Update laser frequency noise noise-shape

Most OMS related noises include the usual low-frequency relaxation. The laser noise, however, is still a white noise accross the whole frequency band, while it's performance model allocation is also relaxed towards low frequencies. This might lead to errors in TDI related studies, which will underestimate residual laser noise at low frequencies.

I would suggest to update laser noise to match the allocation, i.e., to generate a noise with an ASD of 30 Hz/sqrtHz * sqrt(1 + (f / 2e-3)^4).

This might also benefit TDI related studies, since it can lift the residual laser noise terms at low frequencies above the numerical noise floor of the simulation, and will allow to test models for a larger part of the frequency band.

Something related: some noises are more current best estimates (like clock + modulation noise), while others are based on allocations (like OMS noises). We should probably pick one or the other, or have models for both and add a switch to change the whole simulation.

Possibly related to #33.

@j2b.bayle, what do you think?