Thank you very much pbrain for that in-depth reply.
"I think the in-your-face character of the M2 is mostly due to its higher directivity index across the audio band. That’s not atypical of horns in general."
I agree.
My study of acoustics and psychoacoustics and the implications of the size rooms we listen in at home leads me to think that all reflections should be spectrally correct; that early reflections (those arriving within the first ballpark 10 milliseconds) are likely to be detrimental; and that late reflections (those arriving after about 10 milliseconds) are likely to be beneficial. To clarify a bit about the early reflections, they have some benefits and some detriments: They can widen the soundstage and make the speaker less obvious as the sound source, but they can also degrade clarity and cause coloration. Just about everything that late reflections do is beneficial, assuming they are spectrally correct, and assuming you don’t have a significant room acoustics issue like slap echo. A sense of immersion and envelopment comes primarily from the late reflections, which come from all directions; the ear/brain system is able to pick out the ambience information already on the recording from these late-onset reflections, identifying it by its harmonic structure.
So in the shoot-out, with the single speaker far from both side walls, you had no early-onset sidewall reflections to possibly degrade the sound. Aside from the floor and ceiling bounces (which are subjectively relatively benign), all of the reflections were late-onset, and spectrally correct. So what I THINK is that the significantly greater amount of late-onset reflections for the Salon 2’s helped them to sound more natural.
With a more conventional stereo setup, left and right speakers being much closer to their respective side walls, I think the M2’s would have benefitted from their narrower radiation pattern reducing the amount of energy in those early sidewall reflections. But in all fairness, my understanding is that Toole finds these early reflections to be generally beneficial, while Geddes and Griesinger find early reflections to be generally detrimental, so there are well-informed differences of opinion on the subject.
My own somewhat unorthodox approach to horns involves fairly narrow patterns, and then additional drivers aimed such that their reflected outputs arrive "late" (more than 10 milliseconds behind the first-arrival sound). In other words, my best horns speakers might be called "polydirectionals", to use the term coined by the late great Richard Shaninian. I’m trying to minimize the amount of energy in early reflections, but then increase the amount of (spectrally-correct) energy in late-onset reflections.
It’s has not been obvious to me which is the most beneficial: Minimizing early reflections, or increasing late reflections. But the test you participated in seems to suggest that increasing the amount of energy in late reflections probably matters the most, which is something I’ve been working on. And THAT’S why I personally find this test so interesting.
By any chance will you be attending T.H.E. Show in Long Beach in two weeks? If so, I would REALLY welcome your brutally honest critique of what I’m doing. Imo your background makes you uniquely well qualified.
Duke