From: The Cromicus Productions (
Date: Tue Dec 27 2016 - 03:12:38 CST

Hi everyone.

I'm interested in calculating the power spectrum from the velocity
auto-correlation of a whole molecule (not the dispersion relation). To do
this, I'm following example 2 of
where I'm working with the velocities as trajectories. When I work with the
IR-spec from the gui I get some peaks. Then, I do the procedure from the
example and get a data file which, if I understand correctly, has
frequency - amplitude of atom 1 - amplitude of atom 2 - ... amplitude of
atom n - average amplitude
for each frequency. I plot the frequency against what I believe to be the
average amplitude and I'm getting a gaussian distribution!
Am I correct in assuming that the final column is the average amplitude?
I have different atoms, should I consider a weighted average or is the
usual average ok?

Thanks a lot!