Condiciones para que exista la transformada de Fourier cuando:
Es absolutamente integrable
∫
−
∞
∞
|
f
(
t
)
|
d
t
<
∞
(Converge)
{\displaystyle \int \limits _{-\infty }^{\infty }{\left|f\left(t\right)\right|dt<\infty {\text{ (Converge)}}}}
f(t) continua por intervalos [a,b] finito
lim
t
→
t
0
±
f
(
t
)
finito
,
∀
t
0
{\displaystyle {\underset {t\to t_{0}^{\pm }}{\mathop {\lim } }}\,{\text{ }}f(t){\text{ finito}},\forall t_{0}}
lim
f
(
t
)
=
f
(
t
+
)
+
f
(
t
−
)
2
{\displaystyle {\underset {}{\mathop {\lim } }}\,f(t)={\frac {f(t^{+})+f(t^{-})}{2}}}
Para una f(t) real o compleja, con variable real t, se define la transformada de Fourier como:
F
[
f
(
t
)
]
=
F
(
ω
)
=
∫
−
∞
∞
f
(
t
)
.
e
−
j
ω
t
∂
t
{\displaystyle \mathbb {F} [f(t)]=F(\omega )=\int \limits _{-\infty }^{\infty }{f(t).e^{-j\omega t}\partial t}}
Igualmente, tenemos la función inversa de Fourier:
F
−
1
[
F
(
ω
)
]
=
f
(
t
)
=
1
2
π
∫
−
∞
∞
F
(
ω
)
.
e
+
j
ω
t
∂
ω
{\displaystyle \mathbb {F} ^{-1}[F(\omega )]=f(t)={\frac {1}{2\pi }}\int \limits _{-\infty }^{\infty }{F(\omega ).e^{+j\omega t}\partial \omega }}
De forma que se cumple
F
−
1
[
F
[
f
(
t
)
]
⏟
F
(
ω
)
]
=
f
(
t
)
F
[
F
−
1
[
F
(
ω
)
]
⏟
f
(
t
)
]
=
F
(
ω
)
{\displaystyle {\begin{aligned}&\mathbb {F} ^{-1}[\underbrace {\mathbb {F} [f(t)]} _{F(\omega )}]=f(t)\\&\mathbb {F} [\underbrace {\mathbb {F} ^{-1}[F(\omega )]} _{f(t)}]=F(\omega )\\\end{aligned}}}
Es costumbre representar la transformada de fourier de una señal con la letra que lo representa en mayúsculas:
f
(
t
)
↔
F
(
ω
)
,
g
(
t
)
↔
G
(
ω
)
,
x
(
t
)
↔
X
(
ω
)
{\displaystyle f(t)\leftrightarrow F(\omega ),g(t)\leftrightarrow G(\omega ),x(t)\leftrightarrow X(\omega )}
Alguna propiedad:
Integración:
F
(
ω
=
0
)
=
∫
−
∞
∞
f
(
t
)
∂
t
f
(
t
=
0
)
=
1
2
π
∫
−
∞
∞
F
(
ω
)
∂
ω
{\displaystyle {\begin{aligned}&F(\omega =0)=\int _{-\infty }^{\infty }{f(t)\partial t}\\&f(t=0)={\frac {1}{2\pi }}\int _{-\infty }^{\infty }{F(\omega )\partial \omega }\\\end{aligned}}}
La demostración es sencilla:
F
[
f
(
t
)
]
=
F
(
ω
)
=
∫
−
∞
∞
f
(
t
)
.
e
−
j
ω
t
∂
t
F
(
ω
=
0
)
=
∫
−
∞
∞
f
(
t
)
.
e
−
j
ω
t
∂
t
|
ω
=
0
=
∫
−
∞
∞
f
(
t
)
.
e
−
j
0
t
∂
t
=
∫
−
∞
∞
f
(
t
)
∂
t
{\displaystyle {\begin{aligned}&\mathbb {F} [f(t)]=F(\omega )=\int \limits _{-\infty }^{\infty }{f(t).e^{-j\omega t}\partial t}\\&F(\omega =0)=\left.\int \limits _{-\infty }^{\infty }{f(t).e^{-j\omega t}\partial t}\right|_{\omega =0}=\int \limits _{-\infty }^{\infty }{f(t).e^{-j0t}\partial t}=\int \limits _{-\infty }^{\infty }{f(t)\partial t}\\\end{aligned}}}
Análogamente:
F
−
1
[
F
(
ω
)
]
=
f
(
t
)
=
1
2
π
∫
−
∞
∞
F
(
ω
)
.
e
+
j
ω
t
∂
ω
f
(
t
=
0
)
=
1
2
π
∫
−
∞
∞
F
(
ω
)
.
e
+
j
ω
t
=
0
∂
ω
=
1
2
π
∫
−
∞
∞
F
(
ω
)
∂
ω
{\displaystyle {\begin{aligned}&\mathbb {F} ^{-1}[F(\omega )]=f(t)={\frac {1}{2\pi }}\int \limits _{-\infty }^{\infty }{F(\omega ).e^{+j\omega t}\partial \omega }\\&f(t=0)={\frac {1}{2\pi }}\int \limits _{-\infty }^{\infty }{F(\omega ).e^{+j\omega t=0}\partial \omega }={\frac {1}{2\pi }}\int \limits _{-\infty }^{\infty }{F(\omega )\partial \omega }\\\end{aligned}}}
Comentar que en matemáticas es usual dejar el resultado de la transformada de Fourier en función de ω, mientras que en ingeniería es más habitual dejarlo en f, debido a que se mantiene la simetría.
Ambas están relacionadas directamente.
ω
=
2
π
f
F
[
f
(
t
)
]
=
F
(
ω
)
=
∫
−
∞
∞
f
(
t
)
.
e
−
j
ω
t
∂
t
→
F
[
f
(
t
)
]
=
F
(
f
)
=
∫
−
∞
∞
f
(
t
)
.
e
−
j
2
π
f
t
∂
t
F
−
1
[
F
(
ω
)
]
=
f
(
t
)
=
1
2
π
∫
−
∞
∞
F
(
ω
)
.
e
+
j
ω
t
∂
ω
=
1
2
π
∫
−
∞
∞
F
(
f
)
.
e
+
j
2
π
f
t
2
π
∂
f
→
F
−
1
[
F
(
f
)
]
=
f
(
t
)
=
∫
−
∞
∞
F
(
f
)
.
e
+
j
2
π
f
t
∂
f
{\displaystyle {\begin{aligned}&\omega =2\pi f\\&\mathbb {F} [f(t)]=F(\omega )=\int \limits _{-\infty }^{\infty }{f(t).e^{-j\omega t}\partial t}\to \mathbb {F} [f(t)]=F(f)=\int \limits _{-\infty }^{\infty }{f(t).e^{-j2\pi ft}\partial t}{\text{ }}\\&\mathbb {F} ^{-1}[F(\omega )]=f(t)={\frac {1}{2\pi }}\int \limits _{-\infty }^{\infty }{F(\omega ).e^{+j\omega t}\partial \omega }={\frac {1}{2\pi }}\int \limits _{-\infty }^{\infty }{F(f).e^{+j2\pi ft}2\pi \partial f\to }\\&\mathbb {F} ^{-1}[F(f)]=f(t)=\int \limits _{-\infty }^{\infty }{F(f).e^{+j2\pi ft}\partial f}\\\end{aligned}}}
Con la ventaja que en función de f, no tenemos ese divisor de 2π, por lo que mantiene la simetría, que es más comodo al realizar cálculos (sin embargo, no olvidar que la fase de la exponencial esta invertida).
F
[
f
(
t
)
]
=
F
(
ω
)
=
∫
−
∞
∞
f
(
t
)
.
e
−
j
ω
t
∂
t
F
−
1
[
F
(
ω
)
]
=
f
(
t
)
=
1
2
π
∫
−
∞
∞
F
(
ω
)
.
e
+
j
ω
t
∂
ω
{\displaystyle {\begin{aligned}&\mathbb {F} [f(t)]=F(\omega )=\int \limits _{-\infty }^{\infty }{f(t).e^{-j\omega t}\partial t}\\&\mathbb {F} ^{-1}[F(\omega )]=f(t)={\frac {1}{2\pi }}\int \limits _{-\infty }^{\infty }{F(\omega ).e^{+j\omega t}\partial \omega }\\\end{aligned}}}
Pulso rectangular
∏
(
t
T
)
=
{
1
,
|
t
|
≤
T
╱
2
0
,
|
t
|
>
T
╱
2
{\displaystyle \prod {\left({\frac {t}{T}}\right)}=\left\{{\begin{aligned}&1,\left|t\right|\leq {}^{T}\!\!\diagup \!\!{}_{2}\;\\&0,\left|t\right|>{}^{T}\!\!\diagup \!\!{}_{2}\;\\\end{aligned}}\right.}
F
[
∏
(
t
T
)
]
=
2
sin
(
ω
T
╱
2
)
ω
=
T
sinc
(
T
ω
2
)
{\displaystyle \mathbb {F} \left[\prod {\left({\frac {t}{T}}\right)}\right]=2{\frac {\sin \left(\omega {}^{T}\!\!\diagup \!\!{}_{2}\;\right)}{\omega }}=T{\text{sinc}}\left(T{\frac {\omega }{2}}\right)}
Pulso triangular
Λ
(
t
T
)
=
{
1
−
|
t
|
T
,
|
t
|
≤
T
0
,
|
t
|
>
T
{\displaystyle \Lambda \left({\frac {t}{T}}\right)=\left\{{\begin{aligned}&1-{\frac {\left|t\right|}{T}},{\text{ }}\left|t\right|\leq T\\&0{\text{ }},{\text{ }}\left|t\right|>T\\\end{aligned}}\right.}
F
[
Λ
(
t
T
)
]
=
2
(
1
−
cos
(
ω
T
)
)
ω
2
T
=
T
⋅
sinc
2
(
T
ω
2
π
)
{\displaystyle \mathbb {F} \left[\Lambda \left({\frac {t}{T}}\right)\right]={\frac {2(1-\cos(\omega T))}{\omega ^{2}T}}=T\cdot {\text{sinc}}^{2}\left(T{\frac {\omega }{2\pi }}\right)}
sign
(
t
)
=
{
1
,
t
>
0
−
1
,
t
<
0
{\displaystyle \operatorname {sign} (t)=\left\{{\begin{aligned}&1,{\text{ }}t>0\\&-1,{\text{ }}t<0\\\end{aligned}}\right.}
F
[
sign
(
t
)
]
=
2
j
ω
→
ω
=
2
π
f
1
j
π
f
{\displaystyle \mathbb {F} \left[\operatorname {sign} (t)\right]={\frac {2}{j\omega }}{\xrightarrow[{\omega =2\pi f}]{}}{\frac {1}{j\pi f}}}
u
(
t
)
=
{
1
,
t
>
0
0
,
t
<
0
{\displaystyle u(t)=\left\{{\begin{aligned}&1,t>0\\&0,t<0\\\end{aligned}}\right.}
F
[
u
(
t
)
]
=
1
j
ω
+
π
δ
(
ω
)
⏟
a veces se omite
{\displaystyle \mathbb {F} [u(t)]={\frac {1}{j\omega }}+\underbrace {\pi \delta (\omega )} _{\text{a veces se omite}}}
Delta de Dirac
δ
(
t
)
{\displaystyle \delta (t)}
F
[
δ
(
t
)
]
=
1
↔
F
[
1
]
=
2
π
δ
(
−
ω
)
{\displaystyle \mathbb {F} [\delta (t)]=1\leftrightarrow F[1]=2\pi \delta (-\omega )}
cos
(
ω
0
t
)
{\displaystyle \cos(\omega _{0}t)}
F
[
cos
(
ω
0
t
)
]
=
π
δ
(
ω
−
ω
0
)
+
π
δ
(
ω
+
ω
0
)
{\displaystyle \mathbb {F} \left[\cos(\omega _{0}t)\right]=\pi \delta (\omega -\omega _{0})+\pi \delta (\omega +\omega _{0})}
sin
(
ω
0
t
)
{\displaystyle \sin(\omega _{0}t)}
F
[
sin
(
ω
0
t
)
]
=
π
j
δ
(
ω
−
ω
0
)
−
π
j
δ
(
ω
+
ω
0
)
{\displaystyle \mathbb {F} \left[\sin(\omega _{0}t)\right]={\frac {\pi }{j}}\delta (\omega -\omega _{0})-{\frac {\pi }{j}}\delta (\omega +\omega _{0})}
∑
k
=
−
∞
∞
δ
(
t
−
k
T
s
)
{\displaystyle \sum \limits _{k=-\infty }^{\infty }{\delta \left(t-kT_{s}\right)}}
F
[
∑
k
=
−
∞
∞
δ
(
t
−
k
T
s
)
]
=
F
[
1
T
s
∑
k
=
−
∞
∞
e
j
2
π
T
s
k
t
]
=
1
T
s
⋅
∑
k
=
−
∞
∞
2
π
δ
(
ω
−
k
2
π
T
s
)
{\displaystyle \mathbb {F} \left[\sum \limits _{k=-\infty }^{\infty }{\delta \left(t-kT_{s}\right)}\right]=\mathbb {F} \left[{\frac {1}{T_{s}}}\sum \limits _{k=-\infty }^{\infty }{e^{j{\frac {2\pi }{T_{s}}}kt}}\right]={\frac {1}{T_{s}}}\cdot \sum \limits _{k=-\infty }^{\infty }{2\pi \delta }\left(\omega -k{\frac {2\pi }{T_{s}}}\right)}
¿Cuál es la transformada de fourier de una señal periódica? Para saberlo usaremos el tren de deltas utilizado anteriormente, como función auxiliar.
x
(
t
)
=
x
(
t
+
T
s
)
x
0
(
t
)
:
función fundamental
x
(
t
)
=
∑
k
=
−
∞
∞
x
0
(
t
−
k
T
s
)
=
x
0
(
t
)
∗
∑
k
=
−
∞
∞
δ
(
t
−
k
T
s
)
∑
k
=
−
∞
∞
δ
(
t
−
k
T
s
)
=
1
T
s
∑
k
=
−
∞
∞
e
j
2
π
T
s
k
t
→
C
k
=
1
T
s
∀
k
(Propiedades de Trans
. Fourier)
Recordemos las propiedades:
{
F
[
e
+
j
ω
0
t
f
(
t
)
]
=
F
(
ω
−
ω
0
)
F
[
1
]
=
2
π
δ
(
ω
)
→
F
[
e
+
j
ω
0
t
]
=
2
π
δ
(
ω
−
ω
0
)
F
[
f
(
t
)
∗
g
(
t
)
]
=
F
(
ω
)
⋅
G
(
ω
)
}
F
[
x
(
t
)
]
=
F
[
x
0
(
t
)
∗
∑
k
=
−
∞
∞
δ
(
t
−
k
T
s
)
]
=
F
[
x
0
(
t
)
∗
1
T
s
∑
k
=
−
∞
∞
e
j
2
π
T
s
k
t
]
X
(
ω
)
=
X
0
(
ω
)
.
1
T
s
∑
k
=
−
∞
∞
2
π
δ
(
ω
−
k
2
π
T
s
)
=
2
π
T
s
∑
k
=
−
∞
∞
X
0
(
k
2
π
T
s
)
⋅
δ
(
ω
−
k
2
π
T
s
)
{\displaystyle {\begin{aligned}&x(t)=x(t+T_{s})\\&x_{0}(t):{\text{ función fundamental}}\\&x(t)=\sum \limits _{k=-\infty }^{\infty }{x_{0}\left(t-kT_{s}\right)}=x_{0}(t)*\sum \limits _{k=-\infty }^{\infty }{\delta \left(t-kT_{s}\right)}\\&\sum \limits _{k=-\infty }^{\infty }{\delta \left(t-kT_{s}\right)}={\frac {1}{T_{s}}}\sum \limits _{k=-\infty }^{\infty }{e^{j{\frac {2\pi }{T_{s}}}kt}}\to C_{k}={\frac {1}{T_{s}}}{\text{ }}\forall k{\text{ (Propiedades de Trans}}{\text{. Fourier)}}\\&{\text{Recordemos las propiedades: }}\\&\left\{{\begin{aligned}&\mathbb {F} [e^{+j\omega _{0}t}f(t)]=F(\omega -\omega _{0})\\&\mathbb {F} [1]=2\pi \delta (\omega )\to \mathbb {F} [e^{+j\omega _{0}t}]=2\pi \delta (\omega -\omega _{0})\\&\mathbb {F} \left[f(t)*g(t)\right]=F(\omega )\cdot G(\omega )\\\end{aligned}}\right\}\\&\mathbb {F} \left[x(t)\right]=\mathbb {F} \left[x_{0}(t)*\sum \limits _{k=-\infty }^{\infty }{\delta \left(t-kT_{s}\right)}\right]=\mathbb {F} \left[x_{0}(t)*{\frac {1}{T_{s}}}\sum \limits _{k=-\infty }^{\infty }{e^{j{\frac {2\pi }{T_{s}}}kt}}\right]\\&X(\omega )=X_{0}(\omega ).{\frac {1}{T_{s}}}\sum \limits _{k=-\infty }^{\infty }{2\pi \delta \left(\omega -k{\frac {2\pi }{T_{s}}}\right)}={\frac {2\pi }{T_{s}}}\sum \limits _{k=-\infty }^{\infty }{X_{0}\left({\frac {k2\pi }{T_{s}}}\right)\cdot \delta \left(\omega -k{\frac {2\pi }{T_{s}}}\right)}\\\end{aligned}}}
X
(
ω
)
=
2
π
T
s
∑
k
=
−
∞
∞
X
0
(
k
2
π
T
s
)
⋅
δ
(
ω
−
k
2
π
T
s
)
{\displaystyle X(\omega )={\frac {2\pi }{T_{s}}}\sum \limits _{k=-\infty }^{\infty }{X_{0}\left({\frac {k2\pi }{T_{s}}}\right)\cdot \delta \left(\omega -k{\frac {2\pi }{T_{s}}}\right)}}