Re: Re: Good articles about dipoles, and room acoustics
I agree with your first statement.
But the second makes me think a bit.
Maybe it is that a lot of "normal" home listening rooms are actually fairly similar in the reverb vs direct ratio at certain frequencies, UNLESS the room is an "audiophile" room which has been specifically treated to alter the ratio.
hmmmmmm
???.
john k... said:
I agree with your first statement.
But the second makes me think a bit.
Maybe it is that a lot of "normal" home listening rooms are actually fairly similar in the reverb vs direct ratio at certain frequencies, UNLESS the room is an "audiophile" room which has been specifically treated to alter the ratio.
hmmmmmm
???.
jlo said:
Wow! Unbeliveable! Thanks , Jl OHl - ca c'ect un sarce cadeux que tu viens me faire. 1000x Merci!
I had forgotten how odd these recordings sound. Very strange. But they really do point up room acoustics and the FR of the speakers like no other test I've heard. The artificial reverb is fun. The imitation of Boston Symphony hall is right on. Never been there, but have heard plenty of BSO recordings. Love that sound.
I think that most on this thread have ignored thess posts about the Denon CD. You are wrong to do so. I suggest downloading all the tracks and the liner notes. This test CD is no magic bullet, but there is much to be learned from it. And it even gives you an SPL reference! Too cool. Where else will you find a recording with NO acoustics?
Merci encore.
How to measure a driver for use as a dipole
Hi
It seems that it isn't that easy to obtain good plots at the low frequency end of OB measurements
Below a close mic ( 1 cm ) in-room measurement plot of my 30 cm by 46 cm test-OB – woofer only
Below a 1 m mic distance in-room measurement plot of my 30 cm by 46 cm test-OB – woofer only
Below how it SHOULD look like according to edge ( at 1m mic distance ) :
Below how it SHOULD look like according to A, B, C, Dipole ( at 1m mic distance ) :
JohnK has some good papers on his site about measuring techniques and in room behaviour but everything is of no value if the initial measurements aren't correct.
Regarding the above I hardly can find ANYTHING that correlates between simulation and measurement.
Well this thread is also about making faults in the public – as pointed out somewhere else...
But I really wish I cold do better !
-------------
jlo, thanks for the links.
I was always looking for some "dead" recordings.
You already know CARA? It also is capable of convoluting wav-files with simulated rooms.
Greetings
Michael
Hi
It seems that it isn't that easy to obtain good plots at the low frequency end of OB measurements
Below a close mic ( 1 cm ) in-room measurement plot of my 30 cm by 46 cm test-OB – woofer only
An externally hosted image should be here but it was not working when we last tested it.
Below a 1 m mic distance in-room measurement plot of my 30 cm by 46 cm test-OB – woofer only
An externally hosted image should be here but it was not working when we last tested it.
Below how it SHOULD look like according to edge ( at 1m mic distance ) :
An externally hosted image should be here but it was not working when we last tested it.
Below how it SHOULD look like according to A, B, C, Dipole ( at 1m mic distance ) :
An externally hosted image should be here but it was not working when we last tested it.
JohnK has some good papers on his site about measuring techniques and in room behaviour but everything is of no value if the initial measurements aren't correct.
Regarding the above I hardly can find ANYTHING that correlates between simulation and measurement.
Well this thread is also about making faults in the public – as pointed out somewhere else...
But I really wish I cold do better !
-------------
jlo, thanks for the links.
I was always looking for some "dead" recordings.
You already know CARA? It also is capable of convoluting wav-files with simulated rooms.
Greetings
Michael
Hi Michael,
A couple of things: Yes it is not going to be easy to get in room measurements of a dipole woofer, or any other type woofer system for that matter, at least if you are looking for quasi-anechoic response. If you just want in room measurements then stick a mic out and what you see is what you have.
Second, the simulation tools (The Edge and my ABC Dipole worksheet) are based on ideal drivers which have flat on axis, 2Pi response with various approaches to modeling driver directionality. Although, in my worksheet you can input the drivers T/S parameters to see how the native roll off affect the low frequency response. They don't account for the irregularities in the driver's native response at higher frequency. In that regard they must be viewed as a corrections to a driver’s 2Pi response, just as any normal baffle simulation tool. I know that ABC Dipole’s baffle tool is in good agreement with, for example the BDS at the FRDC when used to model a conventional speaker baffle, though I don’t have the high frequency resolution of the BDS in ABC D... (I should also note for those who may be confused that for some reason The Edge place the 0dB level at 6dB below the 2Pi level which is 6dB below the 0dB level in ABC D...)
Near filed measurements are accurate only below roughly C/(Pi*D) and 1M measurements generally aren’t accurate once reflections get into the picture, as I am sure you know. That is why I discuss the near field/1M merge technique to obtain full range quasi-anechoic data for design. You should be able to get good 1M data above 200 or 300 Hz and the near filed data should be accurate below that.
As you know, in ABC dipole the baffle tool is only for ESTAMATING baffled driver response. It yields an accurate representation of the anechoic response at low frequency and can be used for woofer system design with confidence. However, when the midrange comes into play I recommend building a test baffle base on the simulation results and then importing the actually measured SPL (from the merged 1M/near filed dipole response) for design of the equalization and hp pass midrange filter as shown here
The bottom line is that, at least for ABC Dipole, if you design a woofer based on the worksheet and low pass filter it as intended, about 1 octave below the dipole peak or lower, then you can expect it to perform as indicated. But for a midrange driver, where response is used above the dipole predicted peak, the baffle tools are only suitable for evaluation of the low frequency response and estimation of what irregularities the baffle may introduce. There are too many unknowns for these tools to be highly accurate at higher frequency. Driver directional effects can be estimated but the bigger problem is the asymmetry between the front and rear driver response. That is too driver dependent to be reasonable modeled.
What I think would be of more interest is what your 1 M response is when windowed to eliminate reflections. This should give you reasonable data above 200 or 300 Hz. That would be a reasonable comparison to The Edge or ABC D...
FYI, in the ABC plot you presented dark blue is on axis, yellow is an equivalent (at low frequency) circular baffle, green dashed is off axis at what ever angle you specified, maroon dashed is 90 degrees.
A couple of things: Yes it is not going to be easy to get in room measurements of a dipole woofer, or any other type woofer system for that matter, at least if you are looking for quasi-anechoic response. If you just want in room measurements then stick a mic out and what you see is what you have.
Second, the simulation tools (The Edge and my ABC Dipole worksheet) are based on ideal drivers which have flat on axis, 2Pi response with various approaches to modeling driver directionality. Although, in my worksheet you can input the drivers T/S parameters to see how the native roll off affect the low frequency response. They don't account for the irregularities in the driver's native response at higher frequency. In that regard they must be viewed as a corrections to a driver’s 2Pi response, just as any normal baffle simulation tool. I know that ABC Dipole’s baffle tool is in good agreement with, for example the BDS at the FRDC when used to model a conventional speaker baffle, though I don’t have the high frequency resolution of the BDS in ABC D... (I should also note for those who may be confused that for some reason The Edge place the 0dB level at 6dB below the 2Pi level which is 6dB below the 0dB level in ABC D...)
Near filed measurements are accurate only below roughly C/(Pi*D) and 1M measurements generally aren’t accurate once reflections get into the picture, as I am sure you know. That is why I discuss the near field/1M merge technique to obtain full range quasi-anechoic data for design. You should be able to get good 1M data above 200 or 300 Hz and the near filed data should be accurate below that.
As you know, in ABC dipole the baffle tool is only for ESTAMATING baffled driver response. It yields an accurate representation of the anechoic response at low frequency and can be used for woofer system design with confidence. However, when the midrange comes into play I recommend building a test baffle base on the simulation results and then importing the actually measured SPL (from the merged 1M/near filed dipole response) for design of the equalization and hp pass midrange filter as shown here
An externally hosted image should be here but it was not working when we last tested it.
The bottom line is that, at least for ABC Dipole, if you design a woofer based on the worksheet and low pass filter it as intended, about 1 octave below the dipole peak or lower, then you can expect it to perform as indicated. But for a midrange driver, where response is used above the dipole predicted peak, the baffle tools are only suitable for evaluation of the low frequency response and estimation of what irregularities the baffle may introduce. There are too many unknowns for these tools to be highly accurate at higher frequency. Driver directional effects can be estimated but the bigger problem is the asymmetry between the front and rear driver response. That is too driver dependent to be reasonable modeled.
What I think would be of more interest is what your 1 M response is when windowed to eliminate reflections. This should give you reasonable data above 200 or 300 Hz. That would be a reasonable comparison to The Edge or ABC D...
FYI, in the ABC plot you presented dark blue is on axis, yellow is an equivalent (at low frequency) circular baffle, green dashed is off axis at what ever angle you specified, maroon dashed is 90 degrees.
One other option for analyzing the low frequency response of OB speaker systems is my MathCad worksheets. On my site, you can look at sample pdf outputs and compare them to the other tools available. The worksheets include the driver T/S data, a crossover and boost options found in active crossovers, the baffle size and shape geometry, and the floor and rear wall reflections.
ABC Dipole also allows the user to design the woofer equalization and LP filter including calculation of the circuit component values for active circuits. It doesn't include room effects though. I don't believe that room related frequency response anomalies should be accounted for in the basic equalization circuits since they will be dependent on both the woofer and listening positions. But it is of interest to see how these factors affect the response. I use SoundEasy’s finite element room analysis to aid in that aspect of listening room setup. Of course, measuremnts tell the real story.
John,
If GD associated with higher order crossover at low frequency is audible
(1) How do you explain that the NaO also uses high order crossover between mids and woofer and they are crossed at fairly low, not higher than the Orion?
(2) What frequency range do you define as "low frequency" here with regards to the GD associated with higher order crossover?
(3) What is the solution?
I am asking these questions because I am learning to design and build a WWMTMWW+SubSub OB 4 way hybrid active speakers and the single, remaining big difficulty I have is the crossover order between the mids and woofers (at 200Hz), and between the woofers and the subs (at 70Hz). I have been worrying about the audibility of GD. I would use a lower order XO if I could. I have modelled extensively with the driver response files using a number of software including your ABC Dipole and derived a perfectly tailored acoustic Butterworth 3rd order XO at 200Hz that having taken into account the driver and baffle responses and should theoretically work well. But checking the electrical response it is very close to but not even a 2nd order response. We know that 2nd order electrical response would merely keep the cone excursion constant below the XO point, so I expect that the IMD will be significantly higher with such a network! If I want to go higher than 2nd order electrical, the lowest acoustic order would be 4th.
For the time being, I have derived the circuit to shape the driver/baffle response to be perfect acoustic LR4 response for both XOs in a hope that IMD / non linear distortions are a lot more audible than GD / linear distortions. But I am worried about this assumption.
Would you please shed some lights on this? Other experts also please contribute., Lynn, MBK, Jussi, ...
I hope this is not out of the topic. One thing with OB we need to consider is that the driver may usually roll off naturally with a first order, and OB rolls off at 6dB per octave, so within an octave in the XO region, the roll off can be 2nd order already without EQ. We need to give the driver some protection and need to reduce excursion / IMD. So how we can achieve the goal with a low order XO to reduce the group delay?
Regards,
Bill
I have issues with a WWMTMWW layout - specifically, I've yet to hear a WMTMW that I thought had realistic bass. I know it's a popular fad in the ultra high-end sector for some kooky reason, but it's never made any sense to me.
Maybe it's just me, but I think bass coming from a location well above listening height sounds really weird and unnatural. How many bass instruments have acoustic emission from above the listening position? Not many. Bass usually comes from a location close to the floor, from bass viols to bass guitar to pianos to kettledrums. This is hardly surprising, since they can take advantage of the proximity to the floor to get better room coupling.
It's already a major PITA to design something as simple as the MTM of the Ariel thanks to wacko vertical radiation patterns - I can't see WMTMW, or worse, WWMTMWW, being any easier.
At the minimum, the upper bass array is at a different distance to the ceiling than the lower bass array is to the floor, and I don't see that as an advantage, but another source of unpredictable room interaction - particularly of response curves that are strongly distance-dependent. In other words, I would bet that the WWMTMWW measures quite differently at 1, 2, and 3 meters, due to size and proximity to asymmetric floor/ceiling boundaries alone.
Integrating a 3 or 4-way system is very, very difficult, even with a lot of simulation and measurement tools that are ready to hand. Maybe I'm on the slow side, but the "simple" 2-way Ariel took 6 months of twiddling and re-measuring to sound subjectively flat on pink-noise and music.
I've also designed traditional TMW 3-ways, and the subjective system integration takes several times longer - the 100 to 1000 Hz region for the crossover is awkward, since it strongly interacts with the room, and group-delay effects are more audible in this region (the wavelengths are pretty large, after all).
If you must design an MTM, John K's approach makes the most sense - the trick is to get the baffle width to exactly compensate for the vertical directivity of the MTM in the crossover region. Tricky enough to do at one crossover frequency, much less the vertical radiation-pattern complexity of a WWMTMWW.
Maybe it's just me, but I think bass coming from a location well above listening height sounds really weird and unnatural. How many bass instruments have acoustic emission from above the listening position? Not many. Bass usually comes from a location close to the floor, from bass viols to bass guitar to pianos to kettledrums. This is hardly surprising, since they can take advantage of the proximity to the floor to get better room coupling.
It's already a major PITA to design something as simple as the MTM of the Ariel thanks to wacko vertical radiation patterns - I can't see WMTMW, or worse, WWMTMWW, being any easier.
At the minimum, the upper bass array is at a different distance to the ceiling than the lower bass array is to the floor, and I don't see that as an advantage, but another source of unpredictable room interaction - particularly of response curves that are strongly distance-dependent. In other words, I would bet that the WWMTMWW measures quite differently at 1, 2, and 3 meters, due to size and proximity to asymmetric floor/ceiling boundaries alone.
Integrating a 3 or 4-way system is very, very difficult, even with a lot of simulation and measurement tools that are ready to hand. Maybe I'm on the slow side, but the "simple" 2-way Ariel took 6 months of twiddling and re-measuring to sound subjectively flat on pink-noise and music.
I've also designed traditional TMW 3-ways, and the subjective system integration takes several times longer - the 100 to 1000 Hz region for the crossover is awkward, since it strongly interacts with the room, and group-delay effects are more audible in this region (the wavelengths are pretty large, after all).
If you must design an MTM, John K's approach makes the most sense - the trick is to get the baffle width to exactly compensate for the vertical directivity of the MTM in the crossover region. Tricky enough to do at one crossover frequency, much less the vertical radiation-pattern complexity of a WWMTMWW.
How to measure a driver for use as a dipole
Hi
Right, " sticking out a mic and what you see is what you have." is also what is done in PRO (takening 4 multiplexed mic's placed at specific places in the room for THX measurement for example )
In post 797
http://www.diyaudio.com/forums/showthread.php?postid=1214299#post1214299
you can see the same speaker in the same test baffle measured and equalised flat this way - but that is not what I am aiming for right now.
What can be seen in post 1684
http://www.diyaudio.com/forums/showthread.php?postid=1274644#post1274644
is that for the 1M measurement the simulation predicts at least the slope below lets say 300 Hz more or less correct.
BUT with close mic technique, simulation and measurement is COMPLETELY different at the low frequency end if you compare the pictures below ( EDGE and ABC-Dipole at 5 cm distance, measurement at 1 cm )
With close mic-ing I would have expected NO difference at all as room interaction should be neglectible.
Any explanations for that?
Greetings
Michael
Hi
Right, " sticking out a mic and what you see is what you have." is also what is done in PRO (takening 4 multiplexed mic's placed at specific places in the room for THX measurement for example )
In post 797
http://www.diyaudio.com/forums/showthread.php?postid=1214299#post1214299
you can see the same speaker in the same test baffle measured and equalised flat this way - but that is not what I am aiming for right now.
What can be seen in post 1684
http://www.diyaudio.com/forums/showthread.php?postid=1274644#post1274644
is that for the 1M measurement the simulation predicts at least the slope below lets say 300 Hz more or less correct.
BUT with close mic technique, simulation and measurement is COMPLETELY different at the low frequency end if you compare the pictures below ( EDGE and ABC-Dipole at 5 cm distance, measurement at 1 cm )
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
With close mic-ing I would have expected NO difference at all as room interaction should be neglectible.
Any explanations for that?
Greetings
Michael
Michael,
Please post Sd and the T/S parameters for your woofer. I don't know exactly what you are running in ABC but there is obviously a problem as indicated by the glitch at 40 Hz. The Edge just doesn't work right at small mic distances.
He is a plot of the ABC result for a peerless 10" XLS (Sd = 333, Qts = 0.17, Fs = 18 Hz) centered on a 30 x 46 cm baffle at 5 cm (Dark blue trace).
The off axis results aren't meaningful at 5 cm and the on axis response is only valid below about 500 Hz. If you input the correct T/S parameters there is no way ABC will give you anything but a 2nd order HP response consistent with those T/S parameters as the distance approaches zero.
Please post Sd and the T/S parameters for your woofer. I don't know exactly what you are running in ABC but there is obviously a problem as indicated by the glitch at 40 Hz. The Edge just doesn't work right at small mic distances.
He is a plot of the ABC result for a peerless 10" XLS (Sd = 333, Qts = 0.17, Fs = 18 Hz) centered on a 30 x 46 cm baffle at 5 cm (Dark blue trace).
An externally hosted image should be here but it was not working when we last tested it.
The off axis results aren't meaningful at 5 cm and the on axis response is only valid below about 500 Hz. If you input the correct T/S parameters there is no way ABC will give you anything but a 2nd order HP response consistent with those T/S parameters as the distance approaches zero.
HiFiNutNut said:
Good questions. First, there are always trade offs. Both the Orion and the NaO cross to the mid at about 125 Hz. The Orion uses a 4th order electrical LR crossover. The NaO II a modified B3 electrical. As you point out, with OB we do have some limitations. The minimum acoustic slope of the HP crossover to the mid needs to be 3rd order to protect the midrange from having the excursion continuing to increase (for a constant signal level) below the crossover point. So we are pretty much stuck with that as a minimum. However, I wasn't really refering to the mid to woofer crossover. What I was referring to is the addition of a subwoofer.
If you look at the GD of a woofer system with Qtc = 0.5, Fs = 20 and a 4th order 125 Hz LP crossover the response and GD look like this:
When you add a subwoofer with the same Qtc = 0.5 Fs = 20 Hz and a 50 Hz x-o between the main woofer and the sub the response and GD look like this:
The amplitude response remains the same but the GD is significantly increased. The addition of the sub increases the max SPL capability at the expense of more GD at low frequency. There is increasing evidence that this is audible (see the latest edition of High Performance Loudsepakers by Martin Cooloms).
The solution I used for the NaO II is the u-frame woofer system and no sub. The u-frames offer greater max SPL capability than H-frames to start with, as you know, (for the same driver and same physical size), and for those situations where even greater SPL is required at low frequency the NaO II woofer system converts to a sealed box system by installing a rear panel on the u-frames and flipping a switch on the crossover to change the woofer equalization. So you have a total 1332 cm^2 for Sd at low frequency which is 46% more Sd than two 12" XLS woofers. In the Mini I just opted for the sealed box woofer system given the other consideration discussed at my site for the Mini. The 1st gen NaO (or NaO P) does not have the convertable woofer system.
However, going back the the mid/woofer crossover. If you want to reduce the GD then approach I would take is to use a 3rd order acoustic crossover between mids and woofer with inverted polarity. The total GD through the corssover will then be the same as if a 2nd order acoustic crossover was used and you will maintain protection from increasing excursion. The traded off here is that you won't preserve absolute polarity in either the mid or woofer and while the mid would be subject to increasing excursion it will still have greater excursion below the x-o that it would with a 4th order HP. That may effect the HD and IMD of the system at higher SPL.
I know that children should be seen and not heard,so I hope that the noise of this post won't offend the sensitive ears of the Audio Brahmins and other deep thought producers too much:
It seems that the underline "problematic" of various posts scattered here and there throughout this thread, relate in one way or another to the business of re-creating the original "stereo effect" of the recording.Difficulty and dissatisfaction seems to be the order of the day.
Makes me wonder why we bother with it ? Especially when the "stereo effect" is nothing other than a dis-embodied apparition of music. I suppose the goal, through some sort of necromancy, is to re-embody these apparitions; make them "real", "palpable". Will we ever achieve full satisfaction ? I doubt it.......Well at least not until we ourselves become phantoms.
Cilla
It seems that the underline "problematic" of various posts scattered here and there throughout this thread, relate in one way or another to the business of re-creating the original "stereo effect" of the recording.Difficulty and dissatisfaction seems to be the order of the day.
Makes me wonder why we bother with it ? Especially when the "stereo effect" is nothing other than a dis-embodied apparition of music. I suppose the goal, through some sort of necromancy, is to re-embody these apparitions; make them "real", "palpable". Will we ever achieve full satisfaction ? I doubt it.......Well at least not until we ourselves become phantoms.
Cilla
How to measure a driver for use as a dipole
Hi
The driver is a modified Dynaudio 21W54 with ( originally )
Qts = 0.3
Fs = 30 Hz
Sd = 220 cm2
This data ( and some other usefull information about Dynaudio speakers) can be found on
http://www.gattiweb.com/dynaudio.html
But don't stick too much on the data sheets as they are not 100 % reliable as Dynaudio changed " without notice " as others also did....
The material of the cap and also how it was glued to the membrane ( and to the voice coil in one case only ) for example are quite different for the two samples I have.
Also some of the old German data sheets I have are quite different...
Well, I guess the really big problem here was mostly ME MYSELF and I , as I was NOT used to include also the correct T/S speaker parameters for simulation ( I have left the defaults – shame on me ).
Here are the updated results that show MUCH better correlation between measurement and simulation - below around 300 Hz where the driver can be assumed to work piston-like.
AND does not feature T/S parameters, meaning simulation is for a speaker of Fs << 1 Hz ( flat to 0 Hz ) what isn't very much what we have in real world IF we have some especial interested to the low frequency end .
And just for the record the 1M simulation updated with the correct T/S parameters looks like below and also has better correlation in the low frequencies now.
JohnK, thanks for correcting me in my weak point.
Greetings
Michael
Hi
The driver is a modified Dynaudio 21W54 with ( originally )
Qts = 0.3
Fs = 30 Hz
Sd = 220 cm2
This data ( and some other usefull information about Dynaudio speakers) can be found on
http://www.gattiweb.com/dynaudio.html
But don't stick too much on the data sheets as they are not 100 % reliable as Dynaudio changed " without notice " as others also did....
The material of the cap and also how it was glued to the membrane ( and to the voice coil in one case only ) for example are quite different for the two samples I have.
Also some of the old German data sheets I have are quite different...
Well, I guess the really big problem here was mostly ME MYSELF and I , as I was NOT used to include also the correct T/S speaker parameters for simulation ( I have left the defaults – shame on me ).
Here are the updated results that show MUCH better correlation between measurement and simulation - below around 300 Hz where the driver can be assumed to work piston-like.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
AND does not feature T/S parameters, meaning simulation is for a speaker of Fs << 1 Hz ( flat to 0 Hz ) what isn't very much what we have in real world IF we have some especial interested to the low frequency end .
And just for the record the 1M simulation updated with the correct T/S parameters looks like below and also has better correlation in the low frequencies now.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
JohnK, thanks for correcting me in my weak point.
Greetings
Michael
Hi Michael,
That look better. I'd still like to see a 1M measurments with 3 to 5 msec FFT window. I'm sure there will be differences but my spread sheet includes an imperical model of the efffect of asymmetry between front and rear radiation which I based on a variety of my own measurments for a number of different drivers. This is, for example, why the 90 degree off axis response doesn't go to identically zero at all frequencies. Anyway, if you have the time to do a windowed 1 M measurement to get reflections out of the picture I'd be interested to see it.
Maybe also measure the T/S parameters of your driver. As I said, if you design a flat baffle dipoel, H or U frame woofer system based on ABC, other than the H or U frame resonance it's pretty much what you can expect to get anechoically (assuming you correctly damp the U frame). It's all spelled out in the interactive design guide that comes with the spreadsheet.
That look better. I'd still like to see a 1M measurments with 3 to 5 msec FFT window. I'm sure there will be differences but my spread sheet includes an imperical model of the efffect of asymmetry between front and rear radiation which I based on a variety of my own measurments for a number of different drivers. This is, for example, why the 90 degree off axis response doesn't go to identically zero at all frequencies. Anyway, if you have the time to do a windowed 1 M measurement to get reflections out of the picture I'd be interested to see it.
Maybe also measure the T/S parameters of your driver. As I said, if you design a flat baffle dipoel, H or U frame woofer system based on ABC, other than the H or U frame resonance it's pretty much what you can expect to get anechoically (assuming you correctly damp the U frame). It's all spelled out in the interactive design guide that comes with the spreadsheet.
