Why does the "get_phonons_in_qpath" plot strange phonon dispersion?

[biglinn]

Hi, @mesonepigreco:
I find a question about plotting the phonon dispersion.
When I plot the phonon dispersion with plot_dispersion.py(main codes are as follows), it shows strange imaginary frequency:

dyn = CC.Phonons.Phonons("dyn_mat/sscha_dyn_population10_", 6)

qpath, data = CC.Methods.get_bandpath(dyn.structure.unit_cell,
                                      PATH,
                                      SPECIAL_POINTS,
                                      N_POINTS)
xaxis, xticks, xlabels = data # Info to plot correclty the x axis

dispersion = CC.ForceTensor.get_phonons_in_qpath(dyn, qpath)

However, I get dyn0(the file containing all irredicuble q-points, q2r.x needs it) back, then use q2r.x and matdyn.x to plot the dispersion(asr is simple).
In this case, the dispersion is as follows:

I wonder if the get_phonons_in_qpath has any limitations.

[mesonepigreco]

This seems suspicious.
In particular, you have an imaginary phonon mode at Gamma, but this point should not be interpolated. The ZU part of your dispersion is substantially different.

Can you upload the dynamical matrix so that I can try to reproduce the error?
Thanks a lot,
LM

[biglinn]

Thank you for your help!
dyn.zip
These are the dynamical matrix(dyn1~8). The "dyn0" is get from ph.x, containing all irredicuble q-points.