It is possible to generate a very wide sweet spot using asymmetric horns. Read up on Project Everest (I), S2600 and S3100.
There's a review on S2600 in one of the LHF threads that explains it succintly. Link it here if you come across it, please.
Buyers didn't get it, basically, and JBL opted to sell them center channel speakers, instead....
Wow. I learn something every day that I drop in here. Thanks Zilch.
I came across these two reviews, though I'm not sure if either is the one to which you were referring:
New Speakers in Da' House- JBL S2600
JBL 4430 vs S3100 vs S2600 in home
The
picture on the LHF explain asymmetric horn theory well. Many have said that the sweet spot of waveguide/horn speakers such as the Econowaves can be widened by toing-in the speakers such that they cross an imaginary point "in front" of the listener.
All of this leads me to conclude that accurate imaging results when the audio signal from left and right channels arrives at the listening position according to (at least) three theretical aspects:
1. frequency response must be the same (including any reflections)
2. amplitude must be the same
3. arrival time must be the same
It has been observed that signals that are the same in frequency response and that arrive within a relatively small time window, ten(?) milliseconds or so, are combined by our brains into a single sound. So, for practical purposes we might be able to ignore distance differences to each of the two speakers for listeners who are not equidistant from the speakers in a home listening environment.
These asymmetric horns, project Everest I, and the other comments from folks on various fora suggest that when sound arrives in equal amplitude (#2) as well as equal frequency response (#1), good imaging is achieved. Project Everest and these asymmetric horns apparently were an attempt to have output vary with distance off-axis, so that for a listener positioned more directly in front of one speaker would have that speaker's sound attenuated to a level that would match, in level, the sound coming from the speaker that is further away.
If this is all right, then this is very exciting because it tells us that there is a mathematical solution that optimizes imaging for situations where there is more than one listener. Using the off-axis SPL level and the various distances that folks are from the speakers, we can adjust the amount of toe-in or even overall toe-angle in order to put the sweet spot on the listeners.
It would seem the asymmetric horns attempt to accomplish level matching without having to adjust the speaker angle depending on where the listeners in the room are. Brilliant.
Are S2600 horns available?
How do the Everest I/S2600 style asymmetric horns compare with Econowave for the equidistant listener?
Do the asymmetrics create any other problems, such as with the reflected sound that makes them undesirable?