LAMMPS has 'fix phonon', which can be used to estimate the phonon frequencies from a dynamic simulation. I've never used it but I think it might be useful to you: http://lammps.sandia.gov/doc/fix_phonon.html
There's also a third-party add-on for LAMMPS that can calculate hessians (and thus vibrational frequencies) on-the-fly by jusing the finite difference method. You can find it here: https://bitbucket.org/numericalsolutions/lammps-hessian Again, I haven't tested it, but you might want to take a look at it.
Even if both methods are not what you're looking for, maybe they're useful inspiration when building your own solution.
LAMMPS has 'fix phonon', which can be used to estimate the phonon frequencies from a dynamic simulation. I've never used it but I think it might be useful to you: http://lammps.sandia.gov/doc/fix_phonon.html
There's also a third-party add-on for LAMMPS that can calculate hessians (and thus vibrational frequencies) on-the-fly by jusing the finite difference method. You can find it here: https://bitbucket.org/numericalsolutions/lammps-hessian Again, I haven't tested it, but you might want to take a look at it.
Even if both methods are not what you're looking for, maybe they're useful inspiration when building your own solution.
you can get velocity autocorrelation by 'fix vacf' and then you can take FFT by MATLAB code.I do it but I have some problem with graphene ripples which is made due to thermal fluctuation!!