OmniDirectional - work in progress

If you put some sort of barrier so that up-firing tweeter doesn't radiate forward, what you then get is only the treble that's arriving at your listening position via the ceiling and maybe some walls - this gives the arriving wave energy a certain delay that should dramatically improve the 'spaciousness' of your sound - look up 'the Romeyn' sound that's based on the work of that speaker guru (plus other things audio) Floyd Toole.
 
I have been interested in this type of omnidirectional waveguide for some time, but I doubt they work they way they look.

They're commonly used in in-ground outdoor "omnidirectional" speakers. But these speakers tend to have very low SPL and I doubt the waveguide does much more acoustically than a cap would, to keep rain out:

Bose%20Free%20Space%2051-3.jpg


I also see them on tsunami warning towers. Those towers consist of discs with four compression drivers each firing into the center where they wrap around 180 degrees out the pseudospherical waveguide. They usually stack several discs:

An externally hosted image should be here but it was not working when we last tested it.


Inside (click to undistort the image):
US5146508-2.png


It's not that much different than a folded horn, whether you fold it 180 or only 90 degrees. I'm doubtful it can be done with these shapes without significant distortion at any reasonably high SPL.
 
Last edited:
Thanks for the interesting purpose specific examples you provided, I haven't seen these applications before.

You can make an omni that sounds good and has good SPL levels for home hifi speakers. There are very few commercially available, hard to find for listening tests, and tend to be expensive, that's one reason I'm building a pair.

The garden speaker is probably a single full range weatherproof speaker pointing down from the "dustcap", the lower freq (<2Khz) are mostly omnidirectional without help if the driver is large enough and the lower reflector probably helps to maybe 5Khz. Thats enough for garden background music. It probably takes a garden hose and dogs better than most speakers.

The stacked horn array is very interesting. I went looking for how they use it. It's unusual because they seem to use it for voice as well as the usual siren tone modulation pattern. Distortions probably not important as you run for your life :)
https://www.youtube.com/watch?v=T4fcg8E_BPM
 
If you put some sort of barrier so that up-firing tweeter doesn't radiate forward, what you then get is only the treble that's arriving at your listening position via the ceiling and maybe some walls - this gives the arriving wave energy a certain delay that should dramatically improve the 'spaciousness' of your sound - look up 'the Romeyn' sound that's based on the work of that speaker guru (plus other things audio) Floyd Toole.
Interesting suggestion. The domes are directional with frequency, having reasonable wide cone at low freq (120deg@2Khz) and narrow cone at high freq (30deg@15Khz).

A barrier would cause reflections or scattering near the dome. However the proper shape could focus the sound much like a parabolic sound dish for listening. Only it would be used in a reverse sense, trying to focus the energy at the reflection point. I need to think about this one
 
Aperiodic Bi-chamber Bass improvement

After listening to them for awhile, I decided to improve the bass response. The current 16L BR response is weaker than I like and seems to start fading at 100Hz although its F3=60Hz.

I'm restricting the volume to 32L so the design choices were ApBc, BP4, or BP6, there's not enough room to fold a quarter wave horn. The Aperiodic was just easiest, as I already have split the volume and have 2 x 16L chambers. All I needed was to join the chambers with an internal port and then port out the second chamber. So total of 3 x tuning ports (2x external 5x17cm, internal 12x2cm).

The results match HornResp simulation and I get response to 40Hz with an 8" woofer. In fact its +10db higher than the BR at 40Hz. Its an audible improvement. This ApBc also has a dip at 175Hz in the near field measurement. The graphs below at @1m with room acoustics.
 

Attachments

  • DSCN1846 reduced.jpg
    DSCN1846 reduced.jpg
    286.9 KB · Views: 546
  • omni BR at 1m.jpg
    omni BR at 1m.jpg
    124.5 KB · Views: 551
  • omni ApBc @1m.jpg
    omni ApBc @1m.jpg
    130 KB · Views: 270
Last edited:
It sounds good and it surprises people when they see the shape and listen to them. You can walk around the room and get good spacial sound from both speakers nearly everywhere. There's even a few special alcove spots where the reflections can mislead you (like acoustic fun house mirrors). The better position for stereo phantom image (using L+R ) is anywhere in the "sweet line" (not a spot) between the speakers about 2m back and at least 1m from the back wall.
 
After listening to them for awhile, I decided to improve the bass response. The current 16L BR response is weaker than I like and seems to start fading at 100Hz although its F3=60Hz.

I'm restricting the volume to 32L so the design choices were ApBc, BP4, or BP6, there's not enough room to fold a quarter wave horn. The Aperiodic was just easiest, as I already have split the volume and have 2 x 16L chambers. All I needed was to join the chambers with an internal port and then port out the second chamber. So total of 3 x tuning ports (2x external 5x17cm, internal 12x2cm).

The results match HornResp simulation and I get response to 40Hz with an 8" woofer. In fact its +10db higher than the BR at 40Hz. Its an audible improvement. This ApBc also has a dip at 175Hz in the near field measurement. The graphs below at @1m with room acoustics.

Don if you also wanted the top end a bit flatter you can use a fairly simple filter to equalize the mass rolloff of the compression driver on the horn. It is usually a 6dB filter at around 20kHz which takes out some of the gain from the horn. Often called HF compensation.

I would think that it would help to reduce that 2Khz peak with your current crossover.

This thread has a circuit diagram for a different CD but it might give you an idea and you seem pretty handy with spice.

Impressive new compression driver from Peerless - Techtalk Speaker Building, Audio, Video Discussion Forum
 
Thanks, and I agree they still need more EQ to flatten the FR. The electronics are not a problem for me, and I will get to that eventually when things settle. I measure and graph raw for the basic cabinet improvements, but I listen with S/w EQ to flatten things. I'm still evaluating how it radiates and works in the room.

I reduce the 2Khz bump (in S/W) as it makes the speaker sound way too bright. I'm also still not happy (but happier than before) with the bass performance. I think I need a 10inch woofer or maybe 2x8inch to give me better room projection and another 3db down to 40Hz.

Thanks for Peerless link. I think the HF peak/dip may be a function of the waveguide load. I've seen another post with the same waveguide, different CD, and same FR shape.
 
I would be tempted to seal the cabinet you have and use some other drivers in separate smaller cabinets that you can place near your existing ones. Linkwitz has used some fairly cheap Peerless woofers for this and they seem to work quite well, Dayton has similar ones.

It is very hard to get a driver to go above 1 to 2 Khz properly but also have significant output below 100Hz.

If you were willing to go active I think you could get a much better response overall and is easier than changing passive components when tweaking. Downside is the need for more amps.
 
I would be tempted to seal the cabinet you have and use some other drivers in separate smaller cabinets that you can place near your existing ones. Linkwitz has used some fairly cheap Peerless woofers for this and they seem to work quite well, Dayton has similar ones.

It is very hard to get a driver to go above 1 to 2 Khz properly but also have significant output below 100Hz.

If you were willing to go active I think you could get a much better response overall and is easier than changing passive components when tweaking. Downside is the need for more amps.
Agreed, that's why I was limited to the 8" for the sandbox. Even with an LR4@1.8KHz I still need the woofer to behave till 3KHz.

I'm trying to keep the cabinet to a reasonable size, but everything points to bigger for better bass.
 
What you might be able to do is mount bigger woofers to the side or front of your existing cabinet and split the volume. The 8" wouldn't need much if it isn't going low.

A 10" driver like the peerless or even a sub driver like the Dayton Reference should fit. A small cabinet with these and a linkwitz transform can put out some decent bass, not so easy to do passively though.
 
Aperiodic Bi-chamber (ABC) whats that bump?

Early when I tested this, I liked the bass extension but I noticed an inflection point at 175Hz that is not like a Bass Reflex dip. So I remeasured both the nearfield driver, and both ports. Sure enough its still there, so are these artifacts at farfield or is the combined response better.

You can see the graphs [front view, driver near, top port, bottom port, speaker farfield] that the driver and 2 ports combine to remove the bump at 175Hz. There is no evidence of the 175Hz defect at farfield (@1m with room acoustics).
 

Attachments

  • omni aperiodic @1m with more stuffing.jpg
    omni aperiodic @1m with more stuffing.jpg
    128.7 KB · Views: 386
  • omni aperiodic bottom port with more stuffing nearfield.jpg
    omni aperiodic bottom port with more stuffing nearfield.jpg
    116.5 KB · Views: 383
  • omni aperiodic top port with more stuffing nearfield.jpg
    omni aperiodic top port with more stuffing nearfield.jpg
    123 KB · Views: 402
  • omni aperiodic woofer with more stuffing nearfield.jpg
    omni aperiodic woofer with more stuffing nearfield.jpg
    125.2 KB · Views: 403
  • DSCN1846 reduced.jpg
    DSCN1846 reduced.jpg
    286.9 KB · Views: 410
Does more stuffing (damping) help in ABC

When I was trying to remove a bump in the bass response I also tried various amounts of stuffing to see how much I could shift or smooth the response. The bump was never removed, didn't need to be, it was compensated by port(s). However I did get some results from damping in an ABC

Graphs below are @1m with room acoustics. Over several attempts with polyester fill I did manage to eek out a couple db. Nothing major, and I did not see a downside.
 

Attachments

  • omni ApBc @1m.jpg
    omni ApBc @1m.jpg
    130 KB · Views: 81
  • DSCN1848 reduced.jpg
    DSCN1848 reduced.jpg
    484.3 KB · Views: 131
  • DSCN1852 reduced.jpg
    DSCN1852 reduced.jpg
    470.1 KB · Views: 147
  • omni aperiodic @1m with more stuffing rescale.jpg
    omni aperiodic @1m with more stuffing rescale.jpg
    130.5 KB · Views: 130
Don aren't the bi chambers usually separate but joined together with a vent of the same size as the ports?

Your picture looks like a reflex enclosure with two ports.

Have you taken an impedance measurement to see what it looks like and if it drops where you want it tuned to?

The top port looks to have a lot of high frequency stuff getting through unless that is leakage from the driver itself?
 
The port connecting the chambers is that square 24cmx5cm (by the black middle foam) as a cutout. You can tune using all 3 ports depending on what you want. In my case I already had 2x16 chambers so that drove the tuning. Cutting the back rectangle passage was the easiest. It's still a sandbox, and the effect still works, and it matches HornResp Sim. The 175Hz blip is exactly where it should be, I just didn't understand why.

The top chamber noise, I believe is interaction between the driver cone the reflector cone above it. It seems to have gotten worse when I reduced the height of the lower cone (used to be more pointy). This was lower when it was a simple 16L BR using just the top chamber and more "pointy" reflector.
 
Last edited:
I'm following this because I've been messin' around with the idea myself.. ;)

The baffle the woofer is mounted to in my design doesn't extend much farther than the outside diameter of the driver. Could it be possible you are getting some reflection issues between the large baffle and the reflector? Cover it with some felt or something to make it less reflective. Just a thought..
 
Last edited:
That's a good idea, thanks.

The baffle was made that large to allow me to go up to a 10inch driver and still mount the posts. I don't have felt on hand but I could get some 1/4" underpad felt that would soak up reflected mid frequencies.

They don't show up in nearfield woofer measurements. But then again, they may be swamped up top in what appears to be a 500Hz resonance, or it could be mechanical resonances from the "stack".
 
APL_TDA

This was a suggestion from Wesayso, http://www.diyaudio.com/forums/mult...design-soundstage-imaging-14.html#post5032808 in another thread to run a program (APL_TDA) that plots the freq content vs time trying to determine the incident wave and its reflections. It's a very interesting idea.

Below are 2 plots (3D and 2D) for one omni that uses ABC and its at my standard measuring position @1m amp@-25db as a baseline so I can compare it to existing FR plots.

Note : ignore all content below 50Hz, my PC requires 6 fans that run constantly and are throwing LF against a back wall. In the FR plots you can see this as the "LF floor" that does not drop as fast as it should.
 

Attachments

  • Omni left @1m amp-25db mid2D.jpg
    Omni left @1m amp-25db mid2D.jpg
    243.6 KB · Views: 110
  • Omni left @1m amp-25db mid3D.jpg
    Omni left @1m amp-25db mid3D.jpg
    279.3 KB · Views: 206