What are the Top Horn Speakers in the World Today? Vox Olympian vs Avantgarde Trio vs ???

[bonzo75]

How interesting, the AG Duo is also horn-loaded from around 170 Hz upwards.

That is well known. If you like the duo, check out hORNs symphony, you will get it used for 10kish. Not for me, but I can see why some people like it and the top half is better than the duo, and certainly not an unnecessary premium should you get a bigger room and more budget in the future for trios.

 

[bonzo75]

there are also commercial interests and what is available to buy and sell. These may factor in the decision process.

This is key, and is a leading cause of noob linearity.

 

[christensenleif@msn.com]

one illogical trait of the Vox Olympian is the varying dispersion patterns of the different horns

 

[Amir]

I’ve been following discussions across different forums for years, and I’ve seen your posts in the past, often completely contradictory to the positions you’re taking today. You can imagine I might be anyone with any system; that’s irrelevant.

I am sure who you are (Mr. Vahid Sharafi from Tehran) and you try to hide yourself and for this reason you do not share your playback picture.

WBF admin can tell me about your ip address.

You know me and you read my weblog (www.hifi.ir) for many years so you know I wrote about ASR/Vitus around 2007 and did not changed my opinions.

https://www.hifi.ir/2008/10/vitus-audio-in-iran/

Internet let you hide yourself but you can not hide the trust.

What truly matters is consistency. If someone once identified as a follower of the philosophies of people like Romy, Jim Smith, or David Karmeli, and spoke from that foundation, it doesn’t make sense to suddenly pivot and start making marketing-driven claims. Statements like “TAD delivers the true dynamics of horns” come across not as insight but as spin, and that undermines credibility. It would be like me, with firsthand experience of AG horns, later on insisting that Wilson Sasha speakers deliver the same level of dynamics and then stubbornly defending that claim without any solid foundation. That kind of argument doesn’t reflect genuine insight; it just undermines credibility.

As I said before I believe some cone speakers (TAD R1 , Gobel Majestic Wilson WAMM & Alexandria) are close to horns and Romy also believes old big wilsons are very close to horns dynamics.

If you read Romy post you find his opinion is against you.
Romy believes old big Wilsons are as dynamic as his 109db horn.
Read here : http://goodsoundclub.com/Forums/ShowPost.aspx?PageIndex=3&postID=21493


“the biggest Wilsons are the only one mid-sensitively direct radiation know to me speaker that do not compress sound (if driver with proper amplification). At least I, with my DSET driven, nearfiled operating, 109dB sensitive dynamic-spoiled listening habits recognized no dymick problems when I was listening Grand Slam or Alexandria. I have no idea how they did it….

Romy the Cat”

I believe some cone speakers are very dynamic and the “efficiency” is not necessary condition for “dynamics”. I believe not all horns are better than cone speakers, most modern horns like your favorite Avantgarde are very different sounding to vintage horns (like WE horns). If you think using tube amplification can solve the problem of Avantgarde horns then you are in the wrong way.

Not all horns are better than cones, not all cone speakers are dead sounding.

Maybe Romy never stated directly that high-order crossovers kill micro dynamics, but for exactly these obvious reasons he rejects them. The more complex the filter, the more the music has to pass through extra components, each introducing phase shift, time smear, and energy loss. This inevitably blurs micro-dynamics and weakens note decay. That’s why many designers of high efficiency systems, Romy included, lean toward simpler networks: fewer parts in the path, more immediacy and natural flow.

Like a internet robot You just copy/paste and share theories that are popular but meaningless.

I know the meaning of both micro and macro dynamics, which is precisely why I would never place cone-based speakers like my Sasha on the same level as horn loudspeakers with explosive dynamics, lol.

I listened to Avantgarde in the US back when Mike was their representative. Unfortunately, I didn’t take any photos to share with you.

You do not know.

Mike (Audio federation) was never distributor of Avantgarde. Mike was distributor of Acapella

 

[DasguteOhr]

Agreed, with large multi-way horn systems you won’t nail perfect alignment by eye. Physical geometry gets you close, but the precise alignment is finished with microphone-based impulse measurements. The silver timing stripe on some JBL/TAD compression drivers is a useful mechanical reference, yet because the acoustic center shifts with horn loading it can only take you so far. You still need the mic to dial in the last few millimeters. The best practice is mechanical alignment first, and then finalizing the tiniest delays with microphone measurements.

With REW, you can measure group delay (tone burst). You need a microphone with a loop-back function. This means the sound card's output signal is fed back in as the input signal (two-channel measurement). What goes to the amplifier and what is measured at the microphone has no latency. This prevents different latencies from distorting the correct measurement.

 

[DasguteOhr]

My speakers are fully horn loaded but only go to mid 30s Hz. They are big but fit tightly into the corners, so they do not take up much space.

I agree. 30 Hz is no problem in many rooms. Unless you listen loudly and have a lot of large bass drivers that overwhelm the room with sound pressure.

 

[Hornsolutions Swen]

The Vox Olympian is a 4-way horn-loaded satellite: midbass horn, midrange horn, high-frequency horn, and a super tweeter horn. When paired with the optional Vox Elysian sub-bass units (each with dual 12" drivers front-loaded by a 4.6 m folded exponential horn), the system becomes effectively a 5-way all-horn design. By contrast, the Trio G3 is a 3-way horn satellite, with optional bass modules that can be horn-loaded (Basshorns/SpaceHorns) or conventional (Sub 231). The claim that the Olympian has “only two horn-loaded ways” is incorrect, and it’s equally misleading to imply the Trio always uses “conventional” bass , it depends entirely on the bass module selected.

Hello Kozak, very interesting.
To be honest, I didn’t know that a 460 cm folded ‘horn’ is being used here.


I have a few comments about that

In acoustic horn design, the geometry of the flare (the way the cross-section expands along the horn axis) is critical for predictable sound radiation and efficiency.
Exponential law of area growth. In a true exponential horn, the cross-sectional area follows
S(x) = S_0 \, e^{mx}
where S_0 is the throat area, m the flare constant, and x the distance along the horn length.
For a horn of 460 cm length, the area expansion is continuous and smooth from throat to mouth. This ensures that the acoustic impedance gradually transforms from the driver to free air.
Impedance matching and cutoff frequency.The exponential profile produces a well-defined cutoff frequency f_c = \frac{m c}{2\pi}, below which the horn loses efficiency.
A straight-line approximation (polygonal flare) introduces discontinuities in the rate of expansion. These act as acoustic mismatches, generating reflections inside the horn. As a result, the effective cutoff becomes less precise and resonances may occur.
Reduction of internal reflections. Every sharp change in flare angle (as occurs when substituting curved walls with flat boards) creates partial reflections.
In a 460 cm long horn, even small reflections accumulate, producing standing waves and colorations in the mid-bass and lower midrange. The exponential curve minimizes this problem by ensuring continuous impedance transition.
Frequency response and directivity
A true exponential curve yields a smoother low-frequency roll-off and a more predictable radiation pattern.
Straight-line segments approximate the exponential law only roughly; this causes ripples in frequency response and less controlled directivity, especially around the horn’s cutoff region (which, for a 460 cm horn, is typically in the range of ~18–25 Hz depending on flare constant).

Perceptual effect
To the listener, a horn with the correct exponential contour delivers cleaner bass, reduced coloration, and higher efficiency.
A straight-panel approximation may still produce strong output, but with audible irregularities: boominess, uneven tonal balance, or “honkiness.”
In summary: A true exponential 460 cm horn ensures smooth acoustic impedance transformation, minimizes reflections, and produces a cleaner, more natural sound compared to a horn that only imitates the exponential shape with flat, straight segments.

In light of these facts, the sonic result that is delivered should be rated even higher.

 

[Hornsolutions Swen]

Perfect time alignment in horn systems is, of course, only achievable through impulse measurements. Without addressing group delay, it simply doesn’t work. But even if this is done, the alignment changes again as soon as the listening position shifts. For example, someone sitting just 10 cm higher than another listener will automatically experience a different time alignment.
That said, this effect only really becomes significant at crossover frequencies above 10 kHz, as the wavelength gets shorter and shorter. Alternatively, the “sweet spot” can be enlarged through a three-dimensional calibration, which helps reduce the negative influences.


Or, one can simply enjoy the music and give a shit about the perfect phase.

 

[PeterA]

Perfect time alignment in horn systems is, of course, only achievable through impulse measurements. Without addressing group delay, it simply doesn’t work. But even if this is done, the alignment changes again as soon as the listening position shifts. For example, someone sitting just 10 cm higher than another listener will automatically experience a different time alignment.
That said, this effect only really becomes significant at crossover frequencies above 10 kHz, as the wavelength gets shorter and shorter. Alternatively, the “sweet spot” can be enlarged through a three-dimensional calibration, which helps reduce the negative influences.


Or, one can simply enjoy the music and give a shit about the perfect phase.

Is this why people meet with varying levels of success when they add super tweeters to speakers like the JBL Hartfield or Vitavox CN-191?

 

[Hornsolutions Swen]

Is this why people meet with varying levels of success when they add super tweeters to speakers like the JBL Hartfield or Vitavox CN-191?

Exactly. At 10 kHz, the wavelength is already down to 34.3 mm.

At 15 kHz, it shrinks even further to just 22.9 mm.

The shorter the wavelength, the more difficult it becomes to align the super tweeter perfectly with the main speaker. In theory, the solution is straightforward: generate a test tone and move the super tweeter forward or backward until the phase matches precisely.

In the audiophile world, however, things are often done differently. Proper measurements and test equipment are sometimes regarded with suspicion. Instead, people simply solder a capacitor in series and position the super tweeter “by ear.” At such high frequencies, this is virtually impossible — even a tiny misalignment of just a few millimeters can cause severe phase cancellations or lobing effects.

This is why professional loudspeaker designers rely on careful acoustic measurements, time-alignment techniques, and often digital tools such as FIR filters to integrate a super tweeter seamlessly. Without these measures, adding an extra high-frequency driver can do more harm than good, leading to comb filtering, poor imaging, and a less natural soundstage.

Best Regards S

 

[christensenleif@msn.com]

in my opinion a 5 way solution, creates more problems than it solves

 

[Hornsolutions Swen]

in my opinion a 5 way solution, creates more problems than it solves

I completely agree. Every additional crossover in passive filter designs causes more problems than the extended signal path provides solutions or benefits. I have always been a fan of three-way systems, where the first crossover is placed below the critical frequency range (160/170Hz-680/700Hz) , around 160 to 170 Hz. The second crossover should be slightly above that, in the range of about 1 kHz to a maximum of 1.8 kHz. Above this range begins a frequency band where human hearing is particularly sensitive, namely between approximately 2.5 kHz and 3.5 kHz.

Best Regards S

 

[Al M.]

in my opinion a 5 way solution, creates more problems than it solves

Simpler is often, though not aways, better.

My speakers are multi-driver but two-way (quasi line-array).

 

[Salectric]

Exactly. At 10 kHz, the wavelength is already down to 34.3 mm.

At 15 kHz, it shrinks even further to just 22.9 mm.

The shorter the wavelength, the more difficult it becomes to align the super tweeter perfectly with the main speaker. In theory, the solution is straightforward: generate a test tone and move the super tweeter forward or backward until the phase matches precisely.

In the audiophile world, however, things are often done differently. Proper measurements and test equipment are sometimes regarded with suspicion. Instead, people simply solder a capacitor in series and position the super tweeter “by ear.” At such high frequencies, this is virtually impossible — even a tiny misalignment of just a few millimeters can cause severe phase cancellations or lobing effects.

This is why professional loudspeaker designers rely on careful acoustic measurements, time-alignment techniques, and often digital tools such as FIR filters to integrate a super tweeter seamlessly. Without these measures, adding an extra high-frequency driver can do more harm than good, leading to comb filtering, poor imaging, and a less natural soundstage.

Best Regards S

Swen,
I had not heard of the Caeles turntable before seeing your post. That is one very impressive piece of hardware. I can only imagine how good it must sound.

 

[Audiohertz2]

Exactly. At 10 kHz, the wavelength is already down to 34.3 mm.

At 15 kHz, it shrinks even further to just 22.9 mm.

The shorter the wavelength, the more difficult it becomes to align the super tweeter perfectly with the main speaker. In theory, the solution is straightforward: generate a test tone and move the super tweeter forward or backward until the phase matches precisely.

In the audiophile world, however, things are often done differently. Proper measurements and test equipment are sometimes regarded with suspicion. Instead, people simply solder a capacitor in series and position the super tweeter “by ear.” At such high frequencies, this is virtually impossible — even a tiny misalignment of just a few millimeters can cause severe phase cancellations or lobing effects.

This is why professional loudspeaker designers rely on careful acoustic measurements, time-alignment techniques, and often digital tools such as FIR filters to integrate a super tweeter seamlessly. Without these measures, adding an extra high-frequency driver can do more harm than good, leading to comb filtering, poor imaging, and a less natural soundstage.

Best Regards S

Additionally :

There’s not much energy above 12K and you need to be directly on tweeter axis, so location is critical as well as Matching gain , this is were the non measurement group run into trouble IMO , the lack of energy at those frequencies means they are running the super tweeter gain wide open to hear a difference.

 

[DasguteOhr]

Above this range begins a frequency band where human hearing is particularly sensitive, namely between approximately 2.5 kHz and 3.5 kHz.

Best Regards S

Calculated filters are often used to create a dip in the range (separating the midrange and tweeter). This allows for long-term listening without becoming tiring. A flat frequency response creates a fresh, detailed sound, but can be extremely annoying if you switch to another devices.my opion

P.S
then you often hear the statements e.g. the CD player sounds aggressive but real is it the frequency response of the loudspeaker

 

[Al M.]

P.S
then you often hear the statements e.g. the CD player sounds aggressive but real is it the frequency response of the loudspeaker

And/or the room. Room distortions can be horrible and create an effect of harshness and hardness (they also sometimes literally can sound like amplifier clipping). In my experience, room distortions are a huge and vastly underestimated problem, and a major impediment to both enjoyable and realistic sound. Naturally, the problem is greatest at higher SPL.

 

[Hornsolutions Swen]

Calculated filters are often used to create a dip in the range (separating the midrange and tweeter). This allows for long-term listening without becoming tiring. A flat frequency response creates a fresh, detailed sound, but can be extremely annoying if you switch to another devices.my opion

One of the main reasons why one loudspeaker feels fatiguing after 60 minutes while another still sounds relaxed after six hours is, imo relatively straightforward. The human brain must assemble acoustic information, as mixed by the sound engineer, into a coherent soundstage in fractions of a second. This soundstage includes the spatial arrangement of instruments, such as drums behind the singer, bass on the left, or backing vocals on the right. This reconstruction becomes increasingly difficult the more reflections occur from walls, ceilings, or floors.

The brain’s task is significantly easier when fewer reflections are present or – to return to the topic of horns – when the proportion of direct sound increases. Conventional loudspeakers typically have a direct sound share of about 2-3%, while electrostatic speakers achieve only around 0.6 to 1.5% under usual room conditions.
With a direct sound share of approximately 30% or more, as found in perfectly tuned horn loudspeaker desigs, the brain has less work to do in processing the soundstage. As a result, the listener perceives the sound as more pleasant, especially during extended listening sessions. This is one of the main reasons why people who have experienced a perfectly tuned horn loudspeaker system often describe its sound as having a unique sense of ease and clarity.

Of course, the reasons you mentioned are also entirely correct.“

 

[Amir]

And/or the room. Room distortions can be horrible and create an effect of harshness and hardness. In my experience, room distortions are a huge and vastly underestimated problem, and a major impediment to both enjoyable and realistic sound. Naturally, the problem is greatest at higher SPL.

Exactly, those who think 6db crossover is better than 24db never understand the main distortion of sound is speaker room interaction.

When Styrling Trayle properly positioned the speakers they forget the crossover was 6db or 24db.

There are many non-sense debates in this forum about devices specifications. Horn vs cone, 6db vs higher order, …
The main problem of every audio playback is ac power quality, speaker position (room speaker interaction) , amplifier speaker matching and quality of source (both records and transport/turntable) but those are judging speakers by the crossover slop LOL

Gryphon is the king of good specification, sealed box, linear phase crossover, class A amplifier. Zero feedback, … line array bass and midrange, dapolito … but the sound is dead.

Those specifications all are marketing, listen without any pre judgments in proper controlled condition

 

[Amir]

The reason why one loudspeaker feels fatiguing after 60 minutes while another still sounds relaxed after six hours is relatively straightforward. The human brain must assemble acoustic information, as mixed by the sound engineer, into a coherent soundstage in fractions of a second. This soundstage includes the spatial arrangement of instruments, such as drums behind the singer, bass on the left, or backing vocals on the right. This reconstruction becomes increasingly difficult the more reflections occur from walls, ceilings, or floors.

The brain’s task is significantly easier when fewer reflections are present or – to return to the topic of horns – when the proportion of direct sound increases. Conventional loudspeakers typically have a direct sound share of about 2-3%, while electrostatic speakers achieve only around 0.6 to 1.5% under usual room conditions.
With a direct sound share of approximately 30% or more, as found in perfectly tuned horn loudspeaker desigs, the brain has less work to do in processing the soundstage. As a result, the listener perceives the sound as more pleasant, especially during extended listening sessions. This is one of the main reasons why people who have experienced a perfectly tuned horn loudspeaker system often describe its sound as having a unique sense of ease and clarity.

Of course, the reasons you mentioned are also entirely correct.“

Total wrong

Nobody in this world can speak about relation of specifications and our brain system fatigue