I have never listened to a bipole, but wonder what is the advantage (or more likely disadvantage) in terms of boundary interaction when compared to a dipole, at least judging from this diagram?
The summed off-axis energy of a bipole is more spectrally-correct than that of a monopole, because of the additional mids and highs that are going into the reverberant field. As depicted in your diagram the bipole's summed off-axis energy is less spectrally-correct than that of the dipole or omni, this because the bipole's radiation patterns (front and back) generally narrow with increasing frequency.
Some of the corresponding tradeoffs of the bipole include a reduced direct-to-reverberant sound ratio (which may not be a disadvantage in all situations), and a recommendation that the speakers be positioned far enough out from the "front" wall to avoid the backwave energy arriving too soon. Ime this recommendation applies to dipoles and omnis as well.
(Despite the radiation pattern advantages of dipoles and cardioids I choose not to use them because, in my experience, these types convey less "impact" in the bass and lower midrange regions than monopoles and bipoles do.)
Here is an article I wrote for an online magazine which mentions some of the particulars of the bipolar configuration I used:
The Controlled-Pattern Offset Bipole
And here's what they looked like (same drivers on the back, but close to the floor):
What I find interesting about the Fenicia is the fact that the rear-firing array is a speaker that can be orientated, EQ'ed and it's level attenuated, a topology which unlike others gives the user significant control over the effects of "envelopment" and "spaciousness" rather than 'dictating' one particular form of presentation.
I absolutely agree that Sonus Faber's approach with the Fenice is an improvement over a straightforward bipolar loudspeaker. My current designs (such as the Azel you mentioned
above) have a rear-firing driver which is adjustable in level and spectral balance, though admittedly mine lacks the aim-ability of the rear-firing array on the Fenice.
At this point we are veering far from an accurate reproduction or transduction of the signal in the sense that there's a significant level of 'distortion' which results from the room interaction.
Imo this goes straight to the heart of the matter.
If the spectrally-correct rear-firing energy of a bipole or dipole or Fenice/Azel-style multidirectional speaker is "distortion", then isn't the spectrally-incorrect off-axis energy of a monopole loudspeaker even
more of a "distortion"? Yet most people do not prefer quasi-anechoic listening environments for conventional monopole loudspeakers. So if our concept of "accurate" is informed by controlled blind studies of preference, reflections are not necessarily distortions.
Taking the Fenice as an example, presumably its additional rear-firing energy is actually
improving the spectral balance of the reverberant field. Seen through this lens, it is arguably
reducing distortion.
So I take the position that the intelligent deliberate addition of off-axis energy can improve accuracy, to the extent that the spectral balance of the reflections matters to the ears.
Now obviously the deliberate addition of off-axis energy will decrease the direct-to-reverberant ratio, which can be counter-productive, especially if the additional reflections arrive too soon. So imo there are still tradeoffs to be juggled.
Which raises an interesting question:
Is reproducing what the engineers were listening to a reasonable, universal goal?
I can only say that it might be for some people...
The goal of "hearing what the engineers heard" certainly makes intuitive sense, but (at the risk of oversimplifying) the engineers were listening for what needs to be fixed, so they used tools optimized for that purpose, not the least of which is room acoustics which cost them many tens of thousands of dollars to achieve. Imo this not the same goal as "creating the most credible illusion".
... but others may find that particular presentation wanting, and may prefer an "enhanced" version, through the use of dipoles, omnis or even (gasp) upmixing. This results from the fact that stereo is but an effect in itself and that audio reproduction falls short of the realism one experiences with live music.
Yes!!!
Imo there is arguably an inconsistency between "minimizing the circle of confusion" and "tasteful upmixing to embellish the sense of space."
I agree with you, it's wonderful that there are all those different topologies which allow the end user to try out and choose whichever one he or she prefers.
"Standardisation" is in my opinion a huge mistake, as is trying to influence people's preferences through research.
The Fenice/Azel approach offers an unusual degree of adaptability, as the reverberant sound can be optimized (for listener preference or room acoustic situations) without simultaneously changing the direct sound.
I have a feeling you are re-inventing the wheel. You are following the footsteps of Professor Bose who designed the 801 (wooden enclosure, not the plastic pro model) back in the 80s followed by more models with drivers facing all over the place. He gave up at the end.
At first glance, it would appear so.
But looking a bit deeper, here are three things that imo Dr. Bose did not get right:
1. The reason a low direct-to-reverberant sound ratio works well in a concert hall is because the reflection path lengths are very long, allowing the ear adequate time to clearly hear the direct sound stream separately from the reflected sound stream. In general the less time interval between the direct sound and the strong onset of reflections, the lower the level of the reflections needs to be in order to preserve clarity. In all fairness this may not have been known at the time Dr. Bose designed the 901, but the net result is that the direct-to-reverberant sound ratio of the Bose 901 is MUCH TOO LOW for home audio. Too much reverberant energy ends up arriving too early, and clarity is degraded.
2. The spectral balance of the those reflections is significantly different from the spectral balance of the direct sound. The rear-firing energy is generated by a slightly splayed array of 8 fullrange drivers, and the inevitable comb filtering decreases the relative amount of short-wavelength (high frequency) energy in the reflections, not to mention that the spectral balance of the rear-firing drivers started out as their power response rather than their on-axis response to begin with.
3. Dr. Bose's stated goal was 11% direct sound and 89% reflected sound, so he used one driver forward-facing and eight rear-facing. BUT his forward facing driver's output all by itself would have been maybe 1/4 direct sound and 3/4 reflected sound at normal listening distances. So what he ended up with was more like 3% direct sound and 97% reflected sound.
Hopefully I have not re-invented
that wheel.