Why are there so few omnidirectional speakers?

[Duke]

omni speakers have higher sound power in high frequencies than forward firing speakers. Omnis can radiate directional high frequencies with as much output as omnidirectional frequencies like 100hz in a room, as a result an omni speaker with flat on-axis response can theoretically form flat response reflections. That is not the case for conventional speakers.

I agree with this part.

An omni speaker with flat on-axis response sounds significantly brighter than a forward firing speaker.

True, but "flat on-axis" is not the only option available for "voicing" an omni speaker.

Higher sound power of omni speakers can be balanced with rolling off the direct sound with same magnitude and slope where sound power of forward firing speakers start rolling off however direct sound is perceptually dominant in human perception. If direct sound is recessed, reflections can't make it sound balanced. (or if the direct sound is bright, reflections with recessed treble tonality can't make it sound right) Basically there is no way to make recordings sound balanced on omni speakers with recordings that was mixed and mastered on forward firing speakers. Especially treble is doomed to sound wrong on omni speakers.

It sounds to me like you're making the assumption that "flat direct sound, rolled-off-highs in the reflected sound" is the most correct way to deliver a recording to listeners, perhaps because that is preferred in a conventional loudspeaker. Imo it is a "best compromise". Let me explain:

Exactly as you say, "if the direct sound is bright, reflections with recessed treble tonality can't make it sound right". You have just described the situation which arises with conventional loudspeakers! Those most highly rated have the best combinations of "flat direct sound which would be bright all by itself" and "rolled-off reflections which would be dull all by themselves". The best combinations of the bright spectral balance of the flat direct sound and the duller spectral balance of the reflections ends up in a "happy middle ground" voicing which looks like the Harman in-room preference curve (I'm thinking of the thick dotted line but the others are also experimentally-established preference curves):

Omni loudspeakers result in LESS spectral discrepancy between the direct and reflected sound. So if an omni's response (naturally both on and off axis) is conceptually aligned with Harman's in-room preference curve (though probably not identical to it), then its perceived voicing will fall into that "happy middle ground" between bright and dull, BUT with the arguable advantage of reduced spectral discrepancy between the direct and reflected sound.

Here is why I think reducing the spectral discrepancy between the direct and reflected sound is beneficial:

Harman's research shows consistent preference for conventional speakers which minimize the spectral discrepancy between the direct and reflected sound, this characteristic showing up as a really good spin-o-rama. So doing an even better job in that area makes intuitive sense, assuming there are no overriding downsides.

I'm NOT saying that good omnis are across-the-board better than good conventional speakers; they are certainly more challenging to integrate into a room well, and I do not think their inherently low direct-to-reverberant sound ratio is ideal. I have not owned omnis for forty years, nor do I build or sell them. But imo under good conditions (another topic for another day!) they can do some things really well, with conveying natural-sounding timbre being one of them.

 

[dlaloum]

omni speakers have higher sound power in high frequencies than forward firing speakers. Omnis can radiate directional high frequencies with as much output as omnidirectional frequencies like 100hz in a room, as a result an omni speaker with flat on-axis response can theoretically form flat response reflections. That is not the case for conventional speakers. An omni speaker with flat on-axis response sounds significantly brighter than a forward firing speaker. Higher sound power of omni speakers can be balanced with rolling off the direct sound with same magnitude and slope where sound power of forward firing speakers start rolling off however direct sound is perceptually dominant in human perception. If direct sound is recessed, reflections can't make it sound balanced. (or if the direct sound is bright, reflections with recessed treble tonality can't make it sound right) Basically there is no way to make recordings sound balanced on omni speakers with recordings that was mixed and mastered on forward firing speakers. Especially treble is doomed to sound wrong on omni speakers.

Real life experiences do not reflect this...

Here is a review of an omni - and the experience of the reviewer and the measurements show it is very flat - with a light rise up above 10kHz

mbl 101E Radialstrahler loudspeaker Measurements

Sidebar 3: Measurements

www.stereophile.com

 

[Blockader]

I agree with this part.



True, but "flat on-axis" is not the only option available for "voicing" an omni speaker.



It sounds to me like you're making the assumption that "flat direct sound, rolled-off-highs in the reflected sound" is the most correct way to deliver a recording to listeners, perhaps because that is preferred in a conventional loudspeaker. Imo it is a "best compromise". Let me explain:

Exactly as you say, "if the direct sound is bright, reflections with recessed treble tonality can't make it sound right". You have just described the situation which arises with conventional loudspeakers! Those most highly rated have the best combinations of "flat direct sound which would be bright all by itself" and "rolled-off reflections which would be dull all by themselves". The best combinations of the bright spectral balance of the flat direct sound and the duller spectral balance of the reflections ends up in a "happy middle ground" voicing which looks like the Harman in-room preference curve (I'm thinking of the thick dotted line but the others are also experimentally-established preference curves):
View attachment 232281

Omni loudspeakers result in LESS spectral discrepancy between the direct and reflected sound. So if an omni's response (naturally both on and off axis) is conceptually aligned with Harman's in-room preference curve (though probably not identical to it), then its perceived voicing will fall into that "happy middle ground" between bright and dull, BUT with the arguable advantage of reduced spectral discrepancy between the direct and reflected sound.

Here is why I think reducing the spectral discrepancy between the direct and reflected sound is beneficial:

Harman's research shows consistent preference for conventional speakers which minimize the spectral discrepancy between the direct and reflected sound, this characteristic showing up as a really good spin-o-rama. So doing an even better job in that area makes intuitive sense, assuming there are no overriding downsides.

I'm NOT saying that good omnis are across-the-board better than good conventional speakers; they are certainly more challenging to integrate into a room well, and I do not think their inherently low direct-to-reverberant sound ratio is ideal. I have not owned omnis for forty years, nor do I build or sell them. But imo under good conditions (another topic for another day!) they can do some things really well, with conveying natural-sounding timbre being one of them.

I said *Basically there is no way to make recordings sound balanced on omni speakers with recordings that was mixed and mastered on forward firing speakers.*

I said that mixes that is done on forward firing speakers can not sound right on omni speakers. For the same reasons you stated above. Because well designed studio speakers have flat on axis response with a downsloping directivity above from omnidirectional frequencies to directional frequencies. Music that was mixed with constant-change-in-DI speakers can't sound right on flat DI speakers(omni directional speakers)! Flat DI speakers have too much acoustic output compared to forward firing speakers thus they sound bright with *normal* recordings. I am also partial to flat DI speakers because reflections that has closer timbral characteristics to on-axis sound are more beneficial or ASW, timbre and imaging. That being said, standardizations in recording studios have to completely change for flat DI speakers to be relevant in music production and reproduction.

 

[Blockader]

w

Real life experiences do not reflect this...

Here is a review of an omni - and the experience of the reviewer and the measurements show it is very flat - with a light rise up above 10kHz

mbl 101E Radialstrahler loudspeaker Measurements

Sidebar 3: Measurements

www.stereophile.com

And that would make them only sound brighter and less preferable.

 

[Duke]

I said *Basically there is no way to make recordings sound balanced on omni speakers with recordings that was mixed and mastered on forward firing speakers.*

I said that mixes that is done on forward firing speakers can not sound right on omni speakers. For the same reasons you stated above. Because well designed studio speakers have flat on axis response with a downsloping directivity above from omnidirectional frequencies to directional frequencies. In other words, music that was mixed with constant-change-in-DI speakers can't sound right on flat DI speakers like omni speakers are!

I understand the assertion you are making.

"music that was mixed with constant-change-in-DI speakers can't sound right on flat DI speakers like omni speakers are!"

Why?

 

[Blockader]

I understand the assertion you are making.

"music that was mixed with constant-change-in-DI speakers can't sound right on flat DI speakers like omni speakers are!"

Why?

Flat DI speakers have too much acoustic output(higher sound power)in high frequencies compared to flat on-axis forward firing speakers thus they sound bright with *normal* recordings.

 

[Duke]

flat on-axis Flat DI speakers have too much acoustic output(higher sound power)in high frequencies compared to flat on-axis forward firing speakers thus they sound bright with *normal* recordings

Yes, that would be a way to make a bright-sounding omnidirectional speaker (the room also gets a vote but let's leave the room out of it for now).

What about Flat DI speakers whose on-axis curve is like the Harman preference curve? That way they are not bright with "normal" recordings.

Omni speakers can aspire to get both the direct and reflected sound spectrally correct simultaneously. Conventional speakers can't do that; they can only aspire to a best-case-compromise between the two. (That being said, my personal preference leans towards something more like the not-exactly-"conventional" JBL M2, even though its reflected sound has extra energy where the woofer's pattern is widening.)

 

[Blockader]

Yes, that would be a way to make a bright-sounding omnidirectional speaker (the room also gets a vote but let's leave the room out for now).

What about Flat DI speakers whose on-axis curve is like the Harman preference curve? Then they are not bright with "normal" recordings.

Because sounds that arrive later than the direct sound may be perceived as spatially separated auditory images which coexists with the direct sound but the direct sound is perceptually dominant. (precedence effect or echo suppression effect as Haas calls it)

That is why reflections with recessed HF information can't make bright on-axis sound balanced.

How do you plan to avoid direct sound while you have an omni directional speaker in your room? It radiates sounds everywhere.

 

[Duke]

Because sounds that arrive later than the direct sound may be perceived as spatially separated auditory images which coexists with the direct sound but the direct sound is perceptually dominant.

Okay, I'm not sure what your point is here. Could you elaborate?

Edit: Ah, you edited while I was typing (I do that too!). "(precedence effect or echo suppression effect as Haas calls it)". That's a big topic; I'm still not sure what your point is... ?

That is why reflections with recessed HF information can't make bright on-axis sound balanced.

I disagree, because "reflections with recessed HF information" are precisely what make bright (flat) on-axis speakers sound balanced in-room. Take those same speakers outdoors where there are no reflections and they will sound bright. The in-room reflections are not "dominant" in the same sense that the first-arrival sound is, but they DO influence perceived timbre. If you don't believe me, try singing in the shower.

How do you plan to avoid direct sound while you have an omni directional speaker in your room? It radiates sounds everywhere.

Why would I want to "avoid direct sound"? I'm confused.

 

[Blockader]

I disagree, because "reflections with recessed HF information" are precisely what make bright (flat) on-axis speakers sound balanced in-room. Take those same speakers outdoors where there are no reflections and they will sound bright. The in-room reflections are not "dominant" in the same sense that the first-arrival sound is, but they DO influence perceived timbre. If you don't believe me, try singing in the shower.

They do influence the timbre, that is what I am talking about too however for human auditory system direct sound is perceptually dominant. Reflections may add their own timbral colorations to direct sound however during this fusion interval where reflections and direct sound get fused to each other human auditory system is still aware of how the direct sound sounded. If reflections start having significantly different timbre than the direct sound, our hearing registers a new direct sound and starts a new fusion interval.

Toole talks about that in his book in the chapter where he goes into details of precedence effect.

What makes flat on-axis speakers with constant-change-in-DI sound balanced is that they play a recording which was also recorded/mixed/astered with speakers with on-axis sound and constant-change-in-DI. Speakers are, as you know very well, a playback device. There is a source(recording) and speakers apply their transfer function to it. If the transfer function is 1, what we hear is an accurate reproduction of the source.

Basically you can't make a speaker with bright on-axis sound balanced with recessed HF reflections. To our auditory system, direct sound is perceptually dominant. If direct sound sounds bright, you can't make it recessed with recessed reflections. That is not how it works. Our hearing is not linear like microphones have. Speakers with same in room response, may not sound same. That is why it is not recommended to EQ speakers above room transition frequency. But that is a topic for another thread.

 

[Duke]

They do influence the timbre, that is what I am talking about too however for human auditory system direct sound is perceptually dominant. Reflections may add their own timbral colorations to direct sound however during this fusion interval where reflections and direct sound get fused to each other human auditory system is still aware of how the direct sound sounded. If reflections start having significantly different timbre than the direct sound, our hearing registers a new direct sound and starts a new fusion interval.

I agree with you that reflections "do influence the timbre", and can color the direct sound during the fusion interval. A reflection which has the SAME spectral balance as the direct sound will not color the direct sound. A reflection which as a DIFFERENT spectral balance may cause coloration depending on the specifics.

And a reflection whose spectral balance is so degraded that it's no longer recognizable as such has ceased to be "signal" and has become "noise".

What makes flat on-axis speakers with constant-change-in-DI sound balanced is that they play a recording which was also recorded with speakers with on-axis sound and constant-change-in-DI.

Are you saying that the recording can only sound "balanced" when played back on speakers which generate essentially the same (somewhat spectrally incorrect) reflections that the recording engineer's speakers did?

If so, this may be where our disagreement lies. You and I seem to have similar understandings in many relevant areas but for whatever reason we arrive at different conclusions about what would theoretically sound "balanced".

Speakers are, as you know very well, a playback device. There is a source(recording) and speakers apply their transfer function to it. If the transfer function is 1, what we hear is an accurate reproduction of the source.

The input signal is amplitude (voltage) changing over time. The output signal is amplitude (sound pressure) changing over time AND dispersed into three-dimensional space with a pattern that usually changes significantly with frequency. Seems to me that the LESS the dispersion pattern changes with frequency, the closer we are to your one-for-one transfer function target.

Basically you can't make a speaker with bright on-axis sound balanced with recessed HF reflections. To our auditory system, direct sound is perceptually dominant. If direct sound sounds bright, you can't make it recessed with recessed reflections. That is not how it works. Our hearing is not linear like microphones have. Speakers with same in room response, may not sound same. That is why it is not recommended to EQ speakers above room transition frequency. But that is a topic for another thread.

Wait - a minute ago you said that reflections "do influence the timbre", and that "reflections may add their own timbral colorations to direct sound"...

Are you now saying that reflections don't have enough influence on timbre to change "bright" to "recessed", and vice-versa?

If so, then it sounds to me like our difference on this topic boils down to how much effect reflections can have, and I'm not sure that's worth arguing about. If your position includes "reflections do influence timbre" and "reflections may add their own timbral colorations to the direct sound", I think that's sufficient common ground.

But I would still like to know whether you're saying that a recording can only sound "balanced" when played back on speakers which generate essentially the same (somewhat spectrally incorrect) reflections as the recording engineer's speakers.

By the way, THANK YOU for the civil discourse, hopefully we can disagree without being disagreeable. If I cross that line let me know, because sometimes I get carried away.

 

[Blockader]

I am not a native english speaker I think I was not able to explain some details like I wish I could. Basically, reflections influence sound for a certain duration but they can't make you forget how the direct sound sounded like. If the direct sound has balanced tonality, sum of reflections and direct sound of speakers may apply to source such a transfer function that the overall presentation will appear as *bright* to the listener, that being said, bright reflections won't make the listener forget about how balanced the direct sound was. Basically keeping on-axis response flat has higher importance than the factors we have discussed before. On-axis sound should not be modified to fix errors in reflections as long as the sound power response of speakers is not incredibly smooth.

I agree that flat DI speakers have advantages over conventional forward firing speakers. I totally agree with your points below:

The input signal is amplitude (voltage) changing over time. The output signal is amplitude (sound pressure) changing over time AND dispersed into three-dimensional space with a pattern that usually changes significantly with frequency. Seems to me that the LESS the dispersion pattern changes with frequency, the closer we are to your one-for-one transfer function target.

However, omni speakers sound better when they playback a music which was also mixed/mastered with omni speakers.

 

[Duke]

I am not a native english speaker

WOW! I never would have guessed. Very impressive!

Basically, reflections influence sound for a certain duration but they can't make you forget how the direct sound sounded like. If the direct sound has balanced tonality, sum of reflections and direct sound of speakers may apply to source such a transfer function that the overall presentation will appear as *bright* to the listener, that being said, bright reflections won't make the listener forget about how balanced the direct sound was.

I don't know about "brightness", but in my experience "harsh" reflections can have enough of an effect to ruin good-measuring direct sound.

Basically keeping on-axis response flat has higher importance than the factors we have discussed before. On-axis sound should not be modified to fix errors in reflections as long as the sound power response of speakers is not incredibly smooth.

In general I agree with this, but imo there can be exceptions.

Here is an example of "harsh" reflections that can ruin an otherwise good loudspeaker: Suppose we have a speaker with a 6-inch woofer and a 1-inch dome tweeter, crossed over at 2.5 kHz. The woofer's pattern will be maybe 90 degrees wide in the crossover region, while the tweeter's pattern will be 180 degrees, and would have been even wider if not for the baffle. So above the crossover region we have a LOT more off-axis energy than we have in the octave or so below the crossover region. This excess off-axis energy in the 2.5 - 5 kHz region is right where the ear is most sensitive, and it can make female vocals sound edgy and even harsh, and it can result in listening fatigue over time. If I were the designer, I would put a dip in the on-axis response at the bottom end of the tweeter's range. I'd try to make the dip big enough to eliminate that harshness without making it so big that the overtones seem subdued, so it will be a "balancing act" and probably won't be 100% successful. But a dip is much less audible (and objectionable) than a peak, so in my opinion this would be a situtation where a deliberate dip would be beneficial if done right.

However, omni speakers sound better when they playback a music which was also mixed/mastered with omni speakers.

I don't think anybody mixes on omni speakers. The recording engineer needs to very clearly hear what is on the recording so he will probably listen nearfield in a room with a much shorter reverberation time (especially at low frequencies) than our home listening rooms.

Imo the goal of home audio speakers is different. A home audio speaker should get two things right: The direct sound, and the reflected sound. Imo this is true regardless of whether the recording was mixed on Genelecs or Rockits or Altecs or B&W's or headphones or whatever. Get the direct sound right, and get the reflections right, and that's as good as it gets. But "getting the reflections right" includes interaction with the room and is a complicated subject!

 

[Recluse-Animator]

"Normal"..... such an interestingly loaded word...

Normal is a boxed speaker.
Nothing loaded about it.

 

[Recluse-Animator]

The problem with getting good imaging remains - the placement of designs like the Sanders and Ohm close to a wall is intended to provide a desired overall room voicing - a room Target Curve....

But what is not corrected in these approaches is the smearing of the imaging caused by the reflections from a nearby wall, arriving almost simultaneously with the initial speaker signal ... which confuses our psychoacoustic sense of direction on which we depend for imaging.

The only way to resolve that problem is to either absorb/diffuse the reflection (wall treatment) - or delay it (by moving the speaker further from the wall).

Sure, but having to place treatment on the wall behind a speaker is more convenient than having them far out in the room if you don't have a dedicated room.
Personally I would still put treatment on the wall even. Even if it was a box speaker I would put bass traps. You can never have too much bass traps.
Ohm doesn't recommend putting any treatment behind their speakers. Even if you have a window behind them they recommend putting it against the wall.

 

[Recluse-Animator]

The Sanders Sound flat electrostatic panels are highly directional both forwards and backwards, with the typical dipole null to either side. So it's often feasible to set them up such that the backwave energy's first bounce misses the "sweet spot", resulting in an abnormally long (and imo highly desirable) time delay before arriving at the ears.

Have you heard them? What's your opinion?
Both Sanders and all listeners that I've read say that they're the only hybrid ESL that has the woofer integrated correctly.
I have no way of auditioning them nor the money, but These would be the only dipole speakers I would purchase.
I don't think they're overpriced. You can get just the speakers and get your own amp, active crossover and room correction.

 

[Recluse-Animator]

Muraudio PX2 Omni ESL:

 

[Duke]

Have you heard them? What's your opinion?

I heard the Innersound Eros several times, and my understanding is that the current Sanders Sound Model 10 is the same concept just refined. It's one of the best small-sweet-spot speakers I know of, in particular with regard to its world-class imaging given good set-up. It will make you roll your eyes at claims of "electrostatic-like transparency" made by manufacturers of conventional speakers.

Both Sanders and all listeners that I've read say that they're the only hybrid ESL that has the woofer integrated correctly.

The ability to adjust the relative levels of panel and woofer is crucial to good integration for a hybrid electrostat like the Sanders Sound. You see, the woofer behaves like a point source (SPL falls off by 6 dB per doubling of distance) while the panel behaves like a line source (SPL falls off by only 3 dB per doubling of distance). Even with room reflections factored back in, the relative balance of woofer and panel CHANGES with listening distance! I've measured it (with a different hybrid electrostat) and in-room the woofer falls off relative to the panel by almost as much as the theory predicts at distances greater than two meters. (Closer than that and the change with distance is even less for the panel because it starts to approximate a planar source instead of a line source.)

I have no way of auditioning them nor the money, but These would be the only dipole speakers I would purchase.
I don't think they're overpriced. You can get just the speakers and get your own amp, active crossover and room correction.

Maybe keep your eyes open for a used pair? My impression is that they are unusually rugged so a used pair is likely to be in good shape, and the panels can be replaced/upgraded.

Muraudio PX2 Omni ESL:

The omnidirectional (in the horizontal plane) Muraudio is at the opposite end of the electrostatic spectrum from the Sanders Sound hybrids, unfortunately I've never heard a pair. I presume they'd have the same room-interaction challenges as most omnis, but with the inner detail and articulation potential of a good electrostat.

The electostats I sell fall in between those two radiation pattern extremes. SoundLabs have faceted-curved panels which radiate uniformly across either a 90 degree angle or a 45 degree angle, depending on the specifics.

 

[MattHooper]

Basically you can't make a speaker with bright on-axis sound balanced with recessed HF reflections. To our auditory system, direct sound is perceptually dominant. If direct sound sounds bright, you can't make it recessed with recessed reflections. That is not how it works. Our hearing is not linear like microphones have. Speakers with same in room response, may not sound same. That is why it is not recommended to EQ speakers above room transition frequency. But that is a topic for another thread.

Perhaps I'm missing something, but that doesn't seem to make sense given how altering room reflectivity clearly influences the sound.

I have my speakers in a room where I have some control of reflections (HF especially) by, for instance, adding diffusers or pulling thick velvet drapes along the side walls. From the listening position I can easily influence the apparent brightness of the sound by altering the reflections. Pulling the drapes along the side wall reflection point will alter the perceived balance to sound less bright "darker," less airy or strident.

Which is hardly news.

Can you explain what you mean by not being able to alter the apparent HF balance by reducing HF reflections, since that seems to contradict both theory and practice?

 

[Blockader]

Perhaps I'm missing something, but that doesn't seem to make sense given how altering room reflectivity clearly influences the sound.

I have my speakers in a room where I have some control of reflections (HF especially) by, for instance, adding diffusers or pulling thick velvet drapes along the side walls. From the listening position I can easily influence the apparent brightness of the sound by altering the reflections. Pulling the drapes along the side wall reflection point will alter the perceived balance to sound less bright "darker," less airy or strident.

Which is hardly news.

Can you explain what you mean by not being able to alter the apparent HF balance by reducing HF reflections, since that seems to contradict both theory and practice?

maybe you should read the whole conversation between me and Duke to understand what I was talking about.

Here is a brief summary:

you can influence overall tonality of the system with changing how the reflections are formed inside a room however our hearing is not linear like microphones have. To human hearing, direct sound is perceptually dominant. Speakers with bright direct sound will sound bright no matter how the reflections are formed inside the room. Speakers may sound less or more bright than the direct sound based on the tonality of sum of reflections, nevertheless they will sound bright if the direct sound is bright. As I said before, this is why EQing speakers above the room transition frequency is not recommended. Our hearing is not linear. Our hearing does not sum reflections with direct sound directly and process it as a single information. I recommend you to read the small chapter about precedence-haas effect in Toole's book.

EQ your speakers to have bright direct sound and change the form of reflections inside the room, see if you can make them sound recessed to you than before while facing your speakers.