Unsound, thank you for YOUR thoughtful response. I'll try to answer some of your questions:
"Wouldn't such a controlled radiation pattern reduce the size of the sweet spot and reduce the sound-stage, especially for multiple listeners?"
On the contrary, my experience with 90-degree constant-directivity waveguide-style horns indicates that they will give you a much wider than normal sweet spot when set up correctly. I use about 45 degrees of toe-in, such that the axes criss-cross in front of the listening position. Now before I go on, note that the ear localizes a sound source by two mechanisms: Arrival time and intensity. With good pattern control and this strong toe-in, for the off-centerline listener the far speaker is actually louder (greater intensity) because the listener is on-axis relative to it and well off-axis relative to the near (first-arrival) speaker. The two localization mechanisms partially offset, and you still get an enjoyable (though not perfect) soundstage even when well off to one side. The key is the smooth, fairly rapid off-axis roll-off of the near speaker. The tonal balance holds up well throughout the room.
"Wouldn't the amount of reverberant sound be greater in most indoor live venues too?"
Absolutely. In fact, I believe that two key discrepancies between live and reproduced sound are the spectral balance of the reverberant field (addressed by radiation pattern control), and the relative lack of energy in the reverberant field of most home listening rooms as compared with a live indoor performance. The latter leads us towards wide-pattern speakers, or omnis, or "poly-directional" speakers (Richard Shahinian's word). The type of poly-directional speaker I embrace is the controlled-pattern offset bipole.
"So long as the reverberant sound is not too close in time, shouldn't we be able to hear this as a reverberant sound and not as distortion?"
Yes indeed. The concept of "reverberant sound = distortion" is flawed; the ear/brain system actually uses some types of reflections to improve speech intelligibility. Part of the reasoning behind the bipolar approach I prefer is maximizing the time interval between the first-arrival sound and the onset of increased reverberant energy.
"Furthermore, couldn't this reverberant energy be controlled via room treatment and/or room correction?"
In my opinion, control should start with the sound source - the loudspeaker. It is difficult to do frequency-specific room treatment, and room correction via signal processing is microphone-position-dependent. The less spectral variation from one location to another within the room, the greater the listening area in which room correction is clearly beneficial.
"If my budget permitted, I'd guess that I'd move from cones & domes 'n boxes to top quality omnis. Perhaps the antithesis of what you've described as an advantage."
Well, a good omni inherently excels at generating a reverberant field that has the same spectral balance as the first-arrival sound, which is one of my top priorities. The only conceptual problem I have with omnis is the large distance from sidewalls required to avoid a detrimental early sidewall reflection.
"Aren't there already existing remedies for such thermal compression in many cones & domes and not really much of an issue for alternative drivers?"
My understanding is that, in general, drivers have less than 1 dB of thermal compression when used at less than 10% of their RMS rated output. Of course there may well be exceptions to this rule, and it's always the exceptions that are the most interesting.
"Am I correct in assuming that the cross-overs you describe aren't digital and therefore are probably incapable of preserving correct time and phase?"
Yes, I'm talking about passive crossovers. I do not believe that waveform preservation is a top priority, because the ear does not hear waveforms as such. At a certain very high level of performance, where more audible problems have been resolved, then waveform fidelity would be a significant issue. In my opinion, from a perceptual standpoint the tonal balance is the most important issue, but the perceived tonal balance includes the reverberant energy as well as first-arrival sound.
"It would appear to me that this ideal matching of non-horn loaded woofers to the rest of the horn loaded drivers must be rare in deed, all the horn loaded systems I've heard are blaringly bright. I have still yet to hear a horn system that's colouration's are below audibility."
Understood! In my opinion as well, it is rare. I could of course suggest a horn system or two that pays attention to these issues [rolls eyes innocently], but I must also acknowledge that some people's ears are simply more sensitive to horn-type coloration than others. If by any chance you'll be at the Rocky Mountain Audio Fest, please come by and let me know how I'm doing.
Ralph, you are correct - I left out compression driver quality. The field coil Beryllium diaphragm units John Wolff uses are magnificent. I was thinking in terms of more modest price ranges, where diaphragm resonances have to be addressed in the crossover - admittedly less than ideal, but many direct-radiator domes have comparable issues.
I think there are probably size and price floors below which a horn system is unlikely to be competitive with a direct radiator system, but above those thresholds I think good horn systems generally become increasingly competitive.
