Fourier Transform Properties
Last section we calculated some fourier transform. Now we can have some useful properties.
Let's put the formula here,
F(ω)=∫−∞+∞f(t)e−jωtdt
And inversely,
f(t)=2π1∫−∞+∞F(ω)ejωtdω
Linear
Because, integrate is linear, thus fourier transform is linear.
That is to say,
af(t)+bg(t)FaF(ω)+bG(ω)
Duality
If we assume,
f(t)FF(ω)
Let's consider the fourier transform of F(t).
F∗(ω)=∫−∞+∞F(t)e−jωtdt
We also know that,
f(t)=2π1∫−∞+∞F(ω)ejωtdω
We can switch ω and t,
f(ω)=2π1∫−∞+∞F(t)e−jωtdt
Then we replace ω with −ω.
f(−ω)=2π1∫−∞+∞F(t)ejωtdt
Now look at,
F∗(ω)=∫−∞+∞F(t)e−jωtdt
We can conclude that,
F∗(ω)=2πf(−ω)
In all, if,
f(t)FF(ω)
Then,
F(t)F2πf(−ω)
We've seen a real case before, that is,
δ(t)F1
And
1F2πδ(ω)
Shifting Time is Shifting Phase
Suppose,
f(t)FF(ω)
Then,
∫−∞+∞f(t−t0)e−jωtdt=e−jωt0∫−∞+∞f(t−t0)e−jω(t−t0)d(t−t0)=e−jωt0F(ω)
Or,
f(t−t0)Fe−jωt0F(ω)
Shifting Phase is Shifting Frequency
Suppose,
f(t)FF(ω)
∫−∞+∞f(t)e−jω0te−jωtdt=F(ω+ω0)
Or,
f(t)e−jω0tFF(ω+ω0)
Scale
Suppose,
f(t)FF(ω)
∫−∞+∞f(αt)e−jωtdt=∣∣α∣∣1∫−∞+∞f(αt)e−jαωα∣tdαt=∣∣α∣∣1F(αω)
Or,
f(αt)F∣∣α∣∣1F(αω)
Derivative
Suppose,
f(t)FF(ω)
Then,
∫−∞+∞f′(t)e−jωtdt=−jω∫−∞+∞f(t)e−jωtdt=−jωF(ω)
Or,
f′(t)F−jωF(ω)
Integral
Suppose,
f(t)FF(ω)
Because,
∫−∞+∞(∫−∞tf(τ)dτ)e−jωtdt=∫−∞+∞(∫−∞te−jωτdτ)f(t)dt=−jωF(ω)
Or,
∫−∞tf(τ)dτF−jωF(ω)
Convolution
Suppose,
f(t)FF(ω)g(t)FG(ω)
Then,
f(t)∗g(t)=∫−∞+∞f(τ)g(t−τ)dτ
∫−∞+∞(f(t)∗g(t))e−jωtdt=∫−∞+∞(∫−∞+∞f(τ)g(t−τ)dτ)e−jωtdt=∫−∞+∞(∫−∞+∞f(τ)e−jωτdτ)g(t−τ)e−jω(t−τ)dt=∫−∞+∞F(ω)g(t−τ)e−jω(t−τ)dt=F(ω)G(ω)
So,
f(t)∗g(t)FF(ω)G(ω)
Multiplication
Similarly, we can prove,
f(t)g(t)F2π1F(ω)∗G(ω)
With fourier inverse transform.