This will likely get me in some hot water, but here it goes..
When a speaker manufacturer produces a speaker with a built in passive crossover, they (hopefully) try to get the best response from the cabinet/driver combination that they are utilizing. Simply removing the internal passive and going with an active frequency crossover, cutting the signal at the proper point and removing power robbing components from the signal path seems good - but: shelving, notch filtering and any other response correcting processes from the original crossover design are eliminated.
Now what? To get it right - in YOUR space with YOUR other components, you need to use a microphone/analysis process to make things right. Simply going active without the additional step can be counterproductive.
I'm looking very hard at getting the MiniDSP to give it another try. YMMV.