Fourier Transform Theorems • Addition Theorem • Shift Theorem • Convolution Theorem • Similarity Theorem • Rayleigh’s Theorem • Differentiation Theorem Addition Theorem F { f + g} = F + G Proof: F { f + g}(s) = = Z ∞ −∞ f (t)e Z ∞ −∞ − j2πst = F(s) + G(s) [ f (t) + g(t)] e− j2πst dt dt + Z ∞ −∞ g(t)e− j2πst dt Shift Theorem Proof: F { f (t − t0)}(s) = e− j2πst0 F(s) F { f (t − t0)}(s) = Z ∞ −∞ f (t − t0)e− j2πst dt Multiplying the r.h.s. by e j2πst0 e− j2πst0 = 1 yields: = Z ∞ −∞ = e F { f (t − t0)}(s) f (t − t0)e− j2πst e j2πst0 e− j2πst0 dt − j2πst0 Z ∞ −∞ f (t − t0)e− j2πs(t−t0)dt. Substituting u = t − t0 and du = dt yields: F { f (t − t0)}(s) = e − j2πst0 Z ∞ −∞ f (u)e− j2πsudu = e− j2πst0 F(s). Shift Theorem Example F {sin (2π (t + 1/4))} (s) =e = = = = = = j2πs 4 F {sin(2πt)} j e · [δ(s + 1) − δ(s − 1)] 2 i h jπs j jπs e 2 δ(s + 1) − e 2 δ(s − 1) 2h i jπ(+1) j jπ(−1) e 2 δ(s + 1) − e 2 δ(s − 1) 2 j [− jδ(s + 1) − jδ(s − 1)] 2 1 [δ(s + 1) + δ(s − 1)] 2 F {cos(2πs)} jπs 2 Shift Theorem (variation) −1 F Proof: F −1 {F(s − s0)}(t) = e j2πs0 t f (t) {F(s − s0)}(t) = Z ∞ −∞ F(s − s0)e j2πst ds Multiplying the r.h.s. by e j2πs0 t e− j2πs0 t = 1 yields: −1 F = Z ∞ −∞ =e {F(s − s0)}(t) F(s − s0)e j2πst e j2πs0 t e− j2πs0 t ds j2πs0 t Z ∞ −∞ F(s − s0)e j2π(s−s0)t ds. Substituting u = s − s0 and du = ds yields: F −1 {F(s − s0)}(t) = e j2πs0 t Z ∞ −∞ F(u)e j2πut du = e j2πs0 t f (t). Convolution Theorem F { f ∗ g} = F · G Proof: = Z ∞ Z −∞ ∞ −∞ F { f ∗ g}(s) f (u)g(t − u)du e− j2πst dt Changing the order of integration: F { f ∗ g}(s) Z ∞ Z ∞ g(t − u)e− j2πst dt du f (u) = −∞ −∞ By the Shift Theorem, we recognize that Z ∞ g(t − u)e− j2πst dt = e− j2πsuG(s) so that −∞ F { f ∗ g}(s) = Z ∞ f (u)e− j2πsuG(s)du −∞ = G(s) Z ∞ −∞ f (u)e− j2πsudu = G(s) · F(s) Convolution Theorem Example The pulse, Π, is defined as: 1 if |t| ≤ 12 Π(t) = 0 otherwise. The triangular pulse, Λ, is defined as: 1 − |t| if |t| ≤ 1 Λ(t) = 0 otherwise. It is straightforward to show that Λ = Π ∗ Π. Using this fact, we can compute F {Λ}: F {Λ}(s) = F {Π ∗ Π}(s) = F {Π}(s) · F {Π}(s) sin(πs) sin(πs) = · πs πs sin2(πs) = . 2 2 πs Convolution Theorem (variation) F −1 Proof: F = Z ∞ Z −∞ ∞ −∞ {F ∗ G} = f · g −1 {F ∗ G}(t) F(u)G(s − u)du e j2πst ds Changing the order of integration: F −1{F ∗ G}(t) Z ∞ Z ∞ G(s − u)e j2πst ds du F(u) = −∞ −∞ By the Shift Theorem, we recognize that Z ∞ G(s − u)e j2πst ds = e j2πtug(t) so that F −1 −∞ {F ∗ G}(t) = Z ∞ F(u)e j2πtug(t)du −∞ Z ∞ = g(t) −∞ F(u)e j2πtudu = g(t) · f (t) Similarity Theorem Proof: 1 s F { f (at)} (s) = F |a| a F { f (at)}(s) = There are two cases. Z ∞ f (at)e− j2πst dt −∞ • a > 0. Multiplying the integral by a/|a| = 1 and the exponent by a/a = 1 yields: Z 1 ∞ f (at)e− j2π(s/a)at adt F { f (at)}(s) = |a| −∞ We now make the substitution u = at and du = adt: Z 1 ∞ F { f (at)}(s) = f (u)e− j2π(s/a)udu |a| −∞ 1 s = F |a| a Similarity Theorem (contd.) • a < 0. Multiplying the integral by |a|/|a| = 1 and the exponent by −|a|/a = 1 and using the fact that a = −|a| yields: F { f (at)}(s) = 1 t=∞ f (−|a|t)e− j2π(s/a)(−|a|t)|a|dt |a| t=−∞ We now make the substitution u = −|a|t and du = −|a|dt: Z 1 u=−∞ f (u)e− j2π(s/a)udu F { f (at)}(s) = − |a| u=∞ Z 1 ∞ f (u)e− j2π(s/a)udu = |a| −∞ 1 s = F |a| a Z Similarity Theorem Example Let’s compute, G(s), the Fourier transform of: −t 2/9 g(t) = e . We know that the Fourier transform of a Gaussian: −π t 2 f (t) = e is a Gaussian: −π s2 . F(s) = e We also know that : 1 s F { f (at)}(s) = F . |a| a We need to write g(t) in the form f (at): g(t) = f (at) = e Let a = 3√1 π : g(t) = e −t 2/9 =e 1 t −π 3√ π 2 −π(at)2 =f . It follows that: √ −π(3√πs)2 . G(s) = 3 π e 1 √ t . 3 π Rayleigh’s Theorem Z ∞ −∞ Proof: Z | f (t)|2dt = ∞ −∞ 2 | f (t)| dt = Z ∞ −∞ Z ∞ −∞ |F(s)|2ds f ∗(t) f (t)dt Substituting {F −1{F}}∗(t) for f ∗(t): ∗ Z ∞ Z ∞ F(s)e j2πst ds f (t) dt = −∞ −∞ Z ∞ Z ∞ = F ∗(s)e− j2πst ds f (t) dt −∞ −∞ Rayleigh’s Theorem (contd.) Changing the order of integration: Z t=∞ Z s=∞ = F ∗(s)e− j2πst ds f (t) dt t=−∞ s=−∞ Z t=∞ Z s=∞ e− j2πst f (t) dt ds = F ∗(s) = = Zs=−∞ ∞ t=−∞ F ∗(s)F(s) ds Z−∞ ∞ −∞ |F(s)|2 ds Differentiation Theorem F { f 0}(s) = j2πsF(s) Proof: f =F Therefore f (t) = Z ∞ −1 {F} F(s)e j2πst ds −∞ Differentiating both sides with respect to t: f 0(t) = or Z ∞ j2πsF(s)e j2πst ds −∞ f 0 = F −1{ j2πsF(s)} Taking the Fourier transform of both sides: F { f 0}(s) = F {F −1 { j2πsF(s)}} = j2πsF(s) Differentiation Theorem Example F d sin(2πt) (s) dt = j2πsF {sin(2πt)} (s) j = j2πs · [δ(s + 1) − δ(s − 1)] 2 = −πs [δ(s + 1) − δ(s − 1)] = π [sδ(s − 1) − sδ(s + 1)] = π [(+1)δ(s − 1) − (−1)δ(s + 1)] = π [δ(s − 1) + δ(s + 1)] = πF {2 cos(2πt)}(s) = F {2π cos(2πt)}(s)