Apologies. You didn’t state the frequency thing. You stated however the structure of the mode would not change. And ‘fill in the spatial gaps’ (there I went ��
. When, however, you sum the output of two equal sources spaced apart in a certain dimension, something funny happens to the standing wave in that dimension. Of course it all is relative, but you can null a certain modal resonance by creating another one with the same f but different phi. Finding the balance of course is the key. As you stated, it is semantics, I think we agree on the result.
. When, however, you sum the output of two equal sources spaced apart in a certain dimension, something funny happens to the standing wave in that dimension. Of course it all is relative, but you can null a certain modal resonance by creating another one with the same f but different phi. Finding the balance of course is the key. As you stated, it is semantics, I think we agree on the result.
Returning to the original question, has anyone ever experimented with active bass traps?
If I understand correctly, the AVAA C20 forces a sound pressure of close to zero in its front chamber, which is covered by material with a certain flow resistance. I am not sure because the schematic suggests the microphone is placed outside of the front chamber. Roger Roschnik (PSI Audio): "Low frequency active velocity acoustic absorber" - YouTube
Signal processing is done in the analog domain, which I think suggests that a simple feedback controller is used to drive the 10" woofer
At the grille the impedance is said to be 150 Pa*s/m. PSI Audio AVAA Demystified - Chapter 2 - How it works - YouTube
Grille area is 0.2 m^2. PSI Audio AVAA Demystified - Chapter 2 - How it works - YouTube
The distance between grille and corner is in the order of 0.5 m.
With this information, it should be possible to build one yourself.
Maybe relevant, the impedance of a spherical wave is
Z(r) = z_0 / (1 + 1/(i*k*r))
in which
z_0: plane wave impedance (air: 400 Pa*s/m)
i: imaginary unit
k: wave number = 2*pi/wavelength
r: radius
https://ocw.mit.edu/courses/electri...earing-fall-2004/lecture-notes/lec_3_2004.pdf page 3
The equivalent absorption area of the AVAA C20 peaks at 35 Hz. PSI Audio AVAA Demystified - Chapter 2 - How it works - YouTube
Calculating the impedance at the surface of a spherical wave, which somewhat coincides with the grille of a corner placed AVAA C20, results in an impedance of:
Z_0 = 400 / (1+ 1/(i*0.64*0.5)) = 37 + 116i Pa*s/m
which approaches the quoted 150 Pa*s/m, so maybe these are related. A matching impedance means that there is no reflection.
If I understand correctly, the AVAA C20 forces a sound pressure of close to zero in its front chamber, which is covered by material with a certain flow resistance. I am not sure because the schematic suggests the microphone is placed outside of the front chamber. Roger Roschnik (PSI Audio): "Low frequency active velocity acoustic absorber" - YouTube
Signal processing is done in the analog domain, which I think suggests that a simple feedback controller is used to drive the 10" woofer
At the grille the impedance is said to be 150 Pa*s/m. PSI Audio AVAA Demystified - Chapter 2 - How it works - YouTube
Grille area is 0.2 m^2. PSI Audio AVAA Demystified - Chapter 2 - How it works - YouTube
The distance between grille and corner is in the order of 0.5 m.
With this information, it should be possible to build one yourself.
Maybe relevant, the impedance of a spherical wave is
Z(r) = z_0 / (1 + 1/(i*k*r))
in which
z_0: plane wave impedance (air: 400 Pa*s/m)
i: imaginary unit
k: wave number = 2*pi/wavelength
r: radius
https://ocw.mit.edu/courses/electri...earing-fall-2004/lecture-notes/lec_3_2004.pdf page 3
The equivalent absorption area of the AVAA C20 peaks at 35 Hz. PSI Audio AVAA Demystified - Chapter 2 - How it works - YouTube
Calculating the impedance at the surface of a spherical wave, which somewhat coincides with the grille of a corner placed AVAA C20, results in an impedance of:
Z_0 = 400 / (1+ 1/(i*0.64*0.5)) = 37 + 116i Pa*s/m
which approaches the quoted 150 Pa*s/m, so maybe these are related. A matching impedance means that there is no reflection.
Last edited:
Hello TBTL,
I admire your grasp and ability to navigate through the interesting math equations you show in your post. Digestible food for my thought.
I am in your corner [and other DIYers in this thread] who study room acoustics.
Please see my last few posts in Pass Labs Forum thread which I had entitled "DIY the device of US Patent 4,899,387. I continue to grapple with room acoustics and am getting satisfactory results.
Best wishes
Anton
I admire your grasp and ability to navigate through the interesting math equations you show in your post. Digestible food for my thought.
I am in your corner [and other DIYers in this thread] who study room acoustics.
Please see my last few posts in Pass Labs Forum thread which I had entitled "DIY the device of US Patent 4,899,387. I continue to grapple with room acoustics and am getting satisfactory results.
Best wishes
Anton
+1 for DBA, ever while in real room not so perfect as in ideal cuboid. Room was 4.8x4.2x3 m and now is 9x4.2x3. DBA 2+2 red, only 2 front subs green:
Attachments
I think this statement is wrong. If the assumptions are correct, the acoustic impedance is almost completely imaginary. Therefore the presented impedance must be imaginary as well to prevent a reflection (and hence absorb the incoming sound wave).
Thanks for the suggestion, much information in this thread.
DIY the device of US Patent 4,899,387
Yes DBA is a perfect absorber and multisub allows for even more degrees of freedom (given the same amount of sources / absorbers), but I am not after the best result. I wish to learn about absorbers.
Last edited:
I may try something
Thanks for reviving the subject. I wanted to try that since a long time.
About the reflection/resonance debat...
Reflection is the process in which a wave front changes direction due to a change of impedance of the field its propagating in (sound -> goes to a wall -> bounces back).
Resonances are a product of multiple reflections and could have many forms depending on the stimuli. As long as the stimuli signal is not composed of frequencies tuned to a resonance, there is no standing waves but echoes.
Standing waves are a buildup of acoustic energy happening when a source is outputing a wave tuned to one of the resonant frequency of a resonant system. Your room is a resonant system (multiple reflection points and high Q).
Absorbers lower the Q factor of the room, dampening the resonances and making them less spiky.
About multiple subs configuration. They are basically an end-fire array. They are beaming a wavefront in the front while cancelling behind them. The cancellation and beam have bandpass like caracteristics, determined by distance and delay etc... The problem I see is that the beam will still be reflected on the back wall. I never tried them indoor and do not know what's the theory and how to set them up, so I can't tell if it works well and how it sounds. And I'm not ready financially to buy four Sub 6BE.
However, I made an end-fire array outdoor and it sounded enormous.
Back to e-traps.
My intuition is that we need to make the wavefront believe there is no wall (basically, the trap will be showing an impedance hole where it's placed). The wave that goes there would get actively absorbed instead of being reflected by whatever is behind. So, if the pressure is rising, the cone needs to pull. If the pressure is droping, the cone needs to push.
Intuitively again, the best way I could think of doing this would be to simply servo a sub with a PID. The error factor of the PID would simply be the pressure in front of the trap taken with a mic. The PID will then output whatever signal it takes for the measured pressure to stay at zero.
It looks quite realistic to get a stable feedback loop through a DSP while being efficient in our frequency region of interest (under 100Hz). The only problem I see now is the propagation of unwanted wave perpendicular to the offending one, and it could maybe have weird behaviour like frequency shifting etc...
I have an unused sub and a SigmaDSP board laying right under my nose. Let's try stuffs.
Thanks for reviving the subject. I wanted to try that since a long time.
About the reflection/resonance debat...
Reflection is the process in which a wave front changes direction due to a change of impedance of the field its propagating in (sound -> goes to a wall -> bounces back).
Resonances are a product of multiple reflections and could have many forms depending on the stimuli. As long as the stimuli signal is not composed of frequencies tuned to a resonance, there is no standing waves but echoes.
Standing waves are a buildup of acoustic energy happening when a source is outputing a wave tuned to one of the resonant frequency of a resonant system. Your room is a resonant system (multiple reflection points and high Q).
Absorbers lower the Q factor of the room, dampening the resonances and making them less spiky.
About multiple subs configuration. They are basically an end-fire array. They are beaming a wavefront in the front while cancelling behind them. The cancellation and beam have bandpass like caracteristics, determined by distance and delay etc... The problem I see is that the beam will still be reflected on the back wall. I never tried them indoor and do not know what's the theory and how to set them up, so I can't tell if it works well and how it sounds. And I'm not ready financially to buy four Sub 6BE.
However, I made an end-fire array outdoor and it sounded enormous.
Back to e-traps.
My intuition is that we need to make the wavefront believe there is no wall (basically, the trap will be showing an impedance hole where it's placed). The wave that goes there would get actively absorbed instead of being reflected by whatever is behind. So, if the pressure is rising, the cone needs to pull. If the pressure is droping, the cone needs to push.
Intuitively again, the best way I could think of doing this would be to simply servo a sub with a PID. The error factor of the PID would simply be the pressure in front of the trap taken with a mic. The PID will then output whatever signal it takes for the measured pressure to stay at zero.
It looks quite realistic to get a stable feedback loop through a DSP while being efficient in our frequency region of interest (under 100Hz). The only problem I see now is the propagation of unwanted wave perpendicular to the offending one, and it could maybe have weird behaviour like frequency shifting etc...
I have an unused sub and a SigmaDSP board laying right under my nose. Let's try stuffs.
Yes, plenty people have.
It was said in one of the earlier posts, an active basstrap = a multi subwoofer system.
Also similar systems with corner subwoofers have been used for many years by now.
I highly recommend reading Sound Reproduction by Floyd Toole.
The main difference I see between the multi sub system and the etrap would be that you don't need to feed a source to the etrap.
The other thing I was thinking about was to raise the room analysis from simple Fourier analysis to a Laplace based analysis, and try to compensate for room effects with a Volterra kernel, but I'm way over my head on this.
I've found the book, taking a dive into it right now. Thanks
The other thing I was thinking about was to raise the room analysis from simple Fourier analysis to a Laplace based analysis, and try to compensate for room effects with a Volterra kernel, but I'm way over my head on this.
I've found the book, taking a dive into it right now. Thanks
Well yeah, so the biggest difference, and advantage is that any (passive) acoustic system will also treat "non-active sources".
In simple words, anything that isn't related to your audio system like you said.
Useful for like a concert or theater halls, or a little bar or practice place with live music.
Or when there are devices or machines around that act as low frequency sources.
This is the same for an active acoustic system.
When that is not a major requirement, I wouldn't put to much energy in it.
Getting a proper coloration between the listening position(s) and the position of the active acoustic treatment is a real can of worms.
Also getting the feedback loop and stability right is a lot more tricky, in the end it's also a lot more expensive, since you need two active systems.
Practically speaking around twice the price.
While one could use his own active system to do the same thing anyway.
Or get at least similar and very satisfactory results.
In simple words, anything that isn't related to your audio system like you said.
Useful for like a concert or theater halls, or a little bar or practice place with live music.
Or when there are devices or machines around that act as low frequency sources.
This is the same for an active acoustic system.
When that is not a major requirement, I wouldn't put to much energy in it.
Getting a proper coloration between the listening position(s) and the position of the active acoustic treatment is a real can of worms.
Also getting the feedback loop and stability right is a lot more tricky, in the end it's also a lot more expensive, since you need two active systems.
Practically speaking around twice the price.
While one could use his own active system to do the same thing anyway.
Or get at least similar and very satisfactory results.
For a multi-sub system you don't need a DBA.
There are plenty of other tools available, some are even for free.
For example; Multiple Subwoofers: Optimize Them With Multi-Sub Optimizer Software
But I have seen other ones as well.
Plenty of freedom to learn as well!
There are plenty of other tools available, some are even for free.
For example; Multiple Subwoofers: Optimize Them With Multi-Sub Optimizer Software
But I have seen other ones as well.
Plenty of freedom to learn as well!
Thanks for the answer.
MSO is tingling my curiousity. But I don't believe it is an equivalent solution to a DBA or E-Traps.
The EQing of multiple subs would effectively flatten the spectrum in the room, but it could also create other problems.
It's because it only fixes the steady state spectral density, but it doesn't fix the group delays.
In other words, the resonant frequencies of the room will have the right amplitude if they are excited long enough, but they will be lagging.
This also implies that you'll not reach nominal level with short stimuli (like a wavelet) because you eqed them down at the source.
Only a clean wave propagation will fix all the issues at once, and that's what DBA and etraps are offering.
They scoop out the excess energy before it reaches the reflection point. Basically, it's like having no back wall virtually.
Now, there may be some magic I don't see yet in the multisub system. I see the logic but not sure yet about group delays. I'm still diggin the literrature (found the Welti studies).
MSO is tingling my curiousity. But I don't believe it is an equivalent solution to a DBA or E-Traps.
The EQing of multiple subs would effectively flatten the spectrum in the room, but it could also create other problems.
It's because it only fixes the steady state spectral density, but it doesn't fix the group delays.
In other words, the resonant frequencies of the room will have the right amplitude if they are excited long enough, but they will be lagging.
This also implies that you'll not reach nominal level with short stimuli (like a wavelet) because you eqed them down at the source.
Only a clean wave propagation will fix all the issues at once, and that's what DBA and etraps are offering.
They scoop out the excess energy before it reaches the reflection point. Basically, it's like having no back wall virtually.
Now, there may be some magic I don't see yet in the multisub system. I see the logic but not sure yet about group delays. I'm still diggin the literrature (found the Welti studies).
Last edited:
Alright, I stand corrected and I see now how the mode cancelation works in a multisub system:
Cancellation is explained in https://audioroundtable.com/misc/Welti_Multisub.pdf 2.2
Cancellation is explained in https://audioroundtable.com/misc/Welti_Multisub.pdf 2.2
