inverse horn for omni directional?

I think you have to ask yourself first, "what am I trying to do"?

There is one company that uses a compression driver into a horn that is aimed down, there being a complementary curve below, with the exit being 360 degrees rotated about the vertical axis.

You might want to look around for that?

Not quite "omni directional" but definitely doughnut shaped certainly.

At bass freqs most woofers are omnidirectional or at least hemispheric.

As the frequency goes up for any vibrating surface the wavelength gets shorter. As the wavelength gets shorter the surface radiating the sound begins to emit the sound in more and more of a direction perpendicular to the vibrating mode. The critical point is when the wavelength gets around 1/4 wavelength distance compared to a dimension of the vibrating object.

So one answer is to use smaller size surfaces to make sound below 2kHz. Of course the problem with that is merely max output level.

That can be solved by a multi driver line source - which then resolves to a longish ribbon driver of narrow dimension.

You can also look at the Walsh driver solution as well.

Unfortunately there is nothing that is inverse of a horn unless you consider some spherical or egg shaped object with a driver mounted on the surface - but it will have no "wavguide" properties per se.

You could scale up the above described "doughnut" horn...

_-_-bear :Pawprint:
planet10 said:

Here is a pic of my latest surround experiment ... these will hang from the ceiling just behind and a bit to the sides of the listening area. They have a 6" FR w whizzer (costing only labour and some ductseal, these are very much Frugal-phile (tm) material.


bingo this is what I'm looking for. is it just any exponential "sp?" curve for the deflector or is it made like a horn ? I'm making plasma tweeters and want 360º bass/midbass up to 1.2Khz where the tweet will take over. The plan is to use a 7 or 8in Dayton RS series woofer. iv seen the ones for sale but they don't list specs on the horns, only this pic.


  • bl3dcadtop.jpg
    24.8 KB · Views: 540
Ummm... is the plasma tweeter direct radiating out of a horn?

The curve of the HF element shown in that cutaway is likely to conform to one of the usual curves (exponential is one).

The Mid driver is likely only getting reflected from the surface that it faces, as there is not going to be much if any compression taking place, and the mouth size is too small to have any effect at LF... I think those speakers may have another woofer for LF, not sure...

it's not clear from that cutaway but the commercial speaker uses an upward firing mid that sits where the wireframe part of the image posted is...

_-_-bear :Pawprint:
killerfishes said:
how would one design/ make one. I figure an up fireing woofer with a horn shaped cone above it to disperce the sound. I would need to play with XO points but it would be set at under 2khz.

Any ideas?

Are these for your mains/fronts? Have you looked at the Linkwitz Pluto? No "dispersion horn" above the woof - it just fires up, and the XO point is set so that the woof operates in a wide dispersion frq range.

I'm guessing there is some loss of sound pressure/efficiency without a "lens" to concentrate the sound into a vertical field, but these speakers are intended for relatively smaller listening environments, and I think allowing the woof to fire upward "unhindered" may make for a more realistic illusion of the live performance.

I've had many thoughts about these kinds of designs and whenever I finally build my "ultimate" speakers I think they will be something like bi-polar/omni-directional in concept. A few designs use conventional forward-firing drivers plus a ceiling-firing driver mounted in the top of the cabinet - possibly getting good relative efficiency in the primary soundstage with the added onmi dimension up top.

There is also the bi-polar design - for fronts/mains they are mounted in the front and rear baffles of a conventional cabinet, 180 out of phase; I wonder what this would be like with a top-mounted driver in addition - overkill maybe? I've also seen some designs with the woof mounted at about 45 degrees toward the ceiling - I was thinking of trying this with a bi-polar design, tweeter mounted on top between the woofs, facing forward.

Lots of possibilities here...
Re: Re: inverse horn for omni directional?

sdclc126 said:

I'm guessing there is some loss of sound pressure/efficiency without a "lens" to concentrate the sound into a vertical field,

my problem is I want them to look and sound good. I'm going to stack the tweet over the woofer, I know playing into a flat bottom will sound like ****.

the Idea

edit. I think this may be a good time to test out my vacume bagging skills, I will get some HD foam and hot wire it into shape then overlay it with carbon/fiber, kevlar, or CK.


  • speakers.jpg
    23.3 KB · Views: 592
Your rendering/idea looks somewhat like these:

This is actually not a new design - there have been several speakers like this on the market over the years. I think it's quite a challenge for the DIYer, with the curved horn surfaces, etc., but I'd sure love to hear the result.

Good luck with your project - I'm sure we'd all like to see what you've come up with when you're finished. Hope the Duevels give you some ideas.

BTW - have you decided on drivers yet, or will that be decided when the concept is more advanced?


2004-10-11 5:21 pm
sdclc126 said:
Your rendering/idea looks somewhat like these:

This is actually not a new design - there have been several speakers like this on the market over the years. I think it's quite a challenge for the DIYer, with the curved horn surfaces, etc., but I'd sure love to hear the result.

I knew I'd seen it somewhere else, I've been looking for that site all morning. I think there is a company 'Celtic Audio' or similar in the UK that do something very similar but using a conventional tweet, not a compression driver.
Pardon the interruption:

Where may I find details about the applications for omnidirectional speakers? I'm trying to get my mind around the concept; rather, the concept is trying to get around my mind...

My mind works like this: one places two speakers on either side of the "stage" for stereo reproduction. What is omnidirectional doing which is different or better?

In theory a truly omnidirectional speaker will have no "sweet spot" as do normal dynamic speaker systems. The sweet spot is due to the polar response of the drivers narrowing as the freq range climbs - general principle. So, if a speaker is truly omnidirectional, the polar response looks like a ball/sphere, equal in all directions, ergo the sound is the same sounding no matter where you are with respect to the speakers.

Some also like the idea, as with dipoles because they will "splash" some energy off the surrounding walls to the rear, and while dipoles won't, these will also off to the sides - and if truly omni, up and down too...

Hope that explains...

Oh, simplified greatly:

regular speakers = bulb with reflector
omni speakers = bare bulb


Horn-Loaded Omni-Directional

What I see with speakers like the Duevels is more of a disk-shaped dispersal pattern of the sound - yes it's 360 degrees but it may be in a relatively narrow plane due to the focusing effect of the horn(s), so it may still have a "vertical sweetspot", above and below which may be a diminished polar response. Room reflections may, however, give a larger-feeling dimension to the music from speakers like this.

Without having heard anything like this I can only speculate that a dipole or bi-pole design may still render a more realistic soundstage - these may actually sound more 3D than a horn-loaded omni, especially if they are line-source in design and direct more energy at the listener and the rear wall.

I would still consider possibly a non-horn-loaded design, similar to the Pluto, so the sound is projected mostly hemispherically up into the room.
sdclc126 said:
Your rendering/idea looks somewhat like these:
BTW - have you decided on drivers yet, or will that be decided when the concept is more advanced?

That page is where i got the cutaway, but i lost the link.

I was thinking of the Dayton RS225s or RS180s if i go ported, still looking for any other cabinet type. I will not know what cabinet I'm using till the tweet's are made and tested in room, I want as much input on this as you'll give, iv never done any omni project
Iv never even seen Di/Bipoles in person but i may need to try this. should i use a Dipole setup? the tweeter will be inphase 360º so the back side of the setup would be reflecting out of phase and could be detrimental at the XO point.
I would suggest you go to and read the site thoroughly. The Phoenix and Orion are dipoles, the Pluto an omni. Dipoles (and bipoles I suspect) need a relatively large room with around 4+ feet from the rear wall, so take that into account. The Orions are quite expensive including the crossover and amp, but you can design a passive system and don't need to go as high end on the drivers if it's not in your budget.

There are many builders on this site who have built dipoles so plug that term into a search and see what you come up with. I see lots of dipoles and fewer bipoles and omnis. I believe the added dimension these designs put into the soundstage provide a critical aspect of realism to the listening experience.

These are often more complex designs than traditional boxes, as you know, and usually require more complex crossovers and equalization - so this is a pretty big undertaking for a beginner. A dipole might be the easiest and least expensive, as you just mount your drivers on an open baffle, but as I said above, getting them to sound right can be a challenge - you should really have some measuring software to tweak everything.

You can copy most of the designs you see online without buying their kits, but with a kit someone else has done all the designing and testing for you - but again we're talking more money. The Pluto or your own clone of it would probably be your best bet in terms of cost, simplicity, and if you have a smaller listening room than is optimal for dipoles.

Hope this helps some. Other with greater technical knowledge than I will hopefully chime in.
The nature of plasma tweeters it is pointless to use passive as they run off 2Vss, the plans I'm using are from or
my room now would not allow for 4' spacing so that is out plus im not using dipole as the sound from the rear or the woofer would be out of phase with the sound from the tweeter. the problem I'm facing with anything other than omni is that plasma has a 360º spherical output that is in phase every where.

"it may still have a "vertical sweetspot"" that fine, i always sit at the same height but some time closer, like now I'm sitting less than a foot from the speakers, but when I'm just playing music or watching a movie I'm seven feet away.
I still think omni with a tall horn to try have the tallest vertical sweet spot i can. another thing i thought of is a horn that is smaller in size than the cone, if the cone is 7" on the 8"speaker make a 5" horn.
If any one can recommend a good brand speaker for sealed app's i would be thank full. I'm looking to cover from 50-60hz to 1.2Khz . As for cost I'm not letting my self cheap out on the woofers, if I'm spending $250 bucks on parts for the plasmas why would i buy $5 blaplunk car speakers.

You sound confused...

It is important to discriminate between a horn and a reflector.

A horn has GAIN at some frequncies and a mouth size that determines the lowest frequency that the horn will work at. Near and below the mouth frequency the frequency response curve will look like a rollercoaster.

It is possible to make a reflector to bounce sound, but that to has dimension vs. frequency issues to contend with.

In short you can't make a tiny horn go very low.

Think/read a little about the Bose 901 theory - the "direct-reflecting" aspect, and how and why most dipoles (ribbons & ESLs) are situated 3-4 feet from a wall... this may yield some insights and clues.

The bottom line is that at freqs above say 250Hz. it starts to get difficult to get a dynamic driver to exhibit a uniform polar response, much less an "omnidirectional response" or even a "donught" shaped response via a baffle/reflector or horn/waveguide. This is especially true if you want a single driver to cover a range from ~2kHz DOWN to bass freqs, which I seem to read you saying??

No matter what, the lower you want to go, the bigger the things gets.

_-_-bear :Pawprint:
Yes the omni with a cone/horn/lens is a challenging design. I would say try a dipole with your plasma tweeter situated on top, or take a good look at Linkwitz's Pluto - again, no lens/cone, and you can put the plasma tweet right where he has his conventional tweeter. This is an active design so of course your version of it can be active too.

Please consider starting out with a less complex design and working your way up as you gain knowledge and experience - whatever you start with will be your template for the next project - it can just evolve over time as many projects do. You can add/replace components as the design evolves - you won't be "stuck" with whatever you start out with - you can experiment with the basic platform at your whim.

I don't know much about that plasma tweeter either - consider perhaps a conventional tweet to start out with - something relatively low-priced compared to performance, such as Vifa; you'll get good sound and won't blow your budget, leaving room in your wallet for improvement later on.