Omnidirectional speakers

[jim1274]

I’m starting to think there’s something important on the Omni “sound” compromises that somehow I’m just not hearing. I’ve followed all the arguments best I can, but feel like I’m getting into deeper and deeper water without a life preserver, the sharks circling in an ever smaller and smaller circle.

 

[Newman]

I do not criticize upmixing as "adding something to" the original recording, whether or not a case could be made for that characterization. Imo improvements in sound quality and/or spatial quality need not justify themselves against a dogmatic yardstick.

Cool

But note that in the scenario you describe, "having the sound coming from speakers all around you and when your room is nicely damped", the ONLY directions spectrally-correct reflection energy will arrive from are directly from the surround speakers...

I hope you don't mean that in the perceptual sense. That would be like saying the only directions stereo sound comes from is the two speakers, perceptually. Which we know isn't true. Similarly, we would perceive that ambient energy coming from a broad spread.

If I understand Toole, and hence the experimental evidence, correctly, surprisingly few surround channels are needed to satisfyingly generate the perception of envelopment and apparent source width. Maybe four...maybe even two in modestly sized rooms.

 

[MarkS]

That back-handed pot-shot of a comment shows a remarkable ignorance about how science works, and how learning from scientists works. Thanks for sharing: it's very helpful to know. I shall also note who gives it a Like, which also yields quite helpful information.

Let us know when you are ready to move on from the starting point of "everything is relative, and everyone's preferences are different and unique". Some people just can't get past it.

Dude, I'm gonna pull rank on you: I have a PhD in physics from Stanford. I've been a Professor at an R1 research university for 40 years. I've authored a popular graduate-level text and well over 100 papers in refereed journals.

So please take your insufferable Science Church Lady schtick and shove it down your tweeter.

 

[Newman]

I’m starting to think there’s something important on the Omni “sound” compromises that somehow I’m just not hearing. I’ve followed all the arguments best I can, but feel like I’m getting into deeper and deeper water without a life preserver, the sharks circling in an ever smaller and smaller circle.

Which brings me back to the last paragraph of my first post to you: adaptation is your friend. Same for all of us, actually.

Plus, another thing that applies to all of us, is our individual suite of personal biases, which lead to confirmation bias effects, placebo effects, the whole 'sighted listening effects' shooting match.

So don't sweat it too much. You are having a ball with omnis. Long may that continue! Even if it's a dream! My intention isn't to make you change your speakers, or to make you unhappy instead of happy with omnis...that would make me sad.

I'm only trying to bring some of what audio science says about omnis into awareness and into the discussion. Why? Because, when we try to post-rationalise our sighted listening impressions, we are very prone to misattribution. It's a human trait. Such misattribution can lead us down rabbit-holes of logic and conclusions that are flat-out wrong, and when we share them on the internet, wrong-headed myths arise and echo chambers of opinion take hold.

Now, there are audio discussion forums on the internet that are veritable petrie dishes for such myths and opinions, but ASR isn't one of them.

cheers

 

[Newman]

Dude, I'm gonna pull rank on you: I have a PhD in physics from Stanford. I've been a Professor at an R1 research university for 40 years. I've authored a popular graduate-level text and well over 100 papers in refereed journals.

So please take your insufferable Science Church Lady schtick and shove it down your tweeter.

Wow, now you have even less justification for what you wrote. And for the last sentence above, too. Goodness me!

 

[jim1274]

Which brings me back to the last paragraph of my first post to you: adaptation is your friend. Same for all of us, actually.

Plus, another thing that applies to all of us, is our individual suite of personal biases, which lead to confirmation bias effects, placebo effects, the whole 'sighted listening effects' shooting match.

So don't sweat it too much. You are having a ball with omnis. Long may that continue! Even if it's a dream! My intention isn't to make you change your speakers, or to make you unhappy instead of happy with omnis...that would make me sad.

I'm only trying to bring some of what audio science says about omnis into awareness and into the discussion. Why? Because, when we try to post-rationalise our sighted listening impressions, we are very prone to misattribution. It's a human trait. Such misattribution can lead us down rabbit-holes of logic and conclusions that are flat-out wrong, and when we share them on the internet, wrong-headed myths arise and echo chambers of opinion take hold.

Now, there are audio discussion forums on the internet that are veritable petrie dishes for such myths and opinions, but ASR isn't one of them.

cheers

The comments challenging the Omnis and what I’m hearing have been as much or more helpful than the confirmation that they do some things right. Makes me try harder to identify the negatives and think critically. I won’t abandon the Omni just because of attacks on the concept. I’ve been playing them 14 hours a day for a long string of days, enjoying along the way, so that’s not all bad even if they have flaws.

 

[Keith_W]

As yet another scientist on this thread, I should note that disagreements between scientists is normal. Toole says one thing, Linkwitz says another. They are arguing from different perspectives - in a nutshell, Toole is arguing for accurate studio reproduction in the listening room. He wants us to narrow the circle of confusion by using transducers set up with the same principles as the recording studio. Linkwitz is arguing that omnis better approximate how humans perceive sound naturally because sound comes from everywhere. This is why he designed dipoles and omnis. That is fine, one of them may be right, or both may be partially correct.

You can pick apart arguments from both. e.g. it is possible that the circle of confusion will never be narrowed because even studios can not standardize something as basic as a frequency response - see here. IOW we are being asked to standardize our home stereo systems to a standard that our recordings do not meet themselves. I am certainly not advocating throwing out the baby with the bath water here, I pursue accuracy in my system, I use Toole's methods, and I have read his book. I am just pragmatic about what can realistically be achieved.

I also have sympathy with Linkwitz's position on dipoles and omnis. His proposition makes sense - humans are surrounded by an enveloping sound field, this is how we hear. We are adept at detecting the direction of sound and its reflections, and our hearing is tied in with the vestibular system and the position of our head. Yes, I accept that no recording studios use omnis and it is adding an "effect" not found in the recording studio as Toole says. But even Toole believes that reflections are essential and he advocates against over-deadening a room.

So let's look at what both of them have in common. Both believe that a flat response under anechoic conditions AND constant directivity is what makes a good speaker.

So really, the point of disagreement is this: how much should we "omni"? An arc of an omni? The whole thing? If an arc, how much of an arc? Do we accept part of the frequency response as omni, and want an arc for the other part? I suspect that this question will never be settled until a preference study is done.

 

[ShadowFiend]

Great post, Keith. I can add one more important thing is smooth frequency response off-axis and great bass extension and capability

A preference study about dispersion is not easy to be done, and even when it is done, the accuracy is questionable, simply because the preference study is only accurate when the test speakers represent an high quality implementation of the subject and do not violate the preceding criteria. For example with omni speakers, it omnidirectional dispersion need to be up to at least 5-7kHz and all those criteria like flat response on-axis, control directivity aka smooth frequency response off-axis or better constant directivity, low distortion, bass extension capabilities. This requirement rules out the vast majority of omni speakers in the market, leaving only the most expensive one remaining.

In addition to that, those test systems need to be near identical bass capabilities. Which means if the Revel Salon 2 or Magico M3 represents conventional box speaker, then MBL 111F need to be chosen as equivalent, not the less-capable model MBL 126. Thirdly, all those test speakers need to be placed in their optimal position. And lastly, there must be great amount of testers, at least 50-100 people, so that the result can be more or less ok from statistical point of view.

Considering high quality domestic audio is not really lucrative, I doubt this kind of expensive study, if done right and can be consider accurate, can happen anytime soon.

 

[Flaesh]

If an arc, how much of an arc?

90º for rectangular room

 

[Newman]

As yet another scientist on this thread, I should note that disagreements between scientists is normal. Toole says one thing, Linkwitz says another. They are arguing from different perspectives - in a nutshell, Toole is arguing for accurate studio reproduction in the listening room. He wants us to narrow the circle of confusion by using transducers set up with the same principles as the recording studio. Linkwitz is arguing that omnis better approximate how humans perceive sound naturally because sound comes from everywhere. This is why he designed dipoles and omnis. That is fine, one of them may be right, or both may be partially correct.

Sure, and the way to test disagreements, or determine who is ‘righter’ or least partially incorrect, , is to design and execute carefully controlled experiments. I think credible scientists would agree about that.

And, comparing Toole to Linkwitz in that respect, I think there is a clear difference in their efforts to test their ideas.

You can pick apart arguments from both. e.g. it is possible that the circle of confusion will never be narrowed because even studios can not standardize something as basic as a frequency response - see here. IOW we are being asked to standardize our home stereo systems to a standard that our recordings do not meet themselves. I am certainly not advocating throwing out the baby with the bath water here, I pursue accuracy in my system, I use Toole's methods, and I have read his book. I am just pragmatic about what can realistically be achieved.

I also have sympathy with Linkwitz's position on dipoles and omnis. His proposition makes sense - humans are surrounded by an enveloping sound field, this is how we hear.

Yes, and one aspect of adding some realism to the sound field in the listening room, is for speakers to direct sound in a manner that bears some similarity to voice and instruments, however impossible it may be to do perfectly.

We are adept at detecting the direction of sound and its reflections, and our hearing is tied in with the vestibular system and the position of our head. Yes, I accept that no recording studios use omnis and it is adding an "effect" not found in the recording studio as Toole says. But even Toole believes that reflections are essential and he advocates against over-deadening a room.

So let's look at what both of them have in common. Both believe that a flat response under anechoic conditions AND constant directivity is what makes a good speaker.

So really, the point of disagreement is this: how much should we "omni"? An arc of an omni? The whole thing? If an arc, how much of an arc? Do we accept part of the frequency response as omni, and want an arc for the other part? I suspect that this question will never be settled until a preference study is done.

As far as I can tell, Toole’s published work put a lot of effort into examining that question, including preference studies. And the answer was, more or less, that a standard forward-firing speaker has a more suitable dispersion characteristic, as long as the beamwidth is suitably controlled vs frequency.

Toole has gotten to the point where he reckons that, “These days there is little or nothing to be gained in distinctive transduction processes beyond marketing claims.” That is not consistent with the notion that some important preference studies are still needed on the question of “how much omni?”

cheers

 

[Keith_W]

90º for rectangular room
View attachment 340033

Not all listening rooms are like that, though. There are speakers placed in the middle of a large room, non-rectangular rooms, rooms with curved walls, rooms which open to other rooms or the backyard ...

I suspect that there is no single right answer that applies to everybody. I have heard precious few omnis, but quite a few dipoles. Those speakers seem to sound best when they are placed further away from the front wall, and this is a completely subjective opinion. I have never owned omnis or dipoles, because dealing with reflections from monopole speakers gives me enough headaches already! But this is not to say that I don't like them, I just think they are too much trouble for me.

If you ask me "how much should we omni?" my answer would be something along the lines of, "depends on your preference and your room". I would say that you should choose the radiation pattern that suits your needs the best.

 

[Keith_W]

A preference study about dispersion is not easy to be done, and even when it is done, the accuracy is questionable, simply because the preference study is only accurate when the test speakers represent an high quality implementation of the subject and do not violate the preceding criteria.

Bingo! Scientists like variables to be controlled, biases to be eliminated, and interventions to be blind. And, as you rightly point out, the study needs to be funded and nobody is going to fund it without a commercial imperative.

However ... there is one speaker that we could use for such a study ... the B&O Beosound 90. This is an omnidirectional speaker that is able to vary its radiation pattern with DSP. What we need is for Amir or Erin to cosy up to B&O and measure the radiation pattern at different settings on his Klippel. This may not even be needed, I am sure B&O already have this data. Then plonk them in a room and run a succession of people through it. It is easy to blind them to the intervention because no change in speaker is required. It wouldn't even cost very much! Call up a B&O showroom which has the Beosound 90 set up, gather 10-20 friends, and arrange a session. Publish and be famous.

 

[JPA]

Dude, I'm gonna pull rank on you: I have a PhD in physics from Stanford. I've been a Professor at an R1 research university for 40 years. I've authored a popular graduate-level text and well over 100 papers in refereed journals.

So please take your insufferable Science Church Lady schtick and shove it down your tweeter.

And I thought all you had going for you was an affinity for Yes.

 

[Purité Audio]

Bingo! Scientists like variables to be controlled, biases to be eliminated, and interventions to be blind. And, as you rightly point out, the study needs to be funded and nobody is going to fund it without a commercial imperative.

However ... there is one speaker that we could use for such a study ... the B&O Beosound 90. This is an omnidirectional speaker that is able to vary its radiation pattern with DSP. What we need is for Amir or Erin to cosy up to B&O and measure the radiation pattern at different settings on his Klippel. This may not even be needed, I am sure B&O already have this data. Then plonk them in a room and run a succession of people through it. It is easy to blind them to the intervention because no change in speaker is required. It wouldn't even cost very much! Call up a B&O showroom which has the Beosound 90 set up, gather 10-20 friends, and arrange a session. Publish and be famous.

https://bangolufsenassistentgohe.blob.core.windows.net/manuals/SPEAKERS/BEOLAB_90/beolab_90_technical_sound_guide_version_9.pdf

I spent a great deal of time with the 90 when they were released, the slight irony is that in the UK at least B&O shops are a franchise, the owner I spent time with had been retailing washing machines, he had no knowledge of hi-fi ( or consumer electronics ) whatsoever .
Even at the main Hanover Square outlet the speakers were badly set up and no-one had really figured out how to get music to them.
Great speakers lost in a corporate machine, perhaps they are better demonstrated now.
Keith

 

[Flaesh]

B&O Beosound 90

Beolab 90 is not an omni, it's just the name of the mode; The data is even in the manual IIRC.

my answer would be

Mine is the same)).

Not all listening rooms are like that, though.

There is a grain of truth in every joke. True omnis has constant horisontal directivity. The sketch above depicts 90 degree omni sectors and less than normal amount of reflected sound..
Outdoor listening of true omnis will allow subjectively evaluate the usefulness of perfect horizontal directivity separately from room reflections

 

[Keith_W]

Beolab 90 is not an omni, it's just the name of the mode; The data is even in the manual IIRC.

Mine is the same)).

There is a grain of truth in every joke. True omnis has constant horisontal directivity. The sketch above depicts 90 degree omni sectors and less than normal amount of reflected sound..
Outdoor listening of true omnis will allow subjectively evaluate the usefulness of perfect horizontal directivity separately from room reflections

Thanks to Keith (not me, the smarter one) I dug it out from the manual:

Not the smoothest looking omni, but omni enough.

 

[Purité Audio]

It is an interesting speaker to listen to because it really highlights the narrow vs wider dispersion, in narrow the imaging is ultra pin point the sound quality to me was the best of the three but you had to be sitting in the ( unsurprisingly) narrow hot seat, wider was just that much less focused and the sound a little more indistinct.
My wife who was with me that particular found the ‘omni’ mode quite unsettling, she wasn’t in the kitchen.
Keith

 

[NTK]

90º for rectangular room
View attachment 340033

 

[ShadowFiend]

From Beolab90 is only arguably omni up to 3kHz, and the off-axis is rough so in my book it is not a good implementation of omni speaker. So, any opinion about omni speaker based on this Beolab90 is misleading, just like saying dipole is bad based on experience with a fullrange driver in an open-baffle setup

 

[hvbias]

Well here again the music matters. I listen mostly to classical, and the very best recordings IMO are Telarc's minimally miked ones. And I preferred the sound of those recordings on bipolar DefTechs compared to conventional forward-firers when I went to buy some new speakers 20 years ago. (I went to hear the well-measuring Paradigm Studio 60 v3, and was underwhelmed.) I also liked their sound on arena rock, which in real life you hear over giant banks of PA speakers.

However bipolars and omnis have fallen out of popular favor. Mirage (makers of the bipolar M1 owned by Floyd Toole back in the day) is out of business, DefTech's founders left and the new owners dialed down the rear arrays by 6dB. The only bipolar speakers left that I know of are the Axiom LFR series (Axiom calls them omnis but they're really bipolar). True omnis, as noted above, are very rare.

So when one of my DefTechs went wonky three years ago, I ended up with forward-firing replacements, GoldenEar Triton 7s. (I hadn't yet found ASR; if I had, I might have Revel F35s instead.) I'm happy enough with them, though I still sometimes miss the solidity of the sound field that I experienced when moving around the room with the DefTechs.

Have you heard MBL speakers with classical?