Horn Speakers - Is it me or.......

[Jim Taylor]

The sad aspect: whenever I did such experiments, horn-loaded monitors were despised and ranked the worst by almost everyone, even by horn lovers among the engineers in blind tests. The only noteworthy exception was the Strauss main monitor which I have linked above.

(sigh) I understand, but it wasn't always that way .....

 

[Newman]

I didn't expect instant credibility. But trotting out the Toole trope of preference for flat frequency response from 200Hz to 20kHz fails in so many ways (and is really, really tired even for someone (me) who does use measurements).

Baseless claim #1.

1) We have learned so much since Toole's work.

Baseless claim #2.

Toole updates his book regularly. The last update was in 2025. He updates any research and knowledge that supersedes or adds to his work. That's how science works: it isn't out of date or wrong just because of time elapsed: it stands as evidence until better evidence requires it to be treated as out of date. If there were any better evidence then it will be in his latest edition. It isn't, not regarding some other FR from 200-20kHz than flat anechoic, nor its importance to sonic preference.

A frequency sweep in an anechoic or pseudo-anechoic environment or the full Spinorama captures single tone response.

Please don't trot out the 'that's not music' excuse. It doesn't hold up. If you are heading in that direction, I call Baseless Claim #3.

A speaker could meet the group-preference for this frequency response and still sound bad because of IM or multi-tone distortion, compression, or other reasons.

It's possible. But if the speaker is not being misapplied, eg a bookshelf speaker in a huge room with bass boosted to make 20 Hz flat, then it is unlikely to occur.

Toole presented the group-preference frequency response as a starting point. If he now presents it also as the ending point, then he has changed his tune (pun intended).

Baseless Claim #4. Straw man.

2) The data that Toole presents also shows a strong preference for more SPL below 200Hz (not flat).

I already covered that. What exactly did anyone say here that made you feel you need to make this point?

3) The group-preference at what SPL?

Low 80's dbA.

Go much higher and discomfort gets in the way of the assessment, which is not in any way meant to imply that some other FR would be preferred.

Go much lower and the ear's sensitivity curve gets in the way. Everyone knows about the 'loudness' button on the old amps since forever. Nobody denies their utility at low listening levels. Not Toole, not anyone. The exact compensation curve to apply for each listening SPL has attracted some debate, but if you think that's a refutation of the 200-20kHz flat anechoic finding, it simply isn't. It's a well-understood corollary that Toole and pretty much everyone covers. It certainly is not suggesting some other FR may be preferred at normal-loud listening levels nor 'proof' that Toole's work is outdated.

Toole strongly argues for tone controls so you can boost and cut based on preference, volume level, recording, etc.

He argued for them principally to allow some crude adjustment for poor mastering practices. If your speakers have great spinorama, and are being intelligently utilised eg bass correction, good room, etc, then the surprising (yet logical) uniformity of tonal preference (which you very likely share with the mastering engineer if you both have no major hearing damage) means that you are likely to be satisfied with his production, if and only if he has done a good job. The fact that he or she has an occupational hazard of hearing damage, and that listening panels aren't used to detect when this is a problem, means that mastering tonality issues are unfortunately not rare, and a tone control is better than nothing to deal with this. Ironically, so few of us use them, preferring to cast aside poorly mastered recordings and find good ones. That's what I do.

BTW, this is a strong argument why a fixed, locked in, 'boutique' speaker non-neutrality is a terrible idea.

4) An average of subjective ratings is not an objective rating. It is/was useful for designing a speaker for the average person. But it treats all other variance around that preference as error (error variance). There is more than hearing damage (Toole's consistent example) that has, as far as I know, not been explored (other than possibly by boutique speaker makers) that could account for this “error variance.”

Of course it has been explored. It just hasn't been found.

Of course, to the deniers, this is interpreted as 'it hasn't been explored enough', because they are so confident that it exists (based on sighted listening and sighted listening reviews) that the scientific evidence just can't be right.

This is why audiophiles are challenged by audio science. They don't understand the strength of the sighted listening effect, they don't understand the fallacious basis of all their hard-won beliefs about audio playback gear, and they are unwilling to let go of all that in the face of audio science, so they embark on the tried and trusted reaction, ie to shoot the messenger.

There may be (likely are) large sub-populations (groups of people) with consistent preference for other frequency responses. It seemed Toole and others publishing research have had no interest in understanding these sub-populations. That's fine when it isn't your research question, but it assumes every car buyer wants a Toyota Camry (and people then trotting out the flat response trope work from the position that anyone who doesn’t want a Camry because they’ve been told the Camry is the preferred car is an idiot).

The above paragraph is so uninformed that I seriously doubt that you even read Toole's book. If you did, it seems to have been a 'sceptical skim' intending only to agree with any science that confirms your prior views.

But back to my original points about trade-offs. As an example with horns, it is easier to make a more sensitive speaker with less expensive drivers (or at least it has been). Thus, a sizeable population might (does) exist where lower IM and multi-tone distortion, greater reverb, and high SPL are more preferred over the group-preference frequency response, issues with group delay, and other problems.

"Might (does)"? Oh, so you have evidence?

You are following the patterns of the average science denier. "All the things in the current science that I disagree with, are because the science hasn't yet discovered that I am right all along."

Knowing that many prefer more low frequency SPL, lower distortion, and a bit of reverb, a boutique speaker maker could easily justify designing and selling a horn speaker that does these things (all with really well done, private measurements). Do I want that speaker for home theater use or in a studio? Maybe not, but I, personally, much prefer it for my 2-channel system where I might be listening at low levels and from a central seat (boost the treble too while you're at it).

Like I said above, the low-SPL listening compensation curve varies with SPL, so locking one, fixed, (wrong at all other SPL) curve into a boutique speaker is a terrible idea and bound to fail to deliver satisfaction.

You are just making up excuses for boutique speaker makers. Excuses that are not even real in the minds of the boutique makers themselves, who are really just 'tuning by ear' using sighted listening and going down the imagined improvements spiral, and not understanding what they are doing wrong (if indeed they want to produce sound waves that would be preferred in a controlled test).

What's the motivation to share the measurements or send speakers to Erin or ASR for review for a boutique maker? If there is a strong bias toward the group preferred frequency response as the only way to properly design a speaker while ignoring the real markets that exist (and they do exist) for different curves then there's no way I'd want measurement absolutists dragging my name/brand through the mud (not even that Erin is an absolutist). I know if I were a boutique maker I couldn't wait for one of these reviewers to tell their audience of thousands that my product was dog turds (/sarcasm).

Were boutique speakers tuned by ear? Surely. With no measurements? Unlikely for many of the larger makers. Science is not just about "objective measurement" but also has room (large room) for subjective observations.

Another bogus claim. Over and over again, I and others are bringing to your attention that the science has been developed through starting with listening tests ie subjective observations. Then looking at the measurements vs the subjective observations for any correlation. Then develop a hypothesis. Then test the hypothesis by varying the measurement and tracking the subjective impressions to see if they track the hypothesis. If yes, then we have a finding and it is reported.

Yet you assert that there is this big gap in the science that it hasn't looked at enough, a lack of subjective observations? Come off it.

Seeing an anomaly is how many new discoveries are made and often without an objective measurement (induction, deduction, abduction can all occur from qualitative observation). The boutique makers have made these subjective observations and found sub-populations who agree (have a different shared group preference). Measurement absolutists can disagree and scream at the internet all they want. But these larger boutique makers know exactly that they are doing and they have measurements too. They have turned their subjective observation of an anomaly (their preference) into a repeatable (objective) product.

Guru Syndrome. Pure and simple. You are putting them on a pedestal.

As already stated numerous times, the reason audio science is coming up with findings that surprise, is because it is making subjective observations that are real, as opposed to imagined, by controlling experimental subjective observations for non-sonic influences. It is a difficult lesson for many to grasp, but once the lesson is absorbed, the audiophile can free himself from guru syndrome and start to see what has been happening in the hobby space over the years.

That is where ASR can help. But first, you have to absorb the lesson.

cheers

 

[snapsc]

Seeing an anomaly is how many new discoveries are made and often without an objective measurement (induction, deduction, abduction can all occur from qualitative observation). The boutique makers have made these subjective observations and found sub-populations who agree (have a different shared group preference). Measurement absolutists can disagree and scream at the internet all they want. But these larger boutique makers know exactly that they are doing and they have measurements too. They have turned their subjective observation of an anomaly (their preference) into a repeatable (objective) product.

Newman: Guru Syndrome. Pure and simple. You are putting them on a pedestal.​

I suspect that in audio, both of these statement are OFTEN true. A guy modifies his speaker...it sounds better. Next he builds his own speaker...it sounds even better. His buddies encourage him to start a company...now he is building 40-50 pairs a year...he has his own "house" sound and small following of people who like that sound...accurate or not. He earns a good living and a cult following develops around his sound and basically says measurements aren't everything.

Then we also have the "data driven" speaker producers..."Toole, Harman Curve, Klippel" strongly influence how they design and as a result, a lot of people like/prefer the sound of their products...this is the majority of the market...and probably where the market is headed overall.

Without the new discovery people, maybe we never end up with "crazy at the time new products like" electrostats....but without the data people, we probably never end up with "better" electrostats, horns that measure and sound better, purifi drivers, etc.

The hobby needs to embrace both types....and the buyer needs to understand the difference.

 

[Flaesh]

looks very much like an M2

This is exactly luxury M2. The source is visible in one photo - on-mag.fr

 

[mhardy6647]

(sigh) I understand, but it wasn't always that way .....


View attachment 514247


View attachment 514248


View attachment 514249


View attachment 514250

The Altec Fifteen was/is one of the better-kept secrets in vintage audio.

A very nice 12 inch "woofer" (more of a lower midrange, but a really sweet one) and an iconic (again and as usual bad pun intended) horn tracing its lineage back to Western Electric/ERPI

The Tale of Three Horns... Variations of the Western Electric 32A Horn

audio, tube, horn loudspeaker, high efficiency, loudspeaker, turntables, vinyl, field coil,

electravolt.blogspot.com

EDIT: Those others are nice, too -- although, with chagrin, I don't know/haven't heard that bottom one... but it sure looks like it's got it goin' on.

 

[Arindal]

I understand, but it wasn't always that way .....

Certainly true, such horn-loaded big monitors used to be popular in studios particularly in the 1970s and partly in the 1980s. Admittingly I have missed this era in studios, but older colleagues explained to me why these horns had been purchased in the first place and why most of studios suddenly have gotten rid of them eventually:

- the original idea of disco music, electronic pop, early hiphop and others genres was to mix ´for the club´, so a P.A.-like sounding system was closest to the target system
- the more these mixes were played on the radio, via ghettoblaster or in the car or on home systems, the more it became obvious they were not translating well to personal audio of any kind. Hence the ´compatibility check´ many studios asked mastering engineers to perform, like on a ghettoblaster, in-car stereo or alike
- many recording engineers over the years noticed that these systems offered very little subjective detail resolution and transparency unless being played at very high SPL, which was very fatiguing. In the era when CD was introduced, many mixing engineers were paying increasing attention to detail resolution and transparency, so they switched to monitors which would deliver just that at normal SPL

This is exactly luxury M2.

If it is identical in technical terms and sound quality, I have said everything.

Guru Syndrome. Pure and simple. You are putting them on a pedestal.

It is pretty ironic to read such statements from people who stick to findings postulated 35 years ago without isolating root phenomena of audible differences, with the underlying research in the meaning of independent, open experiments and discussion stopping, before speaker designers could keep up with delivering technically flawless products.

But first, you have to absorb the lesson.

Interestingly, I see myself among those who had been absorbing this lesson some 20+ years ago when I was professionally involved in blind listening tests. Ever since I did not find a reason to doubt what Toole et al. were publishing, they sounded logical to me back than and I applied them without much of doubt. Only later I happened to find out that they had just formulated very basic rules and did rather superficial testing with admittingly a lot of participants and effort.

When I did so almost 20 years ago, I found several aspects of reproduction quality which are easily identifiable in blind tests, but previously ignored by ´the science´ of the time. I am still grateful to have met a bunch of experts who were open-minded and showed me this ´world beyond the anechoic frequency graph´, first and foremost A. Jones of TAD Labs, KH Fink, D. Fricke (Ecouton), J. Siegler of KS Digital, S. Linkwitz, Mr. Kiesler (MEG) and many others.

Then we also have the "data driven" speaker producers..."Toole, Harman Curve, Klippel" strongly influence how they design and as a result, a lot of people like/prefer the sound of their products...this is the majority of the market...and probably where the market is headed overall.

Interestingly, I do not see it this way. I don´t see products being overly successful because of their sound quality which had been developed in a ´data driven´ manner. In contrary, I see a lot of products which are regularly meeting appraisal here for good measurements, but do sound extremely different when being listened to in direct comparison (Everyone who don´t believe me: Do an A/B listening test with a KEF Q11 Meta and Kii Three. It is eye-opening!). From the point of reproduction quality, it is simply impossible that both of these can be right at the same time. Those mass-market products which are really successful in terms of sold numbers, like portable bluetooth speakers, wireless headphones and party speakers, aber by no means ´data-driven´ when it comes to achieving sound quality IMHO.

That said, all of the aforementioned experts (at least those being into speaker dev) interestingly follow the same recipe: They do a lot of underlying research and use measurements a lot, but do the final tuning of their products by ear. I have so far not encountered a ´solely data driven´ product which would be superior to theirs.

 

[snapsc]

That said, all of the aforementioned experts (at least those being into speaker dev) interestingly follow the same recipe: They do a lot of underlying research and use measurements a lot, but do the final tuning of their products by ear. I have so far not encountered a ´solely data driven´ product which would be superior to theirs.

Arindal....I agree with you....No doubt....I wasn't very clear, yes, I believe that all commercial speakers developed through measurements get numerous listening tests along the way to determine the final sound they want to speaker to have. Also no doubt, the more comprehensive the measurements along the way, the more likely that issues can be resolved in advance.

 

[Deleted member 92621]

There's more to say but the real world work never stops.

To the OP, why did you like the horns you heard? Because they sounded damn good!

There are always, always trade-offs and you (like many, just check out the thread about the speakers people currently own) like the trade offs that were made.. Attack, dynamics, presence, and you liked it. Great! Me too. I was an engineer in a past career and I've never heard the sound of a single graph. And I am a specialist in human measurement these days. Measurements are wonderfully useful. But your ears matter too.

 

[Newman]

To the OP, why did you like the horns you heard? Because they sounded damn good!

??? The OP actually said that he has yet to hear a horn that he likes (a multi-way horn speaker, not just a waveguide on a tweeter).

Also note that the OP hasn't visited ASR for 5 years.

cheers

 

[Jim Taylor]

EDIT: Those others are nice, too -- although, with chagrin, I don't know/haven't heard that bottom one... but it sure looks like it's got it goin' on.

The bottom one is a modern Westlake speaker in their Reference series ... although for the life of me, I can't remember which one.