Those look tiny
(and very professional )
If you can measure and correct the response (sorta like a Linkwitz Transform), and are planning a steep enough acoustical crossover slope (or a high enough crossover frequency) then a BR is often a better solution as it reduces excursion/distortion.
This can all be simulated using winISD pro (beware that the EQ implementation in this software is a proportional Q type, whereas your BSS uses constant Q...)