Theoretical Gaussian field predictions assume σ=1

[javicarron]

I think the code for computing the theoretical predictions of Gaussian fields (and potentially other fields) is implicitly assuming $\sigma=1$, even for fields where $\sigma$ is explicitly set to a different value.

The correction should be easy, but it requires some careful modifications.