So something like the Genelec "The Ones" design which is noted for equally great vertical dispersion is not subjectively favorable? It would be interesting if you came up with some sort of vertical dispersion cut-off that represents the "ideal" situation.
Like I said, I simply don’t know. But in my KEF R3 vs Ascend Sierra 2EX blind test results thread(s), three major categories of hypotheses were put forward towards explaining the KEF R3's loss (by a large margin) despite having superior (traditionally interpreted) spin measurement results:
Hypothesis 1. It's an outlier: That it was just random chance that both blind test participants overwhelmingly preferred the Ascend 2EX over the KEF R3. Or that by chance my room (or subwoofer crossover) interacted more favorably with the 2EX somehow, even though the KEF R3 may be a bit flatter (though they’re both quite neutral speakers and sound quite similar sonically).
Hypothesis 2. Since it’s unlikely we’re interpreting the measurements wrong, KEF therefore must be “massaging” their published R3 measurements somehow to look better than the performance of the production models they ship.
Hypothesis 3. The KEF R3 published measurements are not wrong, and therefore our model used to predict speaker preference from spin measurements is not quite right. Either there's an important dimension we're missing when predicting preference (e.g. perhaps something that shows up in polar plots or the full sound field, but gets averaged out in the spinorama), or we're incorrectly weighting the degree to which some factors (e.g. directivity index) contribute to overall speaker preference, or a bit of both.
Hypothesis #1 is pretty much always a valid concern (which always comes up in any blind test), but similarly, it’s therefore not productive to dwell on this unless it means actively working to gather more blind test data.
This review from Amir is particularly fascinating because it completely rules out hypothesis #2! We now have
independent confirmation that the KEF R3 does indeed have among the best measurements ever seen in a speaker.
This leaves us only with hypothesis #3 (and #1 relating to not having enough data) to explain why the
KEF R3 didn’t dominate in the blind listening test (or in many other anecdotal sighted comparisons, including Amir’s).
I have absolutely nothing invested in one theory or the other, except towards us finding the truth — we’re all here because we want to find the closest possible thing to objective audio perfection.
Again, the data on the KEF R3 is really fascinating here because if our current interpretation of the measurements is correct, the R3 should be
dominating blind shootouts against just about anything else —
including $20,000 - $30,000 Revel speakers like the Salon 2’s, which don’t measure as well IIRC.
Fortunately, this is something we can test, in theory. So far, multiple anecdotal sighted tests and one blind test (mine) has not reinforced the measurement-predicted outcome that the KEF R3 should dominate pretty much everything else on the market. So we need an explanation for that, and/or more data from the blind testing side.
