Double bass array. Low Q subs. [Quasi]plain wave radiation?

[Flaesh]

Hi there!
Briefly:
Here is the room 14'W*30'L*10H. Speakers not shown as their location may vary. Subs (2F+2R) placed as expected at half of room height.

Here is measured LF response approximately at "sweet spot". As always (unless otherwise specified) no smooting, without eq\correction, with long default window.

Subs are closed 200 l with conventional old fashioned 18" (Fs 29. Qts .29 Vas ~400+ l):

One can say it's not sub at all and yes, for open space this is true. But I usually live at home.
Let's enter the value "0.0 x Pi" for Ang:

and funny acoustical impedance

I'd love to hear @René - Acculution.com 's comments on the above.

 

[René - Acculution.com]

@Flaesh The DBA concept is interesting. I have listened to it in action when working with Adrien years back (https://vbn.aau.dk/en/publications/controlled-acoustic-bass-system-cabs-a-method-to-achieve-uniform-), and you could certainly hear a difference when moving around in the room. I think it is more ideal to have 4F+4R, but placement is not easy then. I imagine you have all 4 on the floor (?). So in a true plane wave you have a specific acoustic impendance of density times sound speed, but the question is of course to which degree you can have a plane wave in room like this. One tries with the DBA/CABS to turn the room into a tube in some sense, and if you excite a plane wave in a tube, then that is the only wave you will have at all frequencies (even above any cut-on frequency for the transverse modes). From you measurement is seems that you have at least a one point in the room where the (specific) acoustic impedance is constant at all frequencies (?), but above that a figure (where I would need a color legend to see the values in the room) where the impedance varies with position. As you have a plane wave in a "tube", you will not experience any room gain, or modes in general for an ideal plane wave, and so your speakers will dictate the roll-off for the frequency response in a point. Leakage in the room will however affect the sound pressure as your 'tube' needs hard walls for all of this to work ideally.

 

[Flaesh]

Good ol' nice sweet pleasant discussion about *BA vs multisubs with Nils and Earl https://www.diyaudio.com/forums/multi-way/280249-controlled-directivity-diys-5.html#post4466544
I have not experimented with sub's configurations in the current room, only DBA; in the previous room 14'W*16'L*10H (rudiments of the transverse wall are visible on the room plan) DBA was subjectivly preferred by uninformed listeners in comparations with another eqalized\delayed\inverted subs. While I don't try four channels, random sub positions and using MSO. Some sence of "front bass" with DBA is good for me, also as nothing like bass in head\separate bass and so on.

I imagine you have all 4 on the floor (?).

No,

Subs (2F+2R) placed as expected at half of room height.

5' from floor, 5' from ceiling. These are not things that are easy to move around. When deviating from 1/4 of the width, the result is the same as the simple model - second width mode appears. Here is REW simple model for empty room without doors\windows:

 

[René - Acculution.com]

Good ol' nice sweet pleasant discussion about *BA vs multisubs with Nils and Earl https://www.diyaudio.com/forums/multi-way/280249-controlled-directivity-diys-5.html#post4466544
I have not experimented with sub's configurations in the current room, only DBA; in the previous room 14'W*16'L*10H (rudiments of the transverse wall are visible on the room plan) DBA was subjectivly preferred by uninformed listeners in comparations with another eqalized\delayed\inverted subs. While I don't try four channels, random sub positions and using MSO. Some sence of "front bass" with DBA is good for me, also as nothing like bass in head\separate bass and so on.

No,

5' from floor, 5' from ceiling. These are not things that are easy to move around. When deviating from 1/4 of the width, the result is the same as the simple model - second width mode appears. Here is REW simple model for empty room without doors\windows:
View attachment 136779

Ah, okay, they are high up. Nice. Yes, I read the post you linked to quite recently. I enjoy these post not just for technical points, but also to see how different experts discuss things; do they get heated and insist on their way being the best, are they open for suggestions, things like that, the psykology of it all ;-) When it comes to "does DBA inherently sound good/bad, natural/unnatual, ..." I really have no insight. I coooould be onboard with the fact that the bass being more 'outdoors-y"/propagating and the rest being "indoors-y" could sound somewhat unnatural to some ears, but I have no idea. I am more into the physics and the mathematics ;-)

 

[Flaesh]

post you linked to quite recently. I enjoy these post not just for technical points, but also to see how different experts discuss things

There are 3 or more pages after in similar manner))

that the bass being more 'outdoors-y"

I agree. It's funny and unexpected that so far none of the guests said: "there is not enough bass" or something like that.

I am more into the physics and the mathematics ;-)

I would like to see a correlation between physics and audible effect . At one time I believed that a wavelet without irregularities characterizes such a bass, but this was a mistake.

 

[Kvalsvoll]

In this room, I would try all 4x subwoofers on front wall, and absorption on rear wall.

Math/model for DBA is very simple - for the perfect situation - and you get perfect frequency response, and a sound field where direction (pv) is predictable and similar to free-space at very large distance, where the wavefront is a plane wave. The problem is, practical rooms do not have rigid wall surfaces and 2x units on one wall is not enough to produce a sufficiently plane wave. And as you approach 100hz and so, the deviation from the model becomes so large, it will be very difficult to tune the system so you avoid hearing the rear sub units.

With 4x and individual processing, you can tune the system using other approaches - experimenting with delay on the rear units are crucial here. the goal should be to achieve a decent frequency response AND timing and preserved sound field intensity/velocity. This can usually be achieved in most rooms with 4x, it does not require a DBA/SBA approach, but it it will require some work on calibration.

 

[Flaesh]

In imperfect rooms, where the upper tens of hertz range is still damaged by openings and furniture, precise positioning of subwoofers may not be all that necessary. See "Subjective Preference of Modal Control Methods in Listening Rooms", Fazenda et al. The name speaks for itself. DBA (CABS) considered and their "SSS" is what called "DBA" by some people)). They also used two different music samples ..
Paper in open access https://www.researchgate.net/public...reference_of_modal_control_in_listening_rooms

In this room, I would try all 4x subwoofers on front wall, and absorption on rear wall.

Yes, I'm going to try sometime. I think you can imagine what effective absorber for say 30-40 hertz looks like . Adding 4 rear subs may be easyer.

 

[René - Acculution.com]

Isn't Geddes' approach different from those mentioned in the paper? If so, it would have been nice to that also...

 

[Kvalsvoll]

I think you can imagine what effective absorber for say 30-40 hertz looks like . Adding 4 rear subs may be easyer.

It can be done, build a helmholtz-wall covering the rear wall, use the online porous-acoustic-absorber calculator to find correct size of openings in the panel. Just doing a simple absorber wall is not so efficient at very low frequencies.

 

[Flaesh]

Geddes' approach

Fazenda et al. are not hardworking enough. As I Undoubtedly it would be interesting to compare with Geddes'\MSO approach. But I think "1.6 Front-Back (FB) Configuration" is something similar in generally.
My friends and I definitely preferred the outdoorish bass in my system with samles like Fazenda's "sample A".
You can consider it as waveform reproduction or so(?))

a helmholtz-wall covering the rear wall

HR are band-limited. Wide range perforated\slotted absorbers aren't very helmholtz-ish. See Bogich Petrovich (RIP) papers.
Nils had rockwool rear wall with SBA and had good results.
One buddy (simple hiend stereo towers) consulted with some "acoustician" build huge helmholtz rear wall. Bad before and slightly less bad after . For hiend-ish price.

 

[Kvalsvoll]

Fazenda et al. are not hardworking enough. As I Undoubtedly it would be interesting to compare with Geddes'\MSO approach. But I think "1.6 Front-Back (FB) Configuration" is something similar in generally.
My friends and I definitely preferred the outdoorish bass in my system with samles like Fazenda's "sample A".
You can consider it as waveform reproduction or so(?))

HR are band-limited. Wide range perforated\slotted absorbers aren't very helmholtz-ish. See Bogich Petrovich (RIP) papers.
Nils had rockwool rear wall with SBA and had good results.
One buddy (simple hiend stereo towers) consulted with some "acoustician" build huge helmholtz rear wall. Bad before and slightly less bad after . For hiend-ish price.

Calling it Helmholtz-wall was not a good choice, I see it can be misleading.

The acoustic properties of such a rockwool wall can be adjusted using slats - more slats -> more reflective at higher frequencies. This is used to adjust the frequency balance of the room. When the slats cover most of the area, so that there are only small gaps left, the absorption efficiency increases a little at low frequencies.

 

[Flaesh]

Umik1 in listening place (red curve about foot from side wall):

 

[Flaesh]

Wavelets are discussed in other threads as an incomprehensible measurement with unclear meaning and potentially arbitrary interpretation. Here is my in MLP:

 

[ernestcarl]

Wavelets are discussed in other threads as an incomprehensible measurement with unclear meaning and potentially arbitrary interpretation. Here is my in MLP:
View attachment 231865

The vast majority of in-room wavelets I've seen have "breaks" or kinks in the response caused by the room. Technically, this means that the transient behavior is less than ideal. In the bass region, one can usually correct some of the excess GD or excess phase with FIR filters serving as "room correction" phase EQ -- though usually this should be combined with regular min-phase equalization as well (i.e. mixed phase EQ).

*As room induced kinks in the response may be highly position dependent/variable, certain corrections may be most effective only at the measured MLP.

One example:

Pure, minimum phase only EQ:

Mixed phase xo EQ:

The room induced node from imperfect room acoustics still exists, of course, but the null is moved a little later in time, improving the overall immediate transient response specifically in the bass region of the system.


Only three phase EQ sliders were used in rePhase here:
*edit: plus one more linearizing 12dB/oct. subsonic filter at 21 Hz

Normal min-phase PEQs for the sub in the DSP chain are kept separate from this FIR filter and relegated elsewhere to "economize".

Quite frankly, in this example, the clarity is already quite high enough that it's arguable whether this type of correction was even necessary to begin with... nevertheless, some "improvements" in the measurements can be seen.

 

[Flaesh]

Поскольку перегибы отклика, вызванные помещением, могут сильно зависеть/изменяться от положения, некоторые корректировки могут быть наиболее эффективными только при измеренном MLP.

DBA has [signifcant] effect in most of the room. I'll show sometimes; can't find file now.
My grafs here

(unless otherwise specified) no smooting, without eq\correction, with long default window.

Definition:

 

[Flaesh]

For Keith_W

 

[Keith_W]

Thank you! Sorry for being a pain, but do you have a "before" and "after". And would you be able to extend the scale beyond 300ms?

 

[Flaesh]

do you have a "before" and "after"

Don't for actual setup. Gonna add measurements in the indefinite future.

Old first try, before

after:

and well audible difference in D50

 

[Keith_W]

From what I can see, the wide Q suckout at 50Hz has been nicely filled in, but it looks as if ringing is the same?

 

[Gadgety]

Thank you Flaesh for posting. Where is your crossover set to the subs, 100 Hz, or 80 Hz, or something else?