I am convolving, or trying to, a Gaussian and a Lorentzian as follows:

Gaussian1[ \[Mu]_, \[Sigma]_, x_] := 
1/(\[Sigma] Sqrt[2 \[Pi]]) Exp[-(1/2) ((x - \[Mu]) / \[Sigma])^2]

Lorentzian[\[Gamma]_, x0_, x_] := 
1/\[Pi] (1/2 \[Gamma])/((x - x0)^2 + (1/2 \[Gamma])^2)

The convolution attempt being

Convolve[Lorentzian[1, \[Gamma], x0, x], Gaussian1[1, \[Mu], \[Sigma], x], x, y]

However this simply returns itself except output formatted:

 (\[Gamma] Convolve[1/((x - x0)^2 + \[Gamma]^2/4), 
      E^(-((x - \[Mu])^2/(2 \[Sigma]^2))), x, y])/(2 Sqrt[2] \[Pi]^(
     3/2) \[Sigma])

If I add assumptions,

Assumptions->Re[\[Gamma]] && Re[\[Sigma]]

The result is the same. I also tried to perform the convolution using Integrate[] as:

            1/\[Pi] (1/2 \[Gamma])/((\[Tau])^2 + (1/2 \[Gamma])^2) * 1/(\[Sigma] Sqrt[2 \[Pi]]) Exp[-(1/2) ((t - \[Tau]) / \[Sigma])^2],
            {\[Tau], -Infinity, +Infinity},
            Assumptions->Re[\[Gamma]] && Re[\[Sigma]] 

No joy! This convolution is well known as a Voigt profile So it has a solution but I don't understand what I am doing wrong.


You could try recasting your formulae to the internal PDF distributions NormalDistribution and CauchyDistribution.

Convolve[PDF[NormalDistribution[0, σ], x], 
 PDF[CauchyDistribution[a, b], x], x, y]
(* -(1/(2 Sqrt[2] π^(3/2) σ))
 I E^((-a^2 + (b - I y)^2 + 2 a (-I b + y))/(
  2 σ^2)) (E^((
     2 I a b)/σ^2) π Erfi[((a - I b - y) Sqrt[
       1/σ^2])/Sqrt[2]] - 
    E^((2 I b y)/σ^2) π Erfi[((a + I b - y) Sqrt[
       1/σ^2])/Sqrt[2]] - 
    E^((2 I a b)/σ^2) Log[a - I b - y] + 
    E^((2 I b y)/σ^2) Log[a + I b - y] - 
    E^((2 I b y)/σ^2) Log[-a - I b + y] + 
    E^((2 I a b)/σ^2) Log[-a + I b + y]) *)
| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.