Sb knows a method to get angle information of my (peaks,dyads etc.) to determine accuracy of direction information/voxel?

Hey all,
I want to estimate the accuracy of the direction information (peaks,dyads/voxel) for different measurements (b-values) in a given roi. Therefore I need a kind of reference value so that I can compare the direction information within a voxel for different measurements. I was thinking maybe it is the easiest to calculate the spherical angles of each peak/voxel and compare them thus I can say something about my accuracy and precision.
So it would be nice if somebody has another good approach or idea how to compare measurements (different b-values) regarding to the direction information in a given roi or can tell me how to get the angles as values of a given peak. I want to use MRtrix3 CSD (peaks) and FSL BedpostX (dyads) for my study.

Thank you.
Best
Max

Two options:

In BEDPOSTX, the means of distribution of angles are stored in files mean_th[i]samples and mean_ph[i]samples.
The uncertainty of estimated direction can be inferred from file(s) dyads[i]_dispersion.

http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/FDT/UserGuide#BEDPOSTX