As a layman...
...and more of a hobbist who barely knows the difference between a Helmholtz resonator and a corn-dog(other than the guess that the resonator has greater nutritional value)...I was quite satisfied to see my thoughts were on the same track as jneutron...
I too, am also of the opinion that the bass module would be the focal point for more the substantial "cabinetry" in order to minimize resonances...however if one looks at the shape of the AM satts, a bit more is revealed as to inherent enclosure strength...Viewed from above, these "cabinets" describe an arch, probably one of the more structurally efficient forms in architecture. I would take a guess that, much like their vertical counterparts, this design more evenly distributes and withstands the stress they would encounter as a containment for the backwave of the enclosed drivers...which in my mind raises another point, a 2.5in driver mounted on a 3in. square "baffle" should have fewer problems arising from cabinet refractions, a subject which seems to have become a bit of a bugaboo in modern loudspeaker design...unless of course you resort to a cabinet design that harkens back to the tail-end of a '58 Olds...
Heading off to the AM drivers, I believe they take their cues from the premise of the 901s, in that nine 4.5in. drivers have the effective radiating area of a 12in.woofer with far less mass. Since there are two of them per/ch...the total effective radiating area would be the equivalent of a >3.75in. mid unit and the higher freqs would be naturally divided and subsequently handled by the individual smaller transducers and ultimately by their dustcaps...much as the "whizzer" cones from the days of yore. Simple, effective and cheap.
Insofar as the frequency response plots exhibited are concerned, they raise a few questions. Are they skewed in some fashion due to the seemingly wide spread and inherent bias against Bose products in general? Given the fact that the satt modules can be "aimed" to provide from very "focused" to wider dispersion patterns, under what circumstances where these plots arrived at? If at max direct/reflect mode, certainly room effects would play a part. Were the data taken from a point commonly termed the "sweet spot" as it would be positioned, based on the recommended placement parameters from the manufacturer? If not, the results would be suspect IMHO. Also take into consideration the net response of the Acoustimas bass module is placement dependent.
Please take note my comments are based on the Acoustimas 5 system...I have little experience with the 3-2-1 or Lifestyle systems, as I personally have a problem(as evidenced by the connectivity quotient of my "collection") with closed-end plug'n'play products of this type.
I have always submitted that fully appreciating the Bose experience requires a different listening mindset and again, my input is limited to, and based on, my personal experience with the 901-lls and is a direct result of the constant bash, bash, bash. I am not urging everyone to sell what they own and head to the nearest Bose store or retailer...I simply suggest giving their products a listen and deciding for yourself, rather than take the opinions gleaned from, of all places, the internet...as it plumbs the depths of it's vast puddle of knowledge.
jimHJJ(...just some more thoughts...)