I made a metamaterial enclosure for the SB Acoustics SB26 ADC, and thrilled with it's performance, started thinking about covering a wall with metamaterial absorbers.
Link to original project: https://www.diyaudio.com/forums/multi-way/380104-unity-horn-metamaterial-tweeter.html
There have been quite a few papers written on metamaterials. I think I've read half of them and understood 10% of what I read.
IMHO, the academic papers are unnecessarily complex. Methinks this is because some of these papers were written by graduate students who were using the math and the physics to demonstrate that they'd learned math and physics.
As far as I can see, metamaterials are simply an array of closed back transmission lines.
Due to this, there are a myriad of ways to fold the array of lines. If you've built a transmission line or a tapped horn, you have a pretty good idea of how to fold a TL.
With that concept in mind, for this absorber, I largely focused on using a specific footprint and a specific range of frequencies, and then I folded the lines to use as much of the space as possible.
IE:
In prior metamaterial projects, I was using software to generate and idealized series of frequencies, and then trying to figure out how to fold the line. A lot of space was wasted.
In this project, I let the size of the object dictate the folding, and I basically used up 100% of the space inside of it.
This thing isn't perfect, it's "version 1.0"
What I am envisioning, is getting this absorber refined enough so that I could 3D print as many as 36 of them, or maybe even more. Enough so that I could cover an 8' x 2' strip along the back wall of my living room.
The way that my living room is designed, the speakers basically fire straight into the bedroom wall of my master bedroom. Which means that I can't watch movies, music or TV any later than about 9-10pm. Which is no bueno.
Link to original project: https://www.diyaudio.com/forums/multi-way/380104-unity-horn-metamaterial-tweeter.html
There have been quite a few papers written on metamaterials. I think I've read half of them and understood 10% of what I read.
IMHO, the academic papers are unnecessarily complex. Methinks this is because some of these papers were written by graduate students who were using the math and the physics to demonstrate that they'd learned math and physics.
As far as I can see, metamaterials are simply an array of closed back transmission lines.
Due to this, there are a myriad of ways to fold the array of lines. If you've built a transmission line or a tapped horn, you have a pretty good idea of how to fold a TL.
With that concept in mind, for this absorber, I largely focused on using a specific footprint and a specific range of frequencies, and then I folded the lines to use as much of the space as possible.
IE:
In prior metamaterial projects, I was using software to generate and idealized series of frequencies, and then trying to figure out how to fold the line. A lot of space was wasted.
In this project, I let the size of the object dictate the folding, and I basically used up 100% of the space inside of it.
This thing isn't perfect, it's "version 1.0"
What I am envisioning, is getting this absorber refined enough so that I could 3D print as many as 36 of them, or maybe even more. Enough so that I could cover an 8' x 2' strip along the back wall of my living room.
The way that my living room is designed, the speakers basically fire straight into the bedroom wall of my master bedroom. Which means that I can't watch movies, music or TV any later than about 9-10pm. Which is no bueno.
There seems to be a near infinite number of ways to create a metamaterial or metasurface absorber. This PDF is one of better summaries of the approaches:
https://www.preprints.org/manuscript/201805.0096/v1/download
If the link above stops working (many of these articles on metamaterials are behind paywalls) just google "preprints201805.0096.v1.pdf"
I'm not going to attach the PDF, due to I.P. laws
https://www.preprints.org/manuscript/201805.0096/v1/download
If the link above stops working (many of these articles on metamaterials are behind paywalls) just google "preprints201805.0096.v1.pdf"
I'm not going to attach the PDF, due to I.P. laws
A few years back, I listened to Jon Whitledge's Magic Bus for the first time.
Jon played a track for me, and for the first twenty seconds of the track, I thought to myself "This is a really good cover of this song."
The song was "Billie Jean" by Michael Jackson. It starts out with a driving bass rhythm by Louie Johnson, which sets the tone for the song.
When listening to it in John's bus, it took me HALF A MINUTE to realize it wasn't a cover - it was the original song. But the bass sounded completely different. The notes were all the same, but the 'space' between each note was so clearly defined. It was almost as if you took the bass player out of your listening room and had him play outside. The definition of the bass line was just night and day different.
This really blew my mind.
I believe the 'secret' to why the bass sounded so distinct is that John uses a series of acoustic treatments in the bus.
I'll ping him to see if he'll chime in, but if I'm not mistaken, they are helmholtz absorbers.
You can learn more about John's Magic Bus at Magic Bus - Car Stereo, Car Stereo Installation, Car Stereo Tuning
The build log is here:
Build Log
Jon played a track for me, and for the first twenty seconds of the track, I thought to myself "This is a really good cover of this song."
The song was "Billie Jean" by Michael Jackson. It starts out with a driving bass rhythm by Louie Johnson, which sets the tone for the song.
When listening to it in John's bus, it took me HALF A MINUTE to realize it wasn't a cover - it was the original song. But the bass sounded completely different. The notes were all the same, but the 'space' between each note was so clearly defined. It was almost as if you took the bass player out of your listening room and had him play outside. The definition of the bass line was just night and day different.
This really blew my mind.
I believe the 'secret' to why the bass sounded so distinct is that John uses a series of acoustic treatments in the bus.
I'll ping him to see if he'll chime in, but if I'm not mistaken, they are helmholtz absorbers.
You can learn more about John's Magic Bus at Magic Bus - Car Stereo, Car Stereo Installation, Car Stereo Tuning
The build log is here:
Build Log
In “Composite honeycomb metasurface panel for broadband sound absorption,” the authors present a solution which seems almost impossibly simple:
1) a honeycomb structure
2) a backing panel
3) a solid panel on top with a series of holes of varying sizes
To me, all of this just seems way too easy to be true. It is far and away simpler than 3D printing 36 metamaterial labyrinths.
But the more I research this, the more that I realize that it should work:
In a traditional helmholtz absorber, like what is pictured in post #3, there is a single chamber with a plethora of holes of equal size.
That absorber is tuned to a single frequency, and absorptive material is inside to broaden the bandwidth.
In the metasurface design (pictured above), there are a plethora of chambers. The reason we use a plethora of chambers is that we can vary the tuning frequency by varying the diameter of the holes.
This works exactly the same way as a ported sub box, but instead of tuning the ported box to the woofer, we're tuning the ported box to the frequencies we want to absorb in the room.
One way of doing this metamaterial would be to use a series of chambers of varying sizes. But it is just as valid to use port holes of varying size, with chamber sizes that are all identical.
The latter solution is dramatically simpler to build.
I believe you could calculate the depth of the honeycomb and the depth of the port holes using this:
Calculate Perforated Helmholtz Panel Absorber
ha ha as someone who has a paper on electromatic metamaterials I couldn't agree more. Quite a lot of papers have maths thats not even useful in them!
That's what the student papers are for.
So are these good for lining a speaker cabinet? Do they have to be 3D printed?
So are these good for lining a speaker cabinet? Do they have to be 3D printed?
They look like Binary Amplitude Diffusors, BAD panels, holes drilled in a top plate over an absorber to make them more high frequency diffusive and lower frequency absorptive.
Acoustic Absorbers and Diffusers
Trevor Cox and Peter D’Antonio
11.1 Planar hybrid surface
The binary amplitude diffsorber, also known as a BADTM panel [1], is a flat hybrid surface having both absorbing and diffusing abilities. The panel simultaneously provides sound diffusion at high and mid-band frequencies, and crosses over to absorption below some cut-off frequency. Figure 11.1 shows a typical construction. Mineral wool is faced with a complex perforated mask, and the panel is fabric wrapped for appearance. The white patches on the mask are holes, and the black patches hard reflecting surfaces. Figure 11.2 shows the random incidence absorption coefficient for a hybrid surface compared to the mineral wool alone, the effect of changing the backing depth is also shown. The additional vibrating mass within the holes of the mask causes the absorption curve to shift down in frequency generating additional low to mid-frequency absorption. At high frequency, the hard parts of the mask reflect some of the sound, preventing absorption happening in some parts of the mineral wool, and so causing the absorption coefficient to reduce. It is at these high frequencies, where the absorption is reduced, that the surface should start to generate significant amounts of diffuse reflections. To accomplish mid- to high frequency dispersion, a 31 33 2D array of absorptive and reflective areas is used. The reflective areas map to the 1 bit and the absorptive areas map to the 0 bit in a binary pseudorandom number sequence. The distribution of these binary elements is based on an optimal binary sequence with a flat power spectrum as this maximizes dispersion. For example, this could be based on an maximum length sequences.
Hybrid surfaces extend the acoustical performance of traditional fabric wrapped absorbers, and allows wide area coverage without excessive deadening at mid- to high frequencies. The surface is used in facilities that need reflection control, and provides it from a simple inexpensive construction. Furthermore, the acoustic function can be hidden which can lessen the conflict between visual aesthetics and acoustic requirements. Alternatively, many architects have expressed interest in seeing the mask, due to its unique appearance and the fact that it offers an alternative to the traditional periodic perforated metal patterns.
The ones that I 3D printed for a tweeter were a home run:
3D Printed Metamaterials
It kinda made me lol that nobody commented on the post, because the improvement in measured performance was unreal.
Third harmonic distortion was reduced to unmeasurable levels above 2khz!
When I tested the SB26ADC with the metamaterial enclosure, it was a trip, I just kept pouring power into it and it sounded as clean at 100dB as it sounded at 80dB. It's a really peculiar invention, it makes the speaker sound like it has unlimited headroom.
Are you saying that much of the measured HD is actually reflections off the cabinet wall
or driver structure from the back side of the cone, and then coming back though the cone?
or driver structure from the back side of the cone, and then coming back though the cone?
Last edited:
Well I'd always suspected that reflected energy in the cabinet might have some ability to impact the incident wave, and apparently that's true.
What the mechanisms are, is beyond my level of understanding.
If someone understands the mechanics of loudspeaker distortion, please chime in.
Paging @SpeakerScott...
What the mechanisms are, is beyond my level of understanding.
If someone understands the mechanics of loudspeaker distortion, please chime in.
Paging @SpeakerScott...
Not a speaker guy, but I've always thought that white noise seemed to be a difficult test
for a speaker, from how they sounded with it.
So people measuring HD in speakers have misunderstood the results of the measurement?
Interesting.
for a speaker, from how they sounded with it.
So people measuring HD in speakers have misunderstood the results of the measurement?
Interesting.
Last edited:
I just put some lossy medical coton behind tweeter metal dome...it works well. Acts as difusor and certainly damp the close reflections back to the rear of the dome not being absorbed by the rear chamber. Or at least distribute it back less focussed maybe.
For the bass purpose I beleive JM Reynaud loudspeakers if I am not wrong uses such solid passive filter brick right behind the woofer.... Don't how it is made...resonator tube drilled into something perhaps.
I wonder if a damped open cornu shape behind a woofer could acheive same result and distribute the force on internal side panels ? Make an overture between the cornu spiral ends and exterior there will go into a meta material cardioid thingy ?
Not to forget what Yamaha is also doing with its closed tube circuitry...
For the bass purpose I beleive JM Reynaud loudspeakers if I am not wrong uses such solid passive filter brick right behind the woofer.... Don't how it is made...resonator tube drilled into something perhaps.
I wonder if a damped open cornu shape behind a woofer could acheive same result and distribute the force on internal side panels ? Make an overture between the cornu spiral ends and exterior there will go into a meta material cardioid thingy ?
Not to forget what Yamaha is also doing with its closed tube circuitry...
I thought this had potential, given the depth was much less than the 1/4 wavelength normally required: Phys. Rev. X 7, 021034 (2017) - Ultrathin Acoustic Metasurface-Based Schroeder Diffuser
I am a newbie to all this and am greatly interested in this thread.
@Patrick Bateman: Thanks a lot for your efforts and making the knowledge accessible to everyone. I got a doubt while reading these posts which I am trying to ask here.
I have been reading about the passive cardioid type speakers on the forum, where one of the ideas used (as per my limited understanding) is to use (by leaking) the back waves of the driver out of the box with enough phase shift such that it interferes and attenuates the radiation at certain angles resulting in cardioid radiation pattern from the speaker in the lower frequencies. I see such designs involve lot of variables and lot experimentation with fibrous material to get the pattern right.
Would it be possible to implement such designs more efficiently with acoustic metamaterial surfaces? In the sense that if we could use something like a carefully tuned 3D printed acoustic metamaterial 'device' to induce phase shifts to a broad range of frequencies in the midrange and make it interfere with direct waves from the driver to create a pattern of radiation that we want?
Thanks
@Patrick Bateman: Thanks a lot for your efforts and making the knowledge accessible to everyone. I got a doubt while reading these posts which I am trying to ask here.
I have been reading about the passive cardioid type speakers on the forum, where one of the ideas used (as per my limited understanding) is to use (by leaking) the back waves of the driver out of the box with enough phase shift such that it interferes and attenuates the radiation at certain angles resulting in cardioid radiation pattern from the speaker in the lower frequencies. I see such designs involve lot of variables and lot experimentation with fibrous material to get the pattern right.
Would it be possible to implement such designs more efficiently with acoustic metamaterial surfaces? In the sense that if we could use something like a carefully tuned 3D printed acoustic metamaterial 'device' to induce phase shifts to a broad range of frequencies in the midrange and make it interfere with direct waves from the driver to create a pattern of radiation that we want?
Thanks
Very interesting stuff!
I don't know about acoustics, but at least on the optics side metamaterial is buzzword which gets tacked on all sorts of stuff to get papers published.
So a 17 cm thick panel for fixing that 100 Hz problem in the room response. Laser-cut plywood perhaps?
I don't know about acoustics, but at least on the optics side metamaterial is buzzword which gets tacked on all sorts of stuff to get papers published.
So a 17 cm thick panel for fixing that 100 Hz problem in the room response. Laser-cut plywood perhaps?
Last edited:
Harbeth on the flagship that has a 12" woofer uses a plain foam block that has almost the internal enclosure side right behind the woofer....there is just some rolled synthetic wool between the foam block and the back of the 12"... perhaps to diffuse or to prevent some non damped waves to go back right behind the cone, or at least with less ernegy. Same technic than solid diffusors. Some even use cotton balls for cloth drier machines here in the forum. Something like an Helmotz internal panel could maybe work. The first layer that prevent reflexion back is certainly the most critical...density choice. Foama la Earl G. ?
Last edited:
Yes, I think that would be viable.
For instance the Linkwitz LX Mini speaker has a cylinder that's full of damping material, situated behind the midrange. The idea is to absorb the back wave.
A cylinder that's a metamaterial would probably work better.
I need to post the measurements, but in my experiments, metamaterials absorbed more sound than a felt blanket does, and also absorbed more sound than simply having the speaker radiate into 360 degrees.
Metamaterials act a bit like a "sponge", absorbing sound.
Thanks a lot for your reply. 🙂
Eagerly looking forward to learning more from your experiments.
There's a very recent paper on a similar structure: https://aip.scitation.org/doi/full/10.1063/1.5025570
In “Composite honeycomb metasurface panel for broadband sound absorption,” the authors present a solution which seems almost impossibly simple:
1) a honeycomb structure
2) a backing panel
3) a solid panel on top with a series of holes of varying sizes
To me, all of this just seems way too easy to be true. It is far and away simpler than 3D printing 36 metamaterial labyrinths.
But the more I research this, the more that I realize that it should work:
In a traditional helmholtz absorber, like what is pictured in post #3, there is a single chamber with a plethora of holes of equal size.
That absorber is tuned to a single frequency, and absorptive material is inside to broaden the bandwidth.
In the metasurface design (pictured above), there are a plethora of chambers. The reason we use a plethora of chambers is that we can vary the tuning frequency by varying the diameter of the holes.
This works exactly the same way as a ported sub box, but instead of tuning the ported box to the woofer, we're tuning the ported box to the frequencies we want to absorb in the room.
One way of doing this metamaterial would be to use a series of chambers of varying sizes. But it is just as valid to use port holes of varying size, with chamber sizes that are all identical.
The latter solution is dramatically simpler to build.
I believe you could calculate the depth of the honeycomb and the depth of the port holes using this:
Calculate Perforated Helmholtz Panel Absorber
This paper is using 3d printed HR and is glued on plates. I think with some openscad scripts you can easily customize a plate for the frequency you need.