Alright.
So, the Fazenda paper is talking about C.A.B.S in section 1.5.
AES E-Library >> Subjective Preference of Modal Control Methods in Listening Rooms
It also mention the Genelec patent of the year 2000. This
patent is why the O800 ARAM from K&H had to be
discountinued.
https://patentimages.storage.googleapis.com/2e/d3/fc/2d0691d7d92983/WO1999066492A1.pdf
DBAs are exactly this.
Double bass array - Wikipedia
Here is the paper from Welti about sub bass optimization. https://audioroundtable.com/misc/Welti_Multisub.pdf
Read it ! It's not based on the same principle. Welti is disturbing the standing waves by cleverly placing the subs, which is equalizing the pressure in the room instead of ensuring a perfect propagation.
It only works for the first two modes with 3 subwoofer.
So, the Fazenda paper is talking about C.A.B.S in section 1.5.
AES E-Library >> Subjective Preference of Modal Control Methods in Listening Rooms
It also mention the Genelec patent of the year 2000. This
patent is why the O800 ARAM from K&H had to be
discountinued.
https://patentimages.storage.googleapis.com/2e/d3/fc/2d0691d7d92983/WO1999066492A1.pdf
DBAs are exactly this.
Double bass array - Wikipedia
Here is the paper from Welti about sub bass optimization. https://audioroundtable.com/misc/Welti_Multisub.pdf
Read it ! It's not based on the same principle. Welti is disturbing the standing waves by cleverly placing the subs, which is equalizing the pressure in the room instead of ensuring a perfect propagation.
It only works for the first two modes with 3 subwoofer.
I explained my thoughts on this very early in the thread. If I'm not mistaken, this is the missing link.. As I see it, DBA and multi-sub are like a Rubik's Cube. DBA is like completing the puzzle, and multi-subs is like leaving it mixed up but changing your perspective so it appears completed.
Yeah. It's a tradeoff between tightness and pressure level.
No but it has other problems that DBAs are solving when they apply.
I'm sure you can achieve great results with MSO. But I will still prefer a planar wave propagation without reflection in my case. Use 4 sealed subs and you'll get some insanely short group delays.
In the end the right solution is the one you can afford and execute in your room. I'm aiming at mastering so yeah... Gotta do what you gotta do.
Loosing some efficiency with a DBA setup doesn't need to be a problem in practice. In a room which would typically be boomy in the bass, = lots of room gain and pressure gain below the lowest room mode, a DBA setup can be really useful. The room itself can make up for the "uneconomical SPL-use" with DBA while the DBA makes the bass more even in larger areas of the room.
Lazy as I am, I copy of a post from another forum:
"Depending on your room surfaces and surface areas of leaky windows and doors, you may notice an increase in SPL for frequencies below the lowest room mode. It depends. In a hermetically sealed bomb shelter in theory +12 dB / octave downwards. If in a tent 0 dB / octave as all the bass will just escape and there will be no pressurization of the room. In a “normal room” it will be somewhere in between 0 and +12 dB / octave. The SPL of these lower frequencies will theoretically be the same at all places in the room, not varying in SPL as you notice with room modes from one place to another.
My room is in the basement, a “shoebox” half of it below ground level and all 6 surfaces in concrete. I wanted to avoid the expected boomy bass and counter the room gain with subs starting to fall off quite high in frequency. The simulation diagram for the sealed DIY-subs at the front wall shows an F3 at 118 Hz (about 117 dB in the diagram) and tuned to about 49 Hz (FB). At 10 Hz in free air the SPL is down to about 91 dB in the simulation, a drop of about 29 dB from max SPL.
The REW-diagram shows actual measurements in the room with 18” subs at front wall and 15” subs at rear wall. Subs are fed the same signal through a Y-splitter. Red curve is with the subs in phase, blue curve has the rear subs with reversed polarity. There is no EQ or time delay applied to the subs, those are “non-fixed” curves. The room is 7,86 m / 25,8 feet long, theoretically I should have the lowest room mode at 21,8 Hz. Microphone is at the middle of the room (50,3% to be exact) and as expected one gets a huge peak for the 2nd length mode with subs in phase (red curve) and a very deep dip with subs having reversed polarity (blue curve). A few hertz below the lowest room mode frequency the SPL starts to increase because of room pressurization. There is a bump at 20 Hz for the blue curve as the front subs are more potent.
The 40 Hz peak can be fixed with EQ but if the rear subs are in reversed polarity and also suitably time delayed versus front subs + some gain adjustments, that would even out all length modes (in theory at least). If the calibration data of the UMIK 1 microphone is correct, it measures +/- 0 dB at 20 Hz and 6 dB too low value at 10 Hz. So, with the red curve SPL is at about 94 + 6 = 100 dB at 10 Hz, close to 9 dB louder than at F3 / 118 Hz. Not some 29 dB lower as from the simulation, a 38 dB difference between measurements and simulation … As I see it, your room surfaces and how the room is built can make a difference for frequencies below the lowest room mode and of course also those between lowest room mode and upwards."
In the measurements the front and rear subs were fed with about 1 W to each array, not much ...
AE's PB 18H+ at the front are about 98 dB /1W / 1 m if I remember correctly. AE's SPB 15 at the rear about 91 dB / 1 W / 1 m. There are 4 subs at each wall. One might think the 40 Hz tone / peak in the red curve lingers on forever but its T60 is actually only about 430 ms according to REW. Even with subs having a much lower efficiency than the AE drivers, in a concrete bunker the needed power from amps could be quite low for a healthy SPL. An "open space" room or one with flimsy stud walls would have different requirements of course.
Lazy as I am, I copy of a post from another forum:
"Depending on your room surfaces and surface areas of leaky windows and doors, you may notice an increase in SPL for frequencies below the lowest room mode. It depends. In a hermetically sealed bomb shelter in theory +12 dB / octave downwards. If in a tent 0 dB / octave as all the bass will just escape and there will be no pressurization of the room. In a “normal room” it will be somewhere in between 0 and +12 dB / octave. The SPL of these lower frequencies will theoretically be the same at all places in the room, not varying in SPL as you notice with room modes from one place to another.
My room is in the basement, a “shoebox” half of it below ground level and all 6 surfaces in concrete. I wanted to avoid the expected boomy bass and counter the room gain with subs starting to fall off quite high in frequency. The simulation diagram for the sealed DIY-subs at the front wall shows an F3 at 118 Hz (about 117 dB in the diagram) and tuned to about 49 Hz (FB). At 10 Hz in free air the SPL is down to about 91 dB in the simulation, a drop of about 29 dB from max SPL.
The REW-diagram shows actual measurements in the room with 18” subs at front wall and 15” subs at rear wall. Subs are fed the same signal through a Y-splitter. Red curve is with the subs in phase, blue curve has the rear subs with reversed polarity. There is no EQ or time delay applied to the subs, those are “non-fixed” curves. The room is 7,86 m / 25,8 feet long, theoretically I should have the lowest room mode at 21,8 Hz. Microphone is at the middle of the room (50,3% to be exact) and as expected one gets a huge peak for the 2nd length mode with subs in phase (red curve) and a very deep dip with subs having reversed polarity (blue curve). A few hertz below the lowest room mode frequency the SPL starts to increase because of room pressurization. There is a bump at 20 Hz for the blue curve as the front subs are more potent.
The 40 Hz peak can be fixed with EQ but if the rear subs are in reversed polarity and also suitably time delayed versus front subs + some gain adjustments, that would even out all length modes (in theory at least). If the calibration data of the UMIK 1 microphone is correct, it measures +/- 0 dB at 20 Hz and 6 dB too low value at 10 Hz. So, with the red curve SPL is at about 94 + 6 = 100 dB at 10 Hz, close to 9 dB louder than at F3 / 118 Hz. Not some 29 dB lower as from the simulation, a 38 dB difference between measurements and simulation … As I see it, your room surfaces and how the room is built can make a difference for frequencies below the lowest room mode and of course also those between lowest room mode and upwards."
In the measurements the front and rear subs were fed with about 1 W to each array, not much ...
AE's PB 18H+ at the front are about 98 dB /1W / 1 m if I remember correctly. AE's SPB 15 at the rear about 91 dB / 1 W / 1 m. There are 4 subs at each wall. One might think the 40 Hz tone / peak in the red curve lingers on forever but its T60 is actually only about 430 ms according to REW. Even with subs having a much lower efficiency than the AE drivers, in a concrete bunker the needed power from amps could be quite low for a healthy SPL. An "open space" room or one with flimsy stud walls would have different requirements of course.
DBA's can certainly measure better.
But I have an L shaped room, so DBA's just don't work for me (I've tried).
Yeah, the Welti/MSO approach looks like a perfect compromise in your case.
I have seen published documents from Bruno Facenda et al, where CABS / DBA was shown to work well also in L-shaped rooms. Speakers were not directly on opposite walls then, as in a shoebox room. So it might work also in L-shaped rooms, Depends on practicalities of course, best place for the rear speakers might end up where you certainly don't want them to be in a living room.
I wouldn't readily accept that because the pseudo-random nature of the multi-sub approach means the result will be smooth if there are a sufficient number of sources, but we usually stop at between 2 and 4 extra subs as it is good enough.
Geddes original plan begins with comprehensive passive damping, and it was said that this is the first line defence and the more, the better.. I'm not sure it's mandatory, multi-subs do their own damping and all within the steady state formation time.
The principle is exactly this yes, but you need to account for phase rotation and frequency response of your mic and subwoofer. So you need at least some filters on the mic signal path, and they need to be tuned such that the system emulates the impedance of air up to a certain frequency.
Last edited:
Interesting, what have you noticed with this. It would be good to have information on matters such as what to expect with gain, and considering it might vary with frequency and so forth.
Well the first actual limit is purely physical and related to distance.
An ABT is a pressure device. It's nullifying whatever pressure it measures in front of it.
If the pressure rises, the cone will go in. If it drops, the cone will go forward. And it does so until the pressure is ideally zero.
Unfortunately, an ABT propagates its own wave, meaning it will nullify the pressure up to a certain frequency vs distance. Your counteracting wavefront needs to reach the offending wavefront more or less in phase. So the system effectiveness can be described as a sphere that diminishes in diameter as you raise your highest target frequency.
If your correcting wave from the ABT doesn't reach an offending wave, then you just put more energy in your room and you will get some 'echo' coming from the ABT on the back.
In conclusion, if your room is big, you'll either need more ABT placed in the room, or you'll have to limit yourself to lower frequencies. A simple way to guess how much you need is to design the ABTs placement such that if they were used a subwoofer they would propagate a planar wavefront.
Your gain and filter settings (can be an EQ) will be determined by your mic sensitivity and response, speaker sensistivity and response... etc. Basically, you want that system to be a servo subwoofer with a perfectly flat response, and you feed the offending pressures as an error signal to that system.
It's quite a complicated system to make and maybe not in the reach of intermediate DIYers if you really want to do it right. You may need a custom DSP or custom filter board in the signal path, the mic need to have a good response or pass through a deconvolution etc...
DBAs are still the best solution if your room is rectangular. MSOs are a good compromise with some opensource software readily available.
An ABT is a pressure device. It's nullifying whatever pressure it measures in front of it.
If the pressure rises, the cone will go in. If it drops, the cone will go forward. And it does so until the pressure is ideally zero.
Unfortunately, an ABT propagates its own wave, meaning it will nullify the pressure up to a certain frequency vs distance. Your counteracting wavefront needs to reach the offending wavefront more or less in phase. So the system effectiveness can be described as a sphere that diminishes in diameter as you raise your highest target frequency.
If your correcting wave from the ABT doesn't reach an offending wave, then you just put more energy in your room and you will get some 'echo' coming from the ABT on the back.
In conclusion, if your room is big, you'll either need more ABT placed in the room, or you'll have to limit yourself to lower frequencies. A simple way to guess how much you need is to design the ABTs placement such that if they were used a subwoofer they would propagate a planar wavefront.
Your gain and filter settings (can be an EQ) will be determined by your mic sensitivity and response, speaker sensistivity and response... etc. Basically, you want that system to be a servo subwoofer with a perfectly flat response, and you feed the offending pressures as an error signal to that system.
It's quite a complicated system to make and maybe not in the reach of intermediate DIYers if you really want to do it right. You may need a custom DSP or custom filter board in the signal path, the mic need to have a good response or pass through a deconvolution etc...
DBAs are still the best solution if your room is rectangular. MSOs are a good compromise with some opensource software readily available.
Hello PureBasic and all. Thanks for your comprehensive posts. Please explain the acronyms ABT, DBA, and MSO. I continue to be interested and practice the suppression of standing waves.
Best wishes
Anton