When in DIY, why test all possible combinations of speaker elements and crossovers? Instead, wouldn't it be more meaningful to test individual speaker elements and effects of different crossovers components over them? There really is not much public distortion measurements of them plain speaker elements, that would also be comparable. This would also be a lot cheaper than testing full builds.
Also another aspect. Should one wish to see how different geometries affect distortion/directivity/etc, the comparison should be made with different constructions, but using the same elements (the "standard element" for bass, mids and treble). That way results would be comparable. There same goes for crossover components.
Also don't forget pro-audio speakers. Many DIYers are using them where space is not restricted.