omnidirectional loudspeakers = best design available

[KA7NIQ]

Few will remember these speakers https://www.stereophile.com/content/hegeman-model-1-omnidirectional-loudspeaker
Stuart Hegeman was a mentor of Dick Shahinian who made the Obelisk Speakers mentioned in this thread.
Back in the late 70's when I worked selling audio equipment, we sold a Speaker by Design Acoustics that was basically a huge ball with drivers firing everywhere, and we were also an EPI Dealer. EPI made some huge towers with a driver array on all 4 sides! In effect, it was an omnidirectional EPI 100, if anyone remembers the EPI 100 ?
But unlike the EPI 100, it was a real room shaker because it not only had 4 - 8 inch woofers, but the woofers were in a large enough sealed box to play way low in frequency.

 

[Duke]

... we were also an EPI Dealer. EPI made some huge towers with a driver array on all 4 sides! In effect, it was an omnidirectional EPI 100, if anyone remembers the EPI 100 ?
But unlike the EPI 100, it was a real room shaker because it not only had 4 - 8 inch woofers, but the woofers were in a large enough sealed box to play way low in frequency.

Yes EPI/Epicure did not skimp on internal volume for their big quasi-omnidirectional tower speakers!

That warm tonal balance of the Epicure 400 and Epicure 1000 was partially due to what might be called "wrap-around". The 400 had four 6" midwoofers and four 1" tweeters, one on each side, while the 1000 had four 8" midwoofers and four 1" tweeters. So you would have one 1" tweeter and one midwoofer facing you, then going down in frequency as the radiation patterns of the midwoofers got wider and wider, more and more of the output of the other three midwoofers would "wrap around" and combine approximately in-phase with the one that was facing you. So on-axis you were essentially getting the output of one tweeter and one midwoofer in the midrange and treble regions, slowly transitioning to the output of ALL FOUR midwoofers in the low bass region! And along the way, any baffle-step issues were offset by the wrap-around energy from the other midwoofers.

The tradeoff was modest efficiency, but on the other hand the reflection field had essentially the same spectral balance as the direct sound, and imo the Epicure 400 stood out for its outstanding timbre. I presume the 1000 did as well but I never heard a pair.

In the late 70's I "rescued" a pair of Epicure 400's from a dealer who was deliberately clipping a 20-watt amplifier into them to demonstrate how much louder his high-efficiency house-brand speakers would play. They had their plusses and minuses but were to my ears a worthwhile improvement over other speakers I had heard.

 

[mhardy6647]

Back in the late 70's when I worked selling audio equipment, we sold a Speaker by Design Acoustics that was basically a huge ball with drivers firing everywhere, and we were also an EPI Dealer. EPI made some huge towers with a driver array on all 4 sides! In effect, it was an omnidirectional EPI 100, if anyone remembers the EPI 100 ?

EPI 1000.

Yeah, I - with great regret - turned down a pair a few years back because I really had no place to put them.
They are not small.

 

[KA7NIQ]

EPI 1000.

Yeah, I - with great regret - turned down a pair a few years back because I really had no place to put them.
They are not small.

View attachment 229195

I had a pair, moved to a small apartment after divorce, forced to sell them. BIG MISTAKE ! So big, so powerful, and also so musical!

 

[david moran]

I have not had the energy to wade through the back comments here, but following are some high-level comments to help a general understanding of the subject.

- all speakers are omni --- if you take a horn outside and measure it from the back, or a planar and measure from the side, there is not zero output, not even close --- but the amount of treble varies with design and sometimes it very very reduced compared w everything

- some famous 'omni' speakers have tried to be equi-omni, but without great directivity-matching care, which is terribly hard to do w lots of drivers, they are loby and therefore present inconsistent imaging and soundstage ... perhaps this is what someone meant by polydirectional - ?

- many omnis favor one direction or another; the first and most famous was the Bose 901, which favored rearward strength for the treble but was terribly sloppy in its horizontal radpat and hence casual in its image recreation

- lots of equi- / full-strength multipanel towers are loby, and how much it is bothersome is partly a matter of taste, but easily measurable --- Design Acoustics, Epi, Shahinian, the Allison flatback designs (half-omni) such as the One, the Two, and the much better-integrated IC20, about which more presently

- planars are a special case and never null to the side altogether, but sometimes are not totally sloppy in their rearward output (include here some Linkwitz and Dahlquist and similar designs)

- true non-loby equi-omni designs can be very cool, in fact most can sound nicely airy inasmuch as they get more treble into the reverberant field --- MBQ, BeoLab5 (excellent work, with treble beamforming), Ohm (after the first model, which was v inconsistent to produce, they went to a tweeter, to shape the horizontal radpat, meaning doing beamforming, above a certain point), and some others

- the excellent trick is do a phased array or similar directivity matching so you can a uniform and constant hor radpat of the flavor you like, as a matter of personal taste and also a matter of program type. The already mentioned Allison IC20 came close, being an attempt to mimic the gold standard of the dbx Soundfield line. That line comprised several omni designs with truly constant but oval-favoring overall, full-bandwidth output, a true phased array. Its shtick was that the oval was twice-loudness broadband in one direction, such that if you aimed a stereo pair toward each other they kept the images centered and stable (no lobes) as one moved offcenter. It came in out-from-wall and also flatback (half-oval) designs. But the standard now has to be the BeoLab90, which had a different radpat goal but achieved it very tightly, and with options for wide and narrow. It is a spectacular hor radpat achievement, but crazy pricy. Lexicon and one other company, name escapes me, starts w I, iirc, essayed user-controllable radpats along similar lines, but never made a real market mark. (The IC20 also had user hor radpat remote control.) Same 35y ago when Canon and some other companies attempted to make lensed horns duplicating the dbx SF work (as dbx was doing at the same time, at its end phase).

The dbx SF line and BL90 are well worth listening to if you get a chance, just for imaging precision and getting the music up in the air, as Allison owners like to put it.

It should go without saying that small baffles and modern narrower-diam drivers can come fairly close to the same effect. Still, the rearward output will be reduced not matter what.

Anyway, attempt to be precise when you say (or hear) 'omni', clarifying how lobal, whether favoring one orientation or other, how constant or beamy, and so on.

Here starting p7 is a ramble through some of these issues, with some measurements.

http://ethanwiner.com/aes/david_moran.pdf

written before the BL90 came out!

 

[mhardy6647]

Since someone mentioned them in an earlier post.
I'll just mention that the Shahinian Obelisks were in production for decades.

I heard a pair when they were first introduced (1977) at a small, funky (but savvy) hifi store in the (old) Baltimore shopping district and found them quite impressive sounding, irrespective of any 'omnidirectionality'.

Picked up the brochure for them -- which I still have.

 

[yourmando]

The dbx SF line and BL90 are well worth listening to

I’d love to hear the Beolab 90, even if I’d never buy $100k+ speakers.

@Kal Rubinson reviewed then and really only loved the “narrow” mode, leaving omni mode for a single parenthetical sentence:

“From that point on, I used Wide mode mostly to demonstrate to myself and to others just how remarkable the BeoLab 90s could sound in Narrow mode, in which I did all of my critical listening. (For its part, Omni mode is, well, party mode. I'll say no more about that—but I will note that the BeoLab 90's amplifiers are claimed to output 8200W, and that I very quickly gave up trying to test the speakers' dynamic limits.)”

https://www.stereophile.com/content/bang-olufsen-beolab-90-loudspeaker-page-2

And of course he’s heard a number of speakers since then that he’s liked better in his room, such as the Perlisten flagship tower and KEF blade for relatively “affordable” options, which are conventional directivity towers with a smooth DI.

 

[DanielT]

Check out this construction, omni :

2022 Parts Express [Speaker Design Competition] - Over $300 Category

Fun, inspiring video. Good variety of different types of design principles and sizes of speakers.:) That passive crossover filter, a speaker in the video, doesn't look very cheap, so to speak::) I would have been curious to listen to those:

audiosciencereview.com

 

[EJ3]

These have been my "dream" speaker for a long time:

BeoLab 5​

BeoLab Loudspeaker
Production: 09/2003 - present (as of when this was printed)
Designer: David Lewis

BeoLab 5 - the flagship model of the range active speakers - was, at the time of its introduction, the most radical loudspeaker ever made. At a huge investment cost the speakers were brimmed to their very top with innovations and fabulous sound.

The conical shape owed something to the Dalek school of industrial design, yet form closely followed function, enabling an enormous 380mm bass driver to be mounted in the base, pointing downwards. All four internal drivers had their own built-in ICEpower amplifiers - 1000W each for the bass drivers, plus 250W each for mid and top; a total of 2500 W of digital amplification per speaker. The mid-range and treble drives are vertical and radiate at horizontal discs, using so-called “acoustic lens technology” (licensed from Sausalito Audio Works) to disperse the sound in a horizontal plane of 180 degrees.

Usually, one can rely on the sound from speakers being dramatically modified by the room in which they are used. What distinguished the BeoLab 5 from every other speaker on the planet, however, was that they were designed to give the same results wherever they are placed, in whatever room they were being used. BeoLab 5’s go-anywhere ability was down to two unique features:

• one used carefully-shaped elliptical acoustic lenses to direct the sound of mid and high frequencies forward towards the listening area. The elaborately shaped top section meant the speaker effectively created its own ‘rear wall’
• the other was the bass end, powerful built-in amplification and the DSP (Digital Signal Processing) used to counter the effects of room gain. Using a microphone on a motor-driven rod, the system was able to set itself up automatically, pinging a test signal around the room and measuring the results with the on-board microphone. The process took around two minutes and could be repeated easily in case of moving them to a new environment or adding additional furniture.

In use, their impression was that of exceptionally precise and detailed stereo imaging, superbly even-handed overall neutrality and very low colouration throughout. The lens system was extremely effective and the speakers themselves seemed to lose themselves in the room, acoustically speaking.

The bass was about as good as it gets, imparting magnificent weight and power whenever required, while avoiding any unwanted thump, thickening or boxiness. A further bonus was that BeoLab 5 went very, very loud indeed, and did so without changing character or getting aggressive.

The BeoLab 5 was a remarkable and genuinely revolutionary loudspeaker and did a lot to enhance Bang & Olufsen’s hi-fidelity credibility. A pair would fill even the biggest room with seriously loud sound, and powerful, even bass. Very expensive but very good.

The speaker broke with Bang & Olufsen’s well-known design history in a key way, but was an outstanding example of how design could be subjugated to technology. BeoLab 5 was a revolution and an acoustic quantum leap. It buildt on a number of technologies where the Acoustic Lens Technology, a special way of creating an acoustic space between the two ‘plates’, had crucial importance. The designer David Lewis was tied down by the curved plate shape and so chose to play creatively on this theme. BeoLab 5 also contained a brilliant bass control which could program the acoustics of the room at the press of a button on the loudspeaker’s top. A small sensor read and adjusted the loudspeaker to the room and then reproduced absolutely pure sound.

Available in black/aluminium upon release, a white version was available from the end of 2008.

Bang & Olufsen Beolab 5: Real-world speakers, out-of-this world looks

Many hi-fi manufacturers manufacture ’statement’ products - machine designed to demonstrate technological and engineering acumen that will, hopefully, reflect some glory onto the rest of the company’s model range. At Bang & Olufsen, however, every product is a ’statement’ product. Never knowingly underdesigned, the company’s output has, for years, been pitched at the theoretical point where art and hi-fi get together.

From the very first, Peter Bang and Svend Olufsen’s company has married technology to elegant, idiosyncratic design: their 1927 ‘Five Lamper’ radio not only caused a sensation with its ability to run constantly from the mains supply but was also the first radio to be presented like an elegant piece of furniture, in a walnut and maple cabinet. Of course, new technological ground can be broken only rarely, but B&O have never strayed from their commitment to make dramatic and beautiful hi-fi, working in close conjunction with some of Denmark’s brightest architects and designers to that end. They have even rivalled Rolls Royce in the ’superiority’ stakes; for years, their advertising slogan read: ‘B&O - for those who discuss taste and quality before price’.

Whether or not the BeoLab 5 loudspeakers are tasteful is a matter of personal taste; what cannot be denied is that their appearance is dramatic. They squat broodingly, like post-modern sculpture or something Dan Dare would battle the Mekon for control of. The 19mm tweeter and 76mm midrange drivers are visible, firing upward at aluminium discs that serve as horizontal baffles, and sound is dispersed using Acoustic Lens Technology (theoretically giving uniform characteristics in front of the speaker). The 165mm upper bass driver is forward-firing, concealed behind the wraparound ’skirt’ of the main unit in a 5-litre chamber, while the lower bass driver is a mighty 381mm module which fires downwards from its 29-litre enclosure at the plinth on which the entire 61kg assembly stands. The plinth, though, does more than support the speaker, as I was soon to find out.

The connection points are concealed at the rear of the speaker. Each speaker contains 4 power amplifiers (one per drive unit), and has a ‘kettle lead’ mains connection. Each speaker also houses two power link connections, an RCA line-level input and RCA digital coaxial in- and outputs. The speakers can also be assigned right- or left-channel duties from the connection array. No power rating is claimed for the amplifiers on board, although as the maximum power output for each driver is quoted at 250W (tweeter and midrange) and 1000W (upper and lower bass), it seems safe to assume that amplification is adequate.

Performance

Positioning the speakers is easy, at least providing you’re able to lift them in the first place. Once they are pointing forward, touching the top of the speaker provokes a small microphone to appear at the plinth. The BeoLab 5’s inbuilt computer then begins the Adaptive Bass Control procedure, which seeks to determine the exact nature of your room’s acoustics. It generates a series of tones, starting with very low frequencies and moving up the range as the microphone extends further out of the plinth, as if the speaker is beginning its final checks before takeoff. Once the computer is satisfied, the mic retracts and the speaker falls silent. Follow the same process for the second speaker and you’re ready to go. Effectively, these loudspeakers can be placed anywhere in your listening room that takes your fancy.

From wherever you place them, the B&Os sound alive and involving. They’ll transmit the weight and scope of a full orchestra without breaking sweat, communicating Prokoflev’s Romeo and Juliet (Andrew Mogrelia and the Czecho-Slovak State P0 on Naxos) with effortless drive and attack. There’s tremendous realism to plucked violins and a fine complexity to the reproduction of the lower strings.

Integration and timing are first-rate, the tricky task of rendering the more sprightly sections of the ballet both jaunty and substantial pulled off with aplomb. In terms of sheer scale, the BeoLab 5s are nonchalant in their authority: dynamic shifts are communicated with thrilling speed and poise.

But they’re not all bombast. Small-ensemble recordings benefit from the same rapic, subtle characteristics that allow the tiniest detail to be retrieved and revealed even amid sonic mayhem. There’s a sense of unflappability and vigilance about the B&Os that makes each listen an event to be savoured. Of course, they’re not without a few sonic idiosyncracies to match their appearance - those who dislike a rapid and crisp top-end may find the BeoLab 5s a trifle fierce, and the midrange can ‘funnel’, giving a fleeting impression of cupped hands around a singer’s mouth, but these complaints are nit-picking in the extreme. The B&Os are speakers that will please for a lifetime - you needn’t know art to know what you like.

BeoLab 5 technical specifications​

Market	Type	Intro. year-month
EU	6881	2003-09
GB	6882	2003-09
USA-CDN-TWN	6883	2003-
J	6884	2003-09
AUS	6885	2003-09
KOR	6886	2003
CH	6887	2003-09
I	6888	2003-09

Connections:

Power Link	2
Digital SPDIF (IEC60958)	2 (PCM 32/44.1/48/88.2 kHz)
RS232 (vol. sync., service)	1
Phono (line)	1

Details:

Maximum sound pressure: 108 dB (stereo, pair)

Total amplification: 2500W

Power amplifier treble: 250W Class D ICEPower®

Power amplifier mid-range: 250W Class D ICEPower®

Power amplifier upper bass: 1000W Class D ICEPower®

Power amplifier lower bass: 1000W Class D ICEPower®

Effective frequency range: 20 - 20 000 kHz

Cabinet principle/ net volume upper bass: Sealed box / 5 litre

Cabinet principle/net volume lower bass: Sealed box / 29 litre

Directivity control treble and mid-range: Acoustic Lens Technology

Treble driver: 1.8cm

Mid-range driver: 7.6cm

Upper bass driver: 16.5cm

Lower bass driver: 38cm

Digital Signal Processor: 32-bit floating point 180 MFLOPS, 512k byte Flash-ROM

Room adaptation: Adaptive Bass Control (ABC)

Connections: 2 x Power Link, Phono (line), 2 x Digital SPDIF

Volume adjustment: Built-in

Dimensions/Weight (H x W): 97 cm x 49 cm / 61kg



Market specifications:

Type	Market	Voltage
6881	EU	230 V
6882	GB	230 V
6883	USA - CND - TWN	120 V
6884	J	100 V
6885	AUS	240 V
6886	KOR	220 V
6887	CH	230V
6888	I	230 V

Most recent S/W: 1.62A

 

[TraumaDisaster]

I just found the Ohm Walsh omni speakers and they seem cool based on the Steve Guttenburg Audiophiliac yt video I saw. I emailed them and want to try an Auro3D vog and center height 6 channel setup. Should fill my tall ceilings up nicely. If not I could try them as the base layer. Either way I'll test them on movie soundtracks that I know where details occur and see what I think.

 

[david moran]

People,
It might be helpful to go back to first principles at least for 2ch stereo playback -- ?

**All** speakers are omni to some extent and with some near uniformity up to some fairly high frequency (as in above middle C).

Forward-facing designs (*NOT* 'forward-firing'; they fire in all directions) beam progressively after a certain omni point, as I say. Dipoles and bipoles too, when you measure them directly sideways, have significant output, but again w treble rolloff (usually) and raggedy response. Horns, when you measure them directly behind, also have considerable broadband output, but again with significant treble rolloff (duh). A conventional tower design of dynamic drivers even moreso --- its rearward output looks quite like its forward output except for treble rolloff, w typically a much higher starting point than horn designs.

(Because the ear is so treble-oriented and audio is so historically treble-biased, we often talk about 'nulls' and 'forward-firing' and that sort of thing, without recognizing the near-omni world below 1k or 500Hz or lower. Middle C, higher than most of us can sing, is 260Hz or thereabouts, as you all know )

Now. What kind of horizontal radiation pattern is nicest for what kinds of program, well, that's a whole other discussion. Fortunately, how constant the hor radpat should be, that's a subject of recent interest and investigation. There is zero reason speakers should ever have been rolled off in the treble as a function of frequency, except that's how dynamic drivers work and thus necessity became this universally accepted virtue. And blind testing has shown that when the hor radpat (rolled off or not) is uniform w angle, people prefer it over a hor radpat that is all lumpy and scalloped and has these big 'flare-ins' where the smaller driver enters.

Directivity matching is better understood and aimed at than ever before, in other words.

Most earlier omni-ish designs were sloppy and loby, to put it nicely. That may not matter crucially when we are far off, but with closer listening it can be problematic, depending as always on how much you care about stable soundstage presentation and image creation and how sensitive you are to it in the first place. The typical tradeoff is for pleasing airiness and spaciousness. The Bose 901 has that, no question, but the image formation is beyond casual, also huge, even as it was based on a bunch of bogus faux-academic handwaving that ignored the changing hor radpat w frequency. Shahinian (Design Acoustics, others) ditto, different but similar. Forward-facing drivers not in a vertical line similarly a little scattershot, somewhat: Rectilinear, AR3, ADS, dozens of sundry like designs. Planars and dipole/bipoles had their own pluses and minuses as to attractive floatiness and front-wall reflections, a related but side set of discussions. The Allison One and Two were fine from afar but loby up close, yet had immense air (driver design, not just the doubling-up). Snells and ARs and Kefs (B&Ws), like others back when, solved the lobal problems but could sound comparatively airless. (Spaciousness is also a function of seating distance, of course, as is room dryness especially laterally.) Ohm was the first serious and successful effort to have broadband omni output with some constancy and no lobing, but there were whopping consistency issues.

The dbx Soundfield initiative attempted a shaped omni (or half-omni) broadband hor radpat without objectionable scallops using conventional drivers in a true computer-implemented phased array. Huge critical success. There were a very few imitators, and some, including dbx itself, undertook lensed horns to achieve the same end.
Since that time BL (and Lexicon and Kii and once upon a time Canon) advanced the shaped- or nearly full omni state of play with their own multidriver phased arrays or lensed horns.

But anyone here can play around with what they have: reorienting the cabinets and possibly turning up the treble, placing cabinets back-to-back or at right angles, and on and on. Experimenting with horizontal radiation pattern is easy (a little schlepping) and cheap and very rewarding.

If I have gripped this lectern and thumped this tub before here, please forgive.

For those who like weeds, there are hor radpat and other measurements in these two windy history recits:

https://www.bostonaudiosociety.org/pdf/bass/BASS-17-06-9005.pdf (p11 hashandy theoretical hor radpat models)

http://ethanwiner.com/aes/david_moran.pdf

 

[NiagaraPete]

The most important aspect of home audio reproduction, is the role of the loudspeaker. Omnidirectional loudspeakers produce the most realistic musical soundstage in the home; however they remain the least understood by the public and audiophools alike. The superiority of the design is easily heard when in person, and when measured, particularly in the off-axis domains both vertically and horizontally. Live music and therefore sound propogation happens omnidirectionally, and is how our ear-brain mechanisms have evolved to understand sound. 98% of loudspeakers are designed incorrectly; yes you read that right and it's easily verifiable w/ measurements. Most loudspeakers beam the sound toward the listener in a totally unnatural way, and suffer what are called 'lobing effects' and again, are easily measured particularly when you start to move off-axis. (marketing department of companies don't measure nor publicize these measurements once out of the sweet spot, because they are horrible). These concepts have been substantially explored by the late Sigfried Linkwitz of Linkwitz Labs; I believe all his research and papers are available at the website. Further explanation is available at the website of Morrison Audio loudspeakers, highly recommended. It is important to note, that there is major distinction between polydirectional speakers, and omnidirectional.

Makers of omnidirectional loudspeakers include:

Ohm Acoustics
German Physiks
MBL
Mirage (out of business; still available used)
Linkwitz Labs (LX Mini is a hybrid omni)
Duevel
Morrison Audio

I would encourage anyone who is serious about music, and wants the most realistic soundstage in their home, to pursue omnidirectional loudspeakers. Contrary to audiphool misunderstanding, they actually excite the listening room LESS than conventional speakers, and require LESS or no special room treatments.

Rules for making a great omni:

-design should be 2-way. Single driver is inadequate; 3-way is unnecessarily complicated
-woofer driver should face UP, with lots of room behind it to reduce or eliminate back wave from radiating back out the cone
-tweeter should also be facing UP
-both woofer and tweeter should be place immediately together, and with dispersion caps or guides to disperse the outgoing sound both vertically and horizontally
-cabinet should be totally inert, as measured by accelerometer
-speaker should be able to be driven using either a passive or active crossover
-inputs should be Neutrik Speakons; Benchmark Media has measured conclusively the lowered distortion of Speakons compared to binding posts (spades or bananas)
-listening height of drivers should be at listener's seated ear level OR LOWER

This considered, the 2 best omni designs, and therefore the 2 best loudspeaker designs in the world right now, are from Duevel in Germany, and Morrison Audio in Canada. All others on the list are "honorable mention".

Yikes

 

[krabapple]

I want a 'uniform hor radpat NOW!' T-shirt

It's a pleasure to see you here, Mr. Moran.

(Readers here may not all be aware this is the David Moran of the Boston Audio Society, AES, and co-author with E. Brad Meyer of the emperor's-clothes-revealing JAES article "Audibility of a CD-Standard A/DA/A Loop Inserted into High-Resolution Audio Playback" from 2007, which provoked often vicious and mostly ignorant backlash from subjectivist circles, and much amusing goalpost-shifting along the lines of 'buh buh they didn't use REAL hi-rez SACDs/DVDAs!'.)

 

[Newman]

I did know that, but nevertheless Mr Moran seems to be historically overinvested in omnidirectionality as a principle.

They are critically constrained to the second tier when pure audible performance is the priority:-

1. Floyd Toole has written of the importance, to naturalness of sound, of having speakers with some commonality in directionality with instruments. Omnis fail entirely in this regard.
2. They ‘play the room’ so much more than most speakers. The exact size shape and treatment of the room thus become far more critical to the result, along with the need for extreme care and attention to room treatment. This issue is exacerbated by point 1 above.

Mr Moran would be well advised to show some more agnosticism, flexibility and adaptability in his viewpoint. The original case for omni speakers grew from some mistaken ideas about music being omni, plus addressing some limitations of 2-channel playback that we all have the opportunity to overcome today by moving to MCH. The idea that, today and looking forward, omni speakers still represent some kind of pinnacle, is IMHO entirely misplaced.

cheers

 

[BenB]

I did know that, but nevertheless Mr Moran seems to be historically overinvested in omnidirectionality as a principle.

They are critically constrained to the second tier when pure audible performance is the priority:-

1. Floyd Toole has written of the importance, to naturalness of sound, of having speakers with some commonality in directionality with instruments. Omnis fail entirely in this regard.
2. They ‘play the room’ so much more than most speakers. The exact size shape and treatment of the room thus become far more critical to the result, along with the need for extreme care and attention to room treatment. This issue is exacerbated by point 1 above.

Mr Moran would be well advised to show some more agnosticism, flexibility and adaptability in his viewpoint. The original case for omni speakers grew from some mistaken ideas about music being omni, plus addressing some limitations of 2-channel playback that we all have the opportunity to overcome today by moving to MCH. The idea that, today and looking forward, omni speakers still represent some kind of pinnacle, is IMHO entirely misplaced.

cheers

The idea that multi-channel satisfactorily addresses the limitations of 2 channel playback is quite contentious. To my ears, stereo playback always sounds much more natural than MCH does.

 

[Newman]

That would be the mix or your gear. The technology itself delivers new levels of perceptual preference.

 

[Sancus]

The idea that multi-channel satisfactorily addresses the limitations of 2 channel playback is quite contentious. To my ears, stereo playback always sounds much more natural than MCH does.

How is it contentious? It's been pretty resoundingly proven by many research sources, read Toole's book. "I don't like it" and "contentious" are not the same things. This is exactly the same as saying "well I don't like flat on-axis response and smooth off-axis response therefore building a speaker this way is contentious". Of course it isn't.

 

[oivavoi]

Mr Moran would be well advised to show some more agnosticism, flexibility and adaptability in his viewpoint.

Mr. Newman would be well advised to show some more humility when interacting with audio experts whose knowledge of audio research in all likelihood exceeds his own.

 

[Newman]

Well it takes him 5 minutes to start calling people “snot” and “dick”, so humility works both ways.

Meanwhile, all I did was summarise the evidence. Let’s talk evidence.

 

[BenB]

How is it contentious? It's been pretty resoundingly proven by many research sources, read Toole's book. "I don't like it" and "contentious" are not the same things. This is exactly the same as saying "well I don't like flat on-axis response and smooth off-axis response therefore building a speaker this way is contentious". Of course it isn't.

I re-read the chapter on Multichannel audio in Toole's book. I failed to find sufficient scientific evidence to support the claim that listeners reliably find commercially available multi-channel mixes to sound more authentic than stereo mixes when played back using commercially available equipment in typically sized home listening spaces. There was a lot of text devoted to summarizing a paper by Muraoka and Nakazato that used non-traditional recording practices to measure an omni-directional loudspeaker on a stage, and then played that back in an anechoic chamber with various multi-channel speaker configurations. Then only one aspect of the sound was assessed (Frequency-Dependent Inter-aural Cross-correlation), which falls very (very) short of establishing an over-all improvement of authenticity by listeners. Additionally, for science to be complete, the work actually has to be independently verified. In short, they recorded something in an way not widely adopted by the producers of multi-channel content, played it back in a space hugely dissimilar to the typical listening room, and measured a single characteristic of the reproduced sound, and there was no mention of any success in re-producing even those limited results.
Less text was devoted to a study by Hiyama in 2002, and fewer details were provided. This appears to be a subjective test where listeners were asked to assess "the degree of impairment in perceived envelopment" of a small number of speakers trying to emulate 24 speakers. Toole fails to mention that once again, this study was performed in an anechoic chamber, which obviously has little to no bearing on home audio reproduction, where reproduction of the characteristics of a large space would have to compete with the actual space around the listener. Additionally, "envelopment" is but one aspect of the sound, and from my own digging it sounds like they used incoherent noise signals on the speakers in order to try and isolate this one characteristic. Music reproduction is so much more complicated than that, so this fails to establish that even in an anechoic chamber it would be possible to authentically re-create the spatial aspects of music performed in a space (symphony hall, jazz club, whatever).
Regarding your dig about the characteristics of an and off-axis sound and how it relates to listener preference, regardless of what I believe regarding that, if you are referencing the work by Harman, you should understand that it's about half done. A short summary of their work goes something like this:
They had a group of listeners record preferences of various speakers when listening to them without the ability to see them. They made measurements of the speakers, and ran some regression analysis to see if they could closely re-create the recorded preference scores by weighting various aspects of the speaker measurements. They found that they could. Then they got another group of listeners together to listen to a larger set of speakers (if I recall correctly). They once again recorded blind listener preference. They used the weighted characteristics from the first test to try and re-create the preference scores from the second test, and found out it didn't work very well. So they applied new regression analysis, and came up with a new model with a better fit to the data. Now, in order for this to constitute "science", an independent group needs to verify that this model holds up when they conduct their own experiments with their own speakers and listeners. As far as I know, this has never been done.