From what I can assess very expensive drive units aren't necessarily more than that: very expensive. I'm not saying they're incapable, but quality drivers can be made for much less (why then spend more?), and implementation is key. It seems to me the drivers in Harbeth speakers are capable of doing what they're meant to do, and a lot of research goes (and has gone) into their voicing and tonality by ear, to a degree that seems unequaled. The late Peter Snell chose rather standard looking drivers for the A/II model from Becker, Philips and Audax, but the speakers allegedly sounded wonderful (have only heard the later A/IIIi which was not of Mr. Snell's voicing, but a friend of mine has extensive experience with both models, and clearly favored the older A/II's). On the other hand, the Living Voice Vox Olympian's/Palladium's use drivers from UK-based Vitavox (among others) which cost downright fortunes (some £3,100/3,300 for the bass and midrange driver respectively, a piece), but these are clearly chosen following extensive tuning and implementation and for very specific purposes.
What's interesting is also Harbeth's attention towards enclosure "tuning," regarding them as an essential part of the overall sonic imprinting and contribution. Many if not most manufacturers seem almost (blindly) hellbent on killing vibrations first and foremost, rather than "playing along" with their contribution in having an ear for their signature, and this way they may end up with detrimental sonic by-products not intended. Going by the non-resonant enclosure approach, when rather successful, I find exposes the sound of the drive units all the more (and not necessarily for the good), and moreover has a tendency to deaden the sound in ways that comes off unnatural to my ears. Harbeth's approach seems a rather smart one, in that they combine a sensitive ear for live, natural sound, and knows how to apply or convert this into engineering and construction that will eventually present a fairly authentic sonic facsimile.
