rotational_minctracc.py should scale simplex to subject resolution

[mcvaneede]

We just found out that because rotational_minctracc.py uses the default value for the simplex (20mm), this stage fails fairly often at SciNet. There is still the question why when the same command run at MICe we don't have this issue, but scaling the -simplex value to something reasonable (1mm), fixes the issue at SciNet.

[bcdarwin]

The simplex is now settable with a default of 1 as of ffab864.