Hi all,
Simple question for concept. I was looking at designs and it made me wonder about the idea of using an open baffle speaker, something 100hz to 20khz bandwidth that is a dipole, and then right behind it, identical drivers in a sealed cabinet with a little space between the drivers, aligned. The backwave of the open baffle and the front wave of the sealed with opposing phase should largely cancel each other out. Does that leave only the front wave? And the dipole pattern handles the side to side off axis output too? I've got to be missing something here, other than the obvious requires DSP to do the phase and delay to get the effect and the double drivers, etc. What huge thing am I missing?
Very best,
Simple question for concept. I was looking at designs and it made me wonder about the idea of using an open baffle speaker, something 100hz to 20khz bandwidth that is a dipole, and then right behind it, identical drivers in a sealed cabinet with a little space between the drivers, aligned. The backwave of the open baffle and the front wave of the sealed with opposing phase should largely cancel each other out. Does that leave only the front wave? And the dipole pattern handles the side to side off axis output too? I've got to be missing something here, other than the obvious requires DSP to do the phase and delay to get the effect and the double drivers, etc. What huge thing am I missing?
Very best,
I think the reason it's not done is the high expense for not much benefit. I'm not quite sure what the benefit would be of an open isobaric. However, I'm sure you could market this as the best loudspeaker ever.
It's a battle against chaos. How do you chase each particle and stop its motion by making it collide with an anti-particle ?! It's just a matter of phase
Since they are separated in space, there will likely be some directional leakage.
If you only want one front wave, what is the reason you are doing this and not just using a box, so we can suggest a solution?
If you only want one front wave, what is the reason you are doing this and not just using a box, so we can suggest a solution?
I made a half hearted attempt at this once and got dubious results. I might try again sometime. I think it’s futile without carefully EQ’d DSP and separate DSP channels and amps driving each woofer.
I think you are looking at constructing a cardioid pattern speaker. A dipole added to an omni pattern. Have a look at the Dutch and Dutch patents or Kii Audio speakers.
I've done exactly this, but unfortunately it was 15 years ago. I remember it being great at the time, but it's difficult to reference that against anything I have today.
With that one, I had 1x OB driver, and then 1x sealed driver positioned behind and just below.
Some notes, partially from memory, partially from what I'd do if I was to approach this project again:
I'd say give it a try and report back. From a Physics point-of-view, the idea is solid within certain bounds.
Chris
With that one, I had 1x OB driver, and then 1x sealed driver positioned behind and just below.
Some notes, partially from memory, partially from what I'd do if I was to approach this project again:
- You're only going to get good rear-wave cancellation if you keep the drivers reasonably close together. 1/4-wavelength ideally.
- Even then, there'll be an upper limit as the OB driver's magnet starts causing issues.
- You want the sealed driver to have an excursion profile that matches the OB driver. You can do this via EQ, or just using a large sealed box.
- My approach involved a passive 2nd-order LPF which rolled off the sealed driver when the wavelengths were getting too short for cancellation. FIR-based DSP would be a better approach IMO.
- If you wanted to make the polar pattern more consistent, you could apply some absorption to the rear of the OB driver, which would attenuate at HF where the sealed driver cannot.
- You'll probably get more LF output by putting both drivers in a sealed box, but that won't have the directivity pattern you're looking for.
I'd say give it a try and report back. From a Physics point-of-view, the idea is solid within certain bounds.
Chris
It works well with xo below 200Hz, but most certainly with careful dsp help. Remember that we need delay for the main speaker (dipole)!
Combination has cardioid response around xo. Measured outdoors at 1,5m distance. In-room positioning and eq is tricky but worth the effort!
Last edited:
The first article is about an electrostat+ subwoofer, the second about those Gradient speakers that you like. That article contains many speculations that aren't true. First of all, the sound in an anechoic room: the sound lacks of the ecolocalization attribute, as there are no reflections, so it's not a real sound, but somethingh near. Claiming that what the listener ears is more or less the response curve of the speaker is another forcing. The so called room interaction is just the sound ( originating from the speakers) that evolves in ambient.
Thank you all so much for the thoughts.
I mainly was thinking about it as a means to minimize room interaction and out of room SPL potentially. With today's plentiful amplification and DSP channels, it's not unreasonable to me to think about having complex active DSP speakers everywhere and so while I enjoy open baffle, I recently had an experience with a large open baffle that had its back wave fighting other things in the room and I started thinking more about that back wave. Then looked into active cancellation. Then thought about, well, what if I did an open baffle monitor with a range of about 100hz to 20khz and it's a dipole and then can the backwave be canceled? Using a DSP, identical drivers in a sealed enclosure with careful frequency response matching and correct delay, the two phases of the sealed arrangement and the back wave of the open baffle should cancel, at least, some or most of it. Not expecting it to all be canceled. But just minimized significantly. Nothing novel. But, since DSP is so inexpensive now and easy to do, maybe this is something fun to try.
I realize it's not as simple as this, but just wanted to see what others have tried or find info on it. It would be neat to make an active cardioid speaker, if it's not too difficult without a bunch of proprietary things or complex cabinet designs. I couldn't really think of an easy way to do it with a sealed enclosure with slits and all that complex stuff, but I did think of a dipole with something to cancel the back wave, and here we are.
I have DSPs and several identical drivers, so if it's possible, I would be happy to build a physical model and test things. I would do separate active DSP and amplification per driver to keep full control at this stage.
So, if I could move forward, ideal spacing between the drivers (the open baffle to the sealed identical drivers behind them) should be less than a quarterwave worth of distance? How about just as close as possible? At the end of the day, it would just be tossing a little time and scrap plywood away to learn something (for me). Frequency response would need to be as closely matched as possible. I assume also level matched? And then whatever delay it takes to align them (should be easy to spot that if things start cancelling).
One more thought... if I were to integrate subs, any particular notes about how to do similar there? Either open baffle + sealed in the same fashion, or two sealed and one of them in opposing polarity with delay to cancel on that side of the sub to keep direction instead of omni? This is the basis of waveform shaping right? Phase and delay active control with subs?
Worth trying?
Or just chaotic useless?
Very best,
I mainly was thinking about it as a means to minimize room interaction and out of room SPL potentially. With today's plentiful amplification and DSP channels, it's not unreasonable to me to think about having complex active DSP speakers everywhere and so while I enjoy open baffle, I recently had an experience with a large open baffle that had its back wave fighting other things in the room and I started thinking more about that back wave. Then looked into active cancellation. Then thought about, well, what if I did an open baffle monitor with a range of about 100hz to 20khz and it's a dipole and then can the backwave be canceled? Using a DSP, identical drivers in a sealed enclosure with careful frequency response matching and correct delay, the two phases of the sealed arrangement and the back wave of the open baffle should cancel, at least, some or most of it. Not expecting it to all be canceled. But just minimized significantly. Nothing novel. But, since DSP is so inexpensive now and easy to do, maybe this is something fun to try.
I realize it's not as simple as this, but just wanted to see what others have tried or find info on it. It would be neat to make an active cardioid speaker, if it's not too difficult without a bunch of proprietary things or complex cabinet designs. I couldn't really think of an easy way to do it with a sealed enclosure with slits and all that complex stuff, but I did think of a dipole with something to cancel the back wave, and here we are.
I have DSPs and several identical drivers, so if it's possible, I would be happy to build a physical model and test things. I would do separate active DSP and amplification per driver to keep full control at this stage.
So, if I could move forward, ideal spacing between the drivers (the open baffle to the sealed identical drivers behind them) should be less than a quarterwave worth of distance? How about just as close as possible? At the end of the day, it would just be tossing a little time and scrap plywood away to learn something (for me). Frequency response would need to be as closely matched as possible. I assume also level matched? And then whatever delay it takes to align them (should be easy to spot that if things start cancelling).
One more thought... if I were to integrate subs, any particular notes about how to do similar there? Either open baffle + sealed in the same fashion, or two sealed and one of them in opposing polarity with delay to cancel on that side of the sub to keep direction instead of omni? This is the basis of waveform shaping right? Phase and delay active control with subs?
Worth trying?
Or just chaotic useless?
Very best,
Wavelength is the key word. One octave is easy, more than two very difficult even with multiway (at least 4)
You can learn a lot by reading #keysir thread that was foundation to D&D 8c speaker https://www.diyaudio.com/community/threads/2-way-waveguide-cardioid-like.192737/
https://www.erinsaudiocorner.com/loudspeakers/dutch_dutch_8c/
You can learn a lot by reading #keysir thread that was foundation to D&D 8c speaker https://www.diyaudio.com/community/threads/2-way-waveguide-cardioid-like.192737/
https://www.erinsaudiocorner.com/loudspeakers/dutch_dutch_8c/
Since a speaker is a mass-spring system, and you use one of the two in open air and the other in an enclosure, you can no longer truly consider them identical drivers. This is because the mass-spring system changes due to the added air mass and the added air spring. The resonance frequency, Q-factor, and thus the phase response of the two drivers are no longer the same, resulting in significant frequency-dependent summation or cancellation. I tried this once and also attempted to correct it with DSP, but the results weren’t really satisfying in the end. In fact, you can create a cardioid subwoofer this way, but at higher frequencies, it's very difficult to control. My knowledge may be limited in this area, but it’s certainly not easy, like Juhazi allready mentioned, wavelength is key here.
I found that Visaton Boxsim software can be used to model cardioid speakers. I was playing around with a cube speaker with a small driver on each face of the cube. If you run the front and rear firing drivers out of phase that can be considered the dipole element. The four drivers on the sided, top and bottom run in phase can be considered the omni- element. With the levels balanced, this should produce a cardioid pattern.
In the example I captured pictures of I was running all the drivers in phase as an omni. In the speaker I built, I was able to get maybe a 15 dB attenuation of the rear radiation relative to the front.
In the example I captured pictures of I was running all the drivers in phase as an omni. In the speaker I built, I was able to get maybe a 15 dB attenuation of the rear radiation relative to the front.
Distance of sources in z-axis (depth) translates to delay. Baffle width also counts, because soundwaves must go around the whole box/baffle to interfere.
t
Damped u-frame is popular choice to make cardioid woofer (Kreskovsky NaO) . Side vents on sides of a closed box will create cardioid pattern in midrange (D&D)
https://musicanddesign.speakerdesign.net/Dipole-offset.html
t
Damped u-frame is popular choice to make cardioid woofer (Kreskovsky NaO) . Side vents on sides of a closed box will create cardioid pattern in midrange (D&D)
https://musicanddesign.speakerdesign.net/Dipole-offset.html
Last edited:
I’d like all the OB folks to take a look at what Juhazi has shared here……..below 100hz, it’s ALL modal and any attempt to unwind this is perfectly impractical. If you MUST strive for an open baffle situation, work on the midbass…..that’s where the goodness is.
My speaker has monopole bass unit, because the huge dipole loss below 100Hz - compensation would require large baffle and multiple drivers (it is possible, look at Linkwitz LX521)
My reasoning to not use dipole low bass is based on utility (laziness, poor funds and aesthetics) - I don't want a large bass unit and don't have distance for the backwave to not get reflected back early, reinforced by the wall area (minus absorption). Monopole bass has same phase all around, so reflected long wavelength has same phase and sums up (with some delay for the max amplitude.) A dipole's backwave sum is zero or negative... Distance to wall and wavelength (and speed of sound) are the basis to calculate this.
https://musicanddesign.speakerdesign.net/Boundary_reflections.html
Monopole source example
Dipole source
ps. Sorry I can't confirm this with my own measurements, because I have never made a dipole low bass unit! I also very much doubt the claim that dipole low bass excites less room modes because of it's directivity - but that is just a guess. Simply just lower spl of low bass (and modes) might sound "cleaner and more precise" Actually when I met SL in 2013 he asked me to at least test dipole bass, but my answer was that my space is too small. His latest home is deeper where dipole low bass obviously works well. But as well modes are low in amplitude too because of high absorption.
My reasoning to not use dipole low bass is based on utility (laziness, poor funds and aesthetics) - I don't want a large bass unit and don't have distance for the backwave to not get reflected back early, reinforced by the wall area (minus absorption). Monopole bass has same phase all around, so reflected long wavelength has same phase and sums up (with some delay for the max amplitude.) A dipole's backwave sum is zero or negative... Distance to wall and wavelength (and speed of sound) are the basis to calculate this.
https://musicanddesign.speakerdesign.net/Boundary_reflections.html
Monopole source example
Dipole source
ps. Sorry I can't confirm this with my own measurements, because I have never made a dipole low bass unit! I also very much doubt the claim that dipole low bass excites less room modes because of it's directivity - but that is just a guess. Simply just lower spl of low bass (and modes) might sound "cleaner and more precise" Actually when I met SL in 2013 he asked me to at least test dipole bass, but my answer was that my space is too small. His latest home is deeper where dipole low bass obviously works well. But as well modes are low in amplitude too because of high absorption.
Last edited:
I will add this….i’ve done a few IB sub installs over the decades and THAT works……really well…..and if you have adjoining space is THE most practical application I can think of. We’ve done thru wall using both a garage and a utility room as the enclosure……we’ve done through ceiling and through floor manifolds into attics and basements…….just smooth, even in room response. The hope is my next, and likely home through retirement will be built in a few years and will include an IB ground up solution for the main listening space.