Fourier Transforms
Fourier Transforms
Fourier Transforms
• Sufficient conditions for the existence of Fourier
Transform
• Sufficient conditions for the existence of Inverse Fourier
Transform
• Definition of Fourier Transform of f(x)
Prepared by:
Prof. Sunil
Department of Mathematics & Scientific Computing
NIT Hamirpur (HP)
Fourier Transform:
Sufficient conditions for the existence of Fourier Transform:
1. f(x) is piecewise continuous on every finite interval, and
2. f(x) is absolutely integrable on the x-axis.
Sufficient conditions for the existence of Inverse Fourier Transform:
1. F(s) is absolutely integrable over ( −∞, ∞ ) .
2. lim F ( s ) = 0
s →∞
**************************
Fourier transform of f(x):
The Fourier integral of f(x) in the complex form given by
1
∞
1 ∞
f (x) = ∫ ∫
2π −∞ 2π −∞
f (t)eist dt e − isx ds
1
∞
∞ − isx
⇒ f (x) = ∫ ∫ f (t)e dt e ds
ist
(i)
2π −∞ −∞
Fourier Transforms: 2
Visit: https://www.sites.google.com/site/hub2education/
Transform of f.
Since t is a dummy variable, we replace t by x.
Thus, we have
∞
1
∫ F ( s ) e ds ,
− isx
f (x) = (ii)
2π −∞
∞
∫ f (x)e
isx
where F (s ) = dx . (iii)
−∞
Remarks:
• F ( s ) and f (x ) are known as Fourier transform pair, which differs only in the sign of
the exponent.
1
• The factor can multiply in the expression of F ( s ) in integral (iii) instead of the
2π
expression of f(x) in integral (ii).
1 1
• Basically, the choice of normalizing factors or in integrals (ii) and (iii) is
2π 2π
optional. To make the two integrals as symmetric as possible, we can multiply
1
in both the expressions.
2π
∞
1
• We call f ( x ) = ∫
2π −∞
F ( s ) e − isx ds as an inversion formula corresponding to
∫ f (x)e
isx
F (s ) = dx .
−∞
∞
1
• Some times, we call f ( x ) = ∫
2π −∞
F ( s ) eisx ds as an inversion formula corresponding
∫ f (x)e
− isx
to F ( s ) = dx .
−∞
1
∞
∞ −isx ∞
1 ∞ −isx
∫ −∞∫ ∫ ∫
ist ist
f (x) = f (t)e dt e ds f (x) = f (t)e dt e ds
2π −∞ −∞ 2π −∞
∞ ∞
can be written as f (x ) = 1 ∫ F(s )e −isx ds , can be written as f (x ) = ∫ F(s )e ds ,
−isx
2π
−∞ −∞
∞ ∞
1
∫ f (x )e ∫ f (x )e dx .
isx isx
where F(s ) = dx . where F(s ) =
2π
−∞ −∞
or
1
∞
∞ −isx 1 ∞
∞
∫ ∫ f (t)e dt e ds
1
ist
f (x) =
∫ ∫
ist − isx
2π −∞ −∞ f (x ) = f ( t ) e dt e ds
2π − ∞ 2π − ∞
∞
can be written as f ( x ) = 1 ∫ F ( s ) eisx ds , ∞
***********************************
Fourier Transforms
1 − x 2 , if x < 1
Q.No.1.: Find the Fourier transform of f (x) =
0, if x > 1
∞
x cos x − sin x x
and use it to evaluate ∫
x 3
cos dx .
2
0
st
1 − x 2 , if x < 1
1 Part: Find Fourier transform of f (x) =
0, if x > 1
4 eis + e− is 4 eis − e − is
= − 2 + 3
s 2 s 2i
∞
4 s cos s − sin s
⇒ f (x) = − ∫ cos sxds ,
π0 s3
π 1 − x ,
∞ 2
s cos s − sin s π if x < 1
⇒∫ cos sxds = − f (x) = −
0 s3 4 4 0, if x > 1
1
Taking x = , we have
2
Fourier Transforms: 5
Visit: https://www.sites.google.com/site/hub2education/
π 1 π 1
∞ 2
s cos s − sin s s π 3 3π
∫0 s3
cos
2
ds = − = − 1 − = − . = − .
1 −
4 2 4 4 4 4 16
Replacing s by x, we get
∞
x cos x − sin x x 3π
∫
x 3
cos dx = − . Ans.
2 16
0
1 for x < 1
Q.No.2.: Find the Fourier transform of f ( x ) = .
0 for x > 1
∞
sin x
Hence evaluate ∫ x
dx .
0
1 for x < 1
1st Part: Find Fourier transform of f ( x ) =
0 for x > 1
sin s
Thus F(s) = 2 , s ≠ 0.
s
This is the required Fourier transform.
For s = 0, we have F(0) = 2.
∞
sin x
∫
nd
2 Part: Evaluate dx .
0
x
1
∞ ∞
1 1 2sin s − isx for x < 1
∫ ⇒ ∫
− isx
f (x) = F ( s ) e ds e ds =
2π −∞ 2π −∞ s 0 for x > 1
∞
sin s − isx π
for x < 1
⇒
−∞
∫ s
e ds =
0 for x > 1
Putting x = 0, we get
Fourier Transforms: 6
Visit: https://www.sites.google.com/site/hub2education/
∞ ∞
sin s sin s π
∫
−∞
s
ds = π ⇒ ∫
0
s
ds = .
2
[since the integrand is even].
∞
sin x π
Replacing s by x, we get ∫
0
x
dx = .
2
∫ f ( x ) .e
isx
F (s ) = dx
−∞
−a a ∞
= ∫ f ( x ) .eisx dx + ∫ f ( x ) .eisx dx + ∫ f ( x ) .eisx dx
−∞ −a a
−a a
1 isx
∞
= ∫ 0dx + ∫ e dx + ∫ 0dx
−∞ −a
2a a
a
1 eisx 1
= = eisa − e− isa
2a ( is ) x =− a 2asi
π
eix − e− ix isx ( eix − e− ix )
=∫ e dx ∵ sin x =
0 2i 2i
π π
1 e( ) e ( ) 1 e( ) e ( )
π i 1+ s x − i 1−s x i 1+ s x − i 1−s x
1
= ∫ e ( ) − e ( ) dx =
i 1+ s x − i 1−s x
− = +
2i 0 2i i (1 + s ) −i (1 − s ) 0 2i i (1 + s ) i (1 − s ) 0
1 e( ) e ( ) 1
i 1+ s π − i 1−s π
1
= + − −
2i i (1 + s ) i (1 − s ) i (1 + s ) i (1 − s )
1 2eisπ 2
=− + since e ±iπ = −1 .
2i i (1 − s 2 ) i (1 − s 2 )
eiπs + 1
⇒ F (s ) =
1 − s2
is the required exponential Fourier transform representation.
1
∞
1
∞
eiπs + 1 − isx
∫ F ( s ) e ds = 2π −∞∫ 1 − s2 e ds .
− isx
Now f ( x ) =
2π −∞
∞
1 cos πs + i sin πs + 1
⇒ f (x) = ∫
2π −∞ 1 − s2
( cos sx − i sin sx ) ds
∞ ∞
1 cos sx + cos s ( x − π ) i sin sx + sin s ( x − π )
⇒ f (x) = ∫
2π ∞ (1 − s )
2
ds − ∫
2π −∞ (1 − s2 )
ds
∞
1 cos sx + cos s ( x − π )
π ∫0
⇒ f (x) = ds .
(1 − s2 )
Home Assignments