convolution which is circular and discrete convolution which has their specific
characteristics. The importance of convolution theorem in our life has been
proven by scientists and engineers that imply this theorem in order to make the
research and technology such as in Radiotherapy Treatment and image
processing. Other than that, this paper also discuss on the mathematics in the
Islamic perspective which describe the Islamic scholars that give great
contributions in mathematics. We must thanks to Allah because give us the
Islamic scholar with the great knowledge of mathematics that help us in real life.
Other than that, we also must take part in study and explore the knowledge in
Quran in order to increase our faith to Allah.

67

25.REFERENCES
(n.a)(2007)Muslim Founder In Mathematics. FSTC Limited. Retrieved from
http://www.muslimheritage.com/article/muslim-founders-mathematics on 7 May 2016.
(n.a)(n.d)Al-Biruni retrieved from http://www.muslimheritage.com/article/al-biruni on
7 May 2016
(n.a). 2006 Circular Convolution. Massachusetts Institute of Technology Department of
Electrical

Engineering

and

Computer

Science.

Pg

1-5.

http://ocw.mit.edu/courses/electrical-engineering-and-computer-science/6-341-discretetime-signal-processing-fall-2005/lecture-notes/lec16.pdf
A. A. Komech and A. I. Komech. 2011.On the Titchmarsh Convolution Theoremfor
Distributions on the Circle Texas A&M University.Institute for Information
Transmission Problems, Moscow.pp.1-6
A. Borsellino and T. Poggio. 1973. Convolution and Correlation Algebras Kybernetik
13, 113-122(1973). pg: 1-10.
Akshay Behl, Akash Bhatia, Avril Puri 2014.Convolution And Applications Of
Convolution. Dronacharya College Of Engineering, GGN. Pg 2122-2126. Internatonal
Journal Of Innovative Research In Technology. 2014 IJIRT | Volume 1 Issue 6 | ISSN
: 2349-6002
Alejandro Dominguez.2015.A History of the Convolution Operation. Retvieved from
http://pulse.embs.org/january-2015/history-convolution-operation/ on 7 May 2016
Ali Abdullah Al-Daffa (1977). The Muslim Contribution to Mathematics
Crutchfield, Steve (October 9, 2010), "The Joy of Convolution", Johns Hopkins
University, retrieved November 19,2010
Daniele Rosso & Victor Ginzburg .(N.D).Convolution Algebras and Applications to
Representation Theory pg 1-9
Ibid.pg:482
68

Ibid.pg:535-538
Erwin Kreyszig. 1999. Advanced Engineering Mathematics. p.279
Jordan.D.W, Smith. P. 2002. Mathematical Techniques- An Introduction for the
Engineering, Physical, And Mathematical Sciences. 3rd edition. Published in the United
States by Oxford University Press Inc., New York. Pp:479-480.
Kammler.D.W. 2007. A First Course in Fourier Analysis. United Kingdom. Pearson
Education, Inc.. 2nd Edition. p.89-115
Madalina Roxana Buneci .2011.Convolution Algebras For Topological Groupoids With
Locally Compact Fibres Opuscula Mathematica _ Vol. 31 _ No. 2. Pg : 1-14
Michael

Ross.2014.

Joy

Of

Convolution

retrieved

from

http://pages.jh.edu/~signals/discreteconv2/ on 7 May 2016

Prof Zemanek. Al-Khowarizmi: His Background, His Personality, His Work and His
Influence. Austria Lecture notes in Computer Science, Vol. 122, 1981. Springer Verlag,
NewYork.
R.R. Wright.1938. Al-Brn, Kitb al-tafhm li-'aw'il sin'at al-tanjm. Book of the
Instruction in the Elements in the Science of Astrology, edited, London: Luzac.
Randy J Read.2009.The Convolution Theorem And Its Applications. Retrieved from
http://www-structmed.cimr.cam.ac.uk/Course/Convolution/convolution.html on 7 May
2016.
Raouf Doss,1990, Two Generalizations Of Titchmarshs Convolution Theorem Pg 1-5
Spanier. E. H. 1966. Real and Complex Analysis. United Kingdom. Pearson Education,
Inc.. 2nd Edition. p.146
V.I. Sobolev.2013.Convolution of Functions. Encyclopedia of Mathematics. Retrieved
from
http://www.encyclopediaofmath.org/index.php?title=Convolution_of_functions&oldid=2
9335 on 7 May 2016.
69

Weisstein,

E. W.

"Books

about

Convolution."

Retrieved

from

http://www.ericweisstein.com/encyclopedias/books/Convolution.html on 7 May 2016

Weisstein, Eric W. "Convolution Theorem." From MathWorld--A Wolfram Web
Resource.http://mathworld.wolfram.com/ConvolutionTheorem.html
Weisstein,

Eric

W. 1999."Convolution."

From MathWorld--A

Wolfram

Web

Resource.retrieved from http://mathworld.wolfram.com/Convolution.html on 7 May

2016.
Zohor Shanan Idrisi.2011. A Brief Look At Islams Contribution to Mathematics.
Retrived

from

http://www.islamreligion.com/articles/4761/brief-look-at-islam-s-

contribution-to-mathematics/ on 7 May 2016.

70