You're measuring in-room, yes? I think the misinterpretation is stemming from my failure to point out the need for an anechoic chamber. Sound Power is the sum of all sounds radiated from the loudspeaker, so you must measure all angles. Combined with a directivity index, direct sound, and listening window, one can pretty well surmise how well a loudspeaker will perform, and be preferred, in-room, assuming one has absorbed 550 pages of Floyd's work.
@Floyd Toole would absolutely agree that for domestic listening purposes, the direct sound and early reflections are important ("dominates" as you put it). However I would say that it is a terrible assumption that "the reflected sound fills in and averages out to roughly mimic the on-axis response". That is the goal, yes, but so rarely is that accomplished!
Interaural Cross Correlation is the reason a linear omni mic is not an accurate representation of what we hear. However, Toole and Olive were able to establish an algorithm with a .86 (86%) coefficient for listener preference, based on the aforementioned measurements and double blind testing. Only those with something to hide (NOT Philharmonic Audio) would shy away from this kind of testing!
Back to the KEF's, I heard the later hissss but not the initial snap of cymbals, but also harshness of some female vocals. They are very pretty speakers, until you have something to A/B compare with! When your AA's, or even Sony Core's were put on, the wonderful experience of the KEF's disappeared.
