You recommend speakers based on measurements alone, and limited ones at that. I recommend speakers based on combination of measurements and listening, with the ultimate deciding factor being my ears.
Human hearing can not hear things that escape measurement. Only things that escape analysis is by fault of the person reviewing the data, presuming a sufficient set of measurements to determine a specific thing is present.
About the perceptive research, I get it. I've read all the papers you have, possibly even more. I have a couple of Toole's books, including his latest published just this year. But he and Harman have done no research on non-linear distortion or their impact, Geddes has but I disagree with his conclusion as do many others in the field. Anybody else want to read a bit without paying for the AES papers, here's some links
http://www.gedlee.com/downloads/The Perception of Distortion.pdf
http://www.harman.com/about_harman/technology_leadership.aspx
I believe the threshold for audibility is a bit lower than Toole has at times indicated, more like the .5-.7% range. But that's fine, I agree with just about all his other findings.
For music program, James Moire, has reviewed and discussed non-linear distortion data as it related to THD for several decades om regards to perceptual testing. His conclusion is that in terms of THD of a standard harmonic spectrum distribution, such as that with the average solid state amplifer operating in the non-clipping range, which happens to have a very similar spectral distribution as compared to the average speaker, that about 2 percent THD is required to be audible to sensitive listeners on music program, and recommends that 1 percent or under be implemented as a safety net. Of course, specific test signals, such as isolated sine waves, can increase audibility by several orders of a magnitude.
So what? The point isn't about that ... the point is foolishness of recommending a speaker based solely on a spattering of measurements taken by a 3rd party.
Sufficient measurements describe the primary characteristics. The speaker in question, based on the measurements, and presuming B&W did not leave gross non-linear distortion, a very safe assumption, the measurements describe a rather neutral mid-treble module with lower cabinet resonances than normal.
That you do not understand very how to correlate measurements to audiblity is no fault of mine. But it apparently is yours, since you don't really respect any type of scientific procedure enough to use this process yourself for accurate evaluation(s). I subject virtually everything to careful blinded/randomized testing for accurate analysis. It may take a lot more time - but it is worth it to remove one's subconscious bias from skewing the result(s). I don't depend on, nor trust my ears beyond anything but a preliminary evaluation for anything important without bias controls.
-Chris
