• Welcome to ASR. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

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

The third inverted horn looks like an odd transmission line to me,I haven't seen the insides though,it may well be a simple horn.
It's just a bass reflex box.

Sorry, that's already been discussed :) By the way, these speakers sound good. And can be a good reference for a DIY project.
 
It's just a bass reflex box.

Sorry, that's already been discussed :) By the way, these speakers sound good. And can be a good reference for a DIY project.
I searched and asked around in Munich show this year but it seems they weren't there.
Friends say they are pretty decent.We'll see :)
 
The in your face sound demanding attention is likely related to much direct sound.

However, this can be changed with the use of phase grating diffusion. As always: There's a big difference between diffusers, so anything called "a diffuser" might not work well.
That’s the way I understand it, yes, appreciate the input. Not sure if it’s related, but looking at non-smoothed measurements of a CD horn vs some Harbeth speakers I had before in the same room, there is more „grass“ in the top end with the horn.

Re: diffusion - I know that it’s very hard to make generalisations without knowledge of the room / speakers, but what type of diffusor would be suitable to this application & are we talking about diffusion at first reflection points or diffusion closer to MLP? Just curious, I’ve come to enjoy the prevalence of direct sound.

Btw - just for continuity’s sake - there was a post I made above asking about modifications to the crossover of my JBL S3100 that were possibly made by the previous owner, and the funky in-room FR that they produced. The reader may please disregard these measurements entirely, as the crossover was, in fact, assembled incorrectly. It has since been fixed, and the speakers produce a much cleaner, curve. Thread to follow
 
Imo the characteristic of a waveguide which warrants use of that specific term IS its approximately constant directivity over most of its bandwidth. In practice the directivity widens at the low end and narrows at the top end, so it's not perfect, but it's arguably quite useful.

Hello Duke

So out of curiosity how would you class the original devices developed by Db Keele at EV back in 1975? This was long before the term waveguide was coined. If you look through the attached paper the directivity was quite good! Then he moved on to JBL to develop the 2342, 2344 and 2360 family. Are these waveguides? Are these a stepping stone towards them?

Rob :)
 

Attachments

In the image below the Oblate Spheroid is the light purple curve, and the mouth round-over is not shown. The Conical, Oblate Spheroid, and Hughes (Peavy) profiles are constant-directivity types, and as you can see they have most of their curvature down near the throat.

Contours4.jpg

My understanding is that the Oblate Spheroid curvature is the one which imposes the least disturbance to the wavefront for a given change in angle, so imo it's arguably the "optimum" for a constant-directivity application.

Nice post Duke, thx.
I've studied that graph and read a lot of horn theory many quite a bit.
The conicals I build (straight purple traces above) , are true constant directivity, but suffer from compression driver to throat transition, and mouth termination.
For the CD-to-throat matching, I've tried the Charlie Hughes throat adaptation. It works, but I don't think it makes enough of a measured or audible difference to fool with, vs a more relaxed bolt in on type fit.
Mouth termination matters more I think. There I've tried secondary horn flares attached to the primary conical ala Keele, and also secondary tractrix shaped flares ala the Klipsch K-402 horn. Both do a good job of smoothing out polar anomalies caused by the poor mouth termination. I've also tried Hughes 'felt around the mouth' approach, and different degrees of simple mouth round overs.

My understanding of OS waveguides appears to match yours. I think its elegant property is how it diffuses internal reflections and diffraction in the most benign way possible.
Someday I hope to use somebody's Mabat /ATH inspired horn (which I think follow along the OS goal path), and use as as the platform to build a synergy. Best of all worlds I think, with the only tradeoff being ports in the horn interfering with wavefronts. From my experience with simple conical MEHs, I feel certain that tradeoff is well worth it.

And beating my own dead horse here, imo "approximately constant beamwidth over most of its passband" [without the use of diffractive features or vanes] is THE defining property of a waveguide.
Yep, my dead horse too :)
Hey, I figure we can each have our differing viewpoint, and all is just fine.

Although @gnarly also said he has also seen very few CD horns, which surprises me. Whenever I browse a horn catalog, it is dominated by CD horns with their specified '80°H x 60°V' (or whatever) directivity objective. They're everywhere. For example, link.

"Constant directivity" has almost been renamed "so-called constant directivity" by many members of the speaker building press

If the horn or waveguide dimensions don't at least meet Keele's formula for lowest frequency specified, you smell bogus from the gitgo. (formula posted earlier)
Next thing to look at is does that lowest frequency cover the range of intended operation of the driver planned to go with it.
Last thing is to build a proto, and curse cause it doesn't hold pattern as low as hoped ...Lol
 
Re: diffusion - I know that it’s very hard to make generalisations without knowledge of the room / speakers, but what type of diffusor would be suitable to this application & are we talking about diffusion at first reflection points or diffusion closer to MLP? Just curious, I’ve come to enjoy the prevalence of direct sound.
Depends on the distance you have. Diffusion over a wider frequency area needs some distance. If you have enough width, diffusion of opposite side wall reflections will give the most pronounced effect.
The rear wall behind the listening position is also a great place for diffusion if there's enough distance.

The RPG Modffractal is the best I've tested. You need a diffuser that diffuses high enough in frequency. Most stop around 4-5 KHz and that's too low.
 
Hello Duke

So out of curiosity how would you class the original devices developed by Db Keele at EV back in 1975? This was long before the term waveguide was coined. If you look through the attached paper the directivity was quite good! Then he moved on to JBL to develop the 2342, 2344 and 2360 family. Are these waveguides? Are these a stepping stone towards them?

Rob :)

I did not read the paper. At first glance it looks to me like he is joining together different horn shapes in order to get the radiation pattern shape he wants in both planes.

I would probably call them constant-beamwidth or constant-directivity horns.

Where the change in the angle of the walls happens reflections will occur, same thing at the fairly abrupt mouth terminations. Some of the reflection energy will propagate back down the horn all the way to the diaphragm and reflect back out again, partially ping-ponging back and forth and decaying as it does so. Imo this is a source of "horn honk".

And I can see how these horns could conceptually be stepping-stones towards modern waveguides, though I do not know what role they may have played in Earl Geddes' thinking on the subject. Earl was using JBL 4430 studio monitors (a landmark design itself imo) when I first met him, and he told me that for about ten years he had been wanting to build the speaker that became the Summa.

I think the day will come when the Summa is appreciated as being a point of inflection in loudspeaker design.
 
Last edited:
Depends on the distance you have. Diffusion over a wider frequency area needs some distance. If you have enough width, diffusion of opposite side wall reflections will give the most pronounced effect.
The rear wall behind the listening position is also a great place for diffusion if there's enough distance.

The RPG Modffractal is the best I've tested. You need a diffuser that diffuses high enough in frequency. Most stop around 4-5 KHz and that's too low.
Thank you - I might give that a try one day. Distances might be a problem, though, what with everything being set up in an apartment and all. 2.5m space behind the MLP at best
 
The conicals I build (straight purple traces above) , are true constant directivity, but suffer from compression driver to throat transition, and mouth termination.
For the CD-to-throat matching, I've tried the Charlie Hughes throat adaptation. It works, but I don't think it makes enough of a measured or audible difference to fool with, vs a more relaxed bolt in on type fit.
Mouth termination matters more I think. There I've tried secondary horn flares attached to the primary conical ala Keele, and also secondary tractrix shaped flares ala the Klipsch K-402 horn. Both do a good job of smoothing out polar anomalies caused by the poor mouth termination. I've also tried Hughes 'felt around the mouth' approach, and different degrees of simple mouth round overs.

I didn't realize you were designing and building and optimizing your own horns! That is indeed... what's the word I'm looking for... oh here we go: "gnarly". I tip my virtual hat to you!

Someday I hope to use somebody's Mabat /ATH inspired horn (which I think follow along the OS goal path), and use as as the platform to build a synergy. Best of all worlds I think, with the only tradeoff being ports in the horn interfering with wavefronts. From my experience with simple conical MEHs, I feel certain that tradeoff is well worth it.

That very seriously might be the ultimate! As I'm sure you know.

I have lately been using Marcel Batik's math with his permission, as his math optimally smooths the transition from the virtually-straight sidewall to the mouth round-over such that the rate of change in the curvature is very gentle, to minimize the mouth reflection.

The profile Marcel prefers is not quite an Oblate Spheroid, as (if I understand correctly) he's found a variation which optimizes for a bit better frequency response smoothness. I prefer ye olde Oblate Spheroid because it has the best match between the on-axis and off-axis curves, and I don't mind a bit more complexity in the crossover filtering. Last weekend I measured a 12", 70-degree-pattern Oblate Spheroid which used Marcel's math and it had the best match between the on-axis and off-axis curves of anything I have ever measured.

"Constant directivity" has almost been renamed "so-called constant directivity" by many members of the speaker building press

Yes I try to remember to use qualifiers because our beloved "constant directivity" waveguides are not quite so, but arguably close enough to be useful.
 
Last edited:
That very seriously might be the ultimate! As I'm sure you know.

I have lately been using Marcel Batik's math with his permission, as his math optimally smooths the transition from the virtually-straight sidewall to the mouth round-over such that the rate of change in the curvature is very gentle, to minimize the mouth reflection.

The profile Marcel prefers is not quite an Oblate Spheroid, as (if I understand correctly) he's found a variation which optimizes for a bit better frequency response smoothness. I prefer ye olde Oblate Spheroid because it has the best match between the on-axis and off-axis curves, and I don't mind a bit more complexity in the crossover filtering. Last weekend I measured a 12", 70-degree-pattern Oblate Spheroid which used Marcel's math and it had the best match between the on-axis and off-axis curves of anything I have ever measured.
Awesome Duke ! it's great to hear you are employing Marcel's work.
Do you have a 3D printer? I don't have one, but may get one soon, especially since I can't get good Baltic (russian) birch anymore to make conicals.
 
...Someday I hope to use somebody's Mabat /ATH inspired horn (which I think follow along the OS goal path), and use as as the platform to build a synergy. Best of all worlds I think
...I have lately been using Marcel Batik's math with his permission, as his math optimally smooths the transition from the virtually-straight sidewall to the mouth round-over such that the rate of change in the curvature is very gentle, to minimize the mouth reflection.
...Awesome Duke ! it's great to hear you are employing Marcel's work.
I mentioned upthread that a commercial product is in the wind that brings together Marcel's and Earl's work.

You won't see any MEH tech in it though. Dr Geddes does not hold it in high regard, and very much does not see it as the 'best of all worlds'. Not for home audio.

cheers
 
Awesome Duke ! it's great to hear you are employing Marcel's work.

Earl recommended Marcel's work to me.

Do you have a 3D printer? I don't have one, but may get one soon, especially since I can't get good Baltic (russian) birch anymore to make conicals.

No, I was involved in the design phase, and I'm doing the crossover so I made the measurements. The CAD work was done by someone with expertise in multiple areas, he also oversaw the manufacturing of the waveguide and did the final assembly of the parts. The 3D print was done in powdered carbon fiber using a process called "laser sintering".

I can put you in touch with him if you'd like, but just so you know, this was not an inexpensive project.
 
No, I was involved in the design phase, and I'm doing the crossover so I made the measurements. The CAD work was done by someone with expertise in multiple areas, he also oversaw the manufacturing of the waveguide and did the final assembly of the parts. The 3D print was done in powdered carbon fiber using a process called "laser sintering".

I can put you in touch with him if you'd like, but just so you know, this was not an inexpensive project.

Thanks for that kind offer.

I'm pretty sure if I go the printed horn route that I'll want to get a printer and wade in to it in DIY fashion. I truly enjoy every step in the DIY process ... the design, building protos (over and over haha), measuring, and implement processing/xovers etc.
Building the final proto version for stereo or LCR, often feels anticlimactic, and seems more like work.


You won't see any MEH tech in it though. Dr Geddes does not hold it in high regard, and very much does not see it as the 'best of all worlds'. Not for home audio.
That's a pity. I'd love to see his audio intellect help refine the synergy concept into a better horn/waveguide than conical.. (Perhaps though, since commercial interests might be in play with anything he works on, patent infringement could be a show stopper from the gitgo...wild guess.)

Doesn't really matter though, as too many people have heard synergy designs, be they Danley's commercial boxes, or DIY one-offs. There's an ever increasing crowd that realizes MEHs are not just the province of PA, but may in fact be the best full-range co-axial implementation currently known.
A lot of really smart folks are trying to use advanced waveguides for MEHs........for home use :) Folks who have already mastered good designs similar to the Summa etc..

Personally, I think MEHs need to be big to reap their main benefit...full-range integration that extends lowest frequency of pattern control.
So far all the "advanced waveguides used for MEHS" designs I've seen have been relatively small. I figure due to 3D printer size constraints and costs of large prints.
I'm not so sure this has any advantage over a well done co-axial like from KEF, or good waveguide-over-cone builds like from Genelec, Neumann, Geddes, etc.
I think every driver section above the subwoofer handoff needs to be incorporated into a MEH for best of all worlds results.

Anyway that's just me,.... my thoughts / speaker preference ....which means no more than Geddes, or yours, or any other person's does.......
.......or the even general population's mean preference (ala Harman).....
 
As CD horns have evolved , how to select one to build a two way system using drivers such as these?

My a mostly empty room is ~ 23 ft x 15. Triangular ceiling; peaks at 11 ft.

My goals are very low distortion, zero horn coloration, great spaciousness but still with very good on axis response, accurate musical timbre excellent sound stage size.

Some current offerings.
such as,

 
My goals are very low distortion, zero horn coloration, great spaciousness but still with very good on axis response, accurate musical timbre excellent sound stage size.

Ime there is at least one tradeoff to be juggled within your set of goals: Soundstage width is enhanced by strong early same-side-wall reflections, BUT ime those same early reflections detract from "great" spaciousness, and can degrade sound image precision (which you didn't mention but which might be a consideration).

The lowest-coloration horn profile I have worked with is the Oblate Spheroid, as advocated by Earl Geddes, and more specifically when using the mouth geometry math developed by Marcel Batik. That being said, I have not worked with either of the audiohorn profiles you linked to. But my understanding is that, mathematically, the Oblate Spheroid introduces the least amount of disturbance to the wavefront for a given coverage angle, which is why Earl chose it.

Imo there is merit in matching the exit angle of the compression driver to the entry angle of the horn, and if you can't get an exact match, then get as close as you reasonably can.
 
Thanks for your reply.

Soundstage width is enhanced by strong early same-side-wall reflections, BUT ime those same early reflections detract from "great" spaciousness, and can degrade sound image precision (which you didn't mention but which might be a consideration).

Indeed Weltersys (Art Welter) and others here has made me ever mindful of the tradeoff between pursuing use of a CD horn to achieve spaciousness at the expense of smeared or otherwise loss of accurate vocal and instrument imaging at threads like these

I certainly did forget to include accurate imaging as an essential goal.

But cannot a combination of room/system measurirng software (e.g. REW) to select room acoustically modifying materials and the use of "convolving" software (e.g. DIRAC Live) mitigate those destructive reflections?

Regarding the importance of matching the driver's exit angle with the horn's throat geometries, yes, Marco_gea and Docali at diyaudio.com have cautioned me to always avoid adapters where possible. For example, while Docali designed an adapter for Pierre's to fit his 1.4" drivers https://www.usspeaker.com/radian 745neoBepb-1.htm to his 2" horns https://www.athosaudio.com/2021/01/01/tad-th-4001/ https://www.diyaudio.com/community/threads/jmlc-and-yuichi-horns-measurements.395046/page-12
resonances and other anomalies can be created, as it's always an approximation. Better to use a 2" driver with a 2" horn.


Are one of these horns very close to the horn you describe?


Presumably, the horn you and Earl seem to overall prefer had performed better than Troy Crowe's attempt.

Is it generally true that exponential horns like Troy Crowe's https://josephcrowe.com/products/3d-cad-files-horn-no-1994-es450-biradial-for-jbl-2446-2-throat
are among the most popular because they are far less prone to coloration or "honking", or are those problems just as often due to poor crossover design?

And, if quite ironically, can it also be true-and which i think might have been claimed by Bjorn Kolbrek -that even horns like the JMLC AH425, if properly placed can produce a convincingly wide and deep sound stage?
 
... cannot a combination of room/system measurirng software (e.g. REW) to select room acoustically modifying materials and the use of "convolving" software (e.g. DIRAC Live) mitigate those destructive reflections?

I don't know.

My instinct is that it's a good idea to get the native acoustic behavior as correct as possible before using DSP to correct for reflections. And I'm skeptical of the benefits of having the reflection-correcting transfer function imposed on the direct sound and on all other reflections in addition to the targeted one(s).

Are one of these horns very close to the horn you describe?


That audiohorn Oblate Spheroid looks very good to me. I haven't tried the elliptical variation on the theme shown at the link. Earl tried an elliptical Oblate Spheroid but did not go into production with it, presumably preferring the round version even after going to the expense of having the mold made. The elliptical will lose pattern control in the vertical at a higher frequency than in the horizontal but will allow closer driver spacing, so it's a juggling of tradeoffs.

Presumably, the horn you and Earl seem to overall prefer had performed better than Troy Crowe's attempt.

Is it generally true that exponential horns like Troy Crowe's https://josephcrowe.com/products/3d-cad-files-horn-no-1994-es450-biradial-for-jbl-2446-2-throat
are among the most popular because they are far less prone to coloration or "honking", or are those problems just as often due to poor crossover design?

Well I've gone in a different direction but have great respect for Troy Crowe.

And, if quite ironically, can it also be true-and which i think might have been claimed by Bjorn Kolbrek -that even horns like the JMLC AH425, if properly placed can produce a convincingly wide and deep sound stage?

I would be inclined to trust Bjorn's educated and experienced observations and analysis.
 
Last edited:
The combination of a good horn and spaciousness requires high quality diffusion. One could also consider a horn with somewhat wider directivity and diffusion on oppoiste side wall reflections. Depends on the rooms size, what's practical possible, etc. as well.

All horns designs have their positive and negative aspects. While I haven't testet everything, my experience is that that a good designed diffractional horn is perhaps the best compromise if it's large enough and with DSP. The reason is that you can achieve constant directivity in both planes and the minor peaks and dips are minimum phase behaviour, meaning EQ will work perfectly.
While there are other horns that can achieve a flatter raw response, they will not be broadband constant in both planes.
 
My living room is ~ 23 ft L x 15 W. Opposite the ~23 ft east wall, about 3/4 of the opposite side empties into a ~ 10 ft x 9 kitchen and a 10 ft L x 4 ft w stair case, which turns right and continues ~ 15 ft down to ground floor. There is a triangular ceiling which peaks at 11 ft extending over the living room, kitchen and staircase. Three ~ 40 inch x 30 inch bay windows at one end of the north wall in living room; so that wall is most trigonal shaped. Most of my living room is empty, save for a 65" flat panel TV, one ~ 30 inch x 30 inch uphostered chair, ~ 4 ft x 2.5 ft open glass table and four 12" Rythmik F12 subwoofers. Living room covered with old thin pile carpeting, over which is ~ 45% new 1/2 inch remnants.

I can treat the walls, ceiling, windows and floor coverings in what ever would be recommended.

The combination of a good horn and spaciousness requires high quality diffusion. One could also consider a horn with somewhat wider directivity and diffusion on oppoiste side wall reflections. Depends on the rooms size, what's practical possible, etc. as well.
Do you mean, wider directivity than the AH425 horn if the opposite wall and side walls were acoustically treated to create some measured degree of diffusion? If yes, are you describing a horn like this?


Or https://josephcrowe.com/products/3d-cad-files-horn-no-1994-es450-biradial-for-jbl-2446-2-throat

Or perhaps, however much less likely, a higher directivity horn? https://img.usaudiomart.com/uploads/2022/02/10/649826210_thumb_f770641ce3ac2ce1d7f6e5c79eac3675.jpg

All horns designs have their positive and negative aspects. While I haven't testet everything, my experience is that that a good designed diffractional horn is perhaps the best compromise if it's large enough and with DSP. The reason is that you can achieve constant directivity in both planes and the minor peaks and dips are minimum phase behaviour, meaning EQ will work perfectly.

While there are other horns that can achieve a flatter raw response, they will not be broadband constant in both planes.
By diffraction horn do you mean like these?

Those JBL and/or EV model horns may still be available. Would they or which other commercially available diffraction or other horn (s) might be best to achieve spaciousness in my room, plus with any necessary acoustical treatment?
 
My living room is ~ 23 ft L x 15 W. Opposite the ~23 ft east wall, about 3/4 of the opposite side empties into a ~ 10 ft x 9 kitchen and a 10 ft L x 4 ft w stair case, which turns right and continues ~ 15 ft down to ground floor. There is a triangular ceiling which peaks at 11 ft extending over the living room, kitchen and staircase. Three ~ 40 inch x 30 inch bay windows at one end of the north wall in living room; so that wall is most trigonal shaped. Most of my living room is empty, save for a 65" flat panel TV, one ~ 30 inch x 30 inch uphostered chair, ~ 4 ft x 2.5 ft open glass table and four 12" Rythmik F12 subwoofers. Living room covered with old thin pile carpeting, over which is ~ 45% new 1/2 inch remnants.

I can treat the walls, ceiling, windows and floor coverings in what ever would be recommended.

Do you mean, wider directivity than the AH425 horn if the opposite wall and side walls were acoustically treated to create some measured degree of diffusion? If yes, are you describing a horn like this?


Or https://josephcrowe.com/products/3d-cad-files-horn-no-1994-es450-biradial-for-jbl-2446-2-throat

Or perhaps, however much less likely, a higher directivity horn? https://img.usaudiomart.com/uploads/2022/02/10/649826210_thumb_f770641ce3ac2ce1d7f6e5c79eac3675.jpg

By diffraction horn do you mean like these?

Those JBL and/or EV model horns may still be available. Would they or which other commercially available diffraction or other horn (s) might be best to achieve spaciousness in my room, plus with any necessary acoustical treatment?
Your room is wide enough for diffusion on opposite side wall reflection points, something that will create a very large, spacious and addicted result with a broadsband diffuser. You would need to contact me for acousic consultancy: https://www.audiosciencereview.com/forum/index.php?threads/acoustic-consultancy.45771/

I'm no fan of short horns because they loose the vertical directivity high in frequency. I strongly believe in having broadband constant directivity, which means very big horns. If not, I think there are other speaker designs are better than horns. The old diffraction slot horns all had their issues. I'm talking about a modern diffaction slot horn, and I'm biased here since we have developed one. You can read about it here: https://www.audiosciencereview.com/...audio-midrange-horn-polar-measurements.42732/

We also did a radial diffraction slot horn. Oh please note that JBL used radial and biradial completely incorrect! JBL never sold a biradial horn. A biradial horn revolves in both places, while a radial only in one plane. Radial and biradial doesn't really have anything to do with a diffraction slot in the true meaning. Below is the prototype radial horn we did, which is a great horn with broadband CD in both planes but would require a similar shaped bass bin.
Notice how short the diffraction slot is. A horn like this has better raw response than for example the Klipsch K-402 and holds the vertical directivity much lower in frequency.

Proto biradial front (Medium).jpg


Proto biradial side (Medium).jpg
 
Last edited:
Back
Top Bottom