No, not anechoic, but first-order lateral reflections were >15 msec. I find it difficult to look at the description of preference testing like this and make any generalization about wider directivity or downward-sloping off-axis response being preferred or "more pleasing" in this setting. I suspect that the bass extension of the Array 1400 may have compensated for the less even on-axis and listening window response.
I was coming up with about 12 milliseconds for the first sidewall reflections (single speaker in the middle of a 21 foot wide room, ten feet listening distance), which is imo enough time delay to make it benign (assuming it's spectrally correct). It might still impart some image widening, since spatial attributes are discernible in Harman's single-speaker tests. But it's very encouraging to me that the narrow-horizontal-pattern Array 1400 is a solid contender.
I presume that the Array 1400's pattern is wider in the vertical than in the horizontal based on eyeballing the horn. Some researchers find that early reflections in the vertical are more benign than early reflections in the horizontal because they arrive at both ears simultaneously, and then (if not absorbed) that wide vertical energy comes back as beneficial later-arriving reflections. So I speculate that the unorthodox configuration of the Array 1400 might minimize early sidewall reflections while still sneaking a fair amount of energy in the later reflections.
In other words, the Array 1400's unorthodox configuration MIGHT actually be an improvement over a more conventional configuration which pattern-matches in the horizontal plane. Up until now the Array 1400 has looked ridiculous to me, but now I'm seeing it through a different lense.
I thought it was interesting that this study suggested "that listeners can be categorized into two different preference classes. Some listeners prefer clarity over reverberance and the others love strong, reverberant and wide sound," also that they assessed three classes of attributes grouped into "reverberance, loudness and width;" "timbre"; "definition and clarity" but that "the individually collected attributes were also different depending on the choice of music."
One of the things which can be adjusted with the system I'm doing is the direct-to-reverberant sound ratio. There is a threshold below which the additional reverberant energy does not degrade clarity, and for the "clarity" group, that would be the "sweet spot". The "reverberance" group may prefer to cross that threshold and trade off some clarity for more envelopment.
My feeling is that when it comes to speakers, probably there is a third, much smaller listener preference class related to timbre, for which the tonality of the speaker and its ability to reproduce the "correct" sound of an orchestra and specific instruments like violin or piano is paramount (REG comes to my mind here), hence issues like floor dip may become extremely undesirable and off-axis BBC dip might seem to compensate for microphone proximity.
I'm guessing that Tapio Lokki (what a cool name!!) did not include "timbre" as a third preference category because it is assumed to be correct since the source in a concert hall is live, unamplified instruments. On the other hand in home audio, timbre depends heavily on the quality of the source, which in this context is the loudspeaker system (ignoring for now Toole's "circle of confusion"). My foray into polydirectional speakers started out as a pursuit of more natural timbre, and it was only later that I became aware of possible spatial benefits. But too much of a good thing ceases to be a good thing if clarity is a top priority.
So I'm doomed to never build a "best" speaker, as it seems I'm always juggling tradeoffs. Maybe what I really need is a better marketing department.