I would like to have tested a bit more but decided to upload Beta31 before my vacations (which start now). It contains a bugfix yesterday reported (hopefully). In eulcidean clusterin denoising. I made modifications to the QSM modelling step.
For those of you who do not know what the switches I added to, leave everything to standard. That is how I modelled in the verisons before. If you like to detect yourself geometrical outliers (wrong fits) rather than my own routines, you can deactivate and try to improve the models yourself in R.