Re 1) if b-values (or b-vectors) are different across subjects, I hope you do NOT plan to use them in the same study, as such a study would not be appropriate, diffusion based properties (such as FA or GFA), are far to sensitive to changes in those.

If you have (e.g. for purpose of testing) different b-vectors or values, and don't want DTIPrep to test for that, simply disable the diffusion information check in the protocol file. 

In general, I think there is a misconception how DTIPrep operates. One should not change parameters etc on the command line, but rather create protocol files where all parameters (and there are many) can be adapted to the given study. It is simple to enable/disable motion correction in the protocol files, and that's the place where it should be modified (not on the command line).