Lyngdorf "boundary woofer" concept: experiences?

[jhaider]

For quite a while I've been intrigued by the "boundary woofer" concept promoted by Lyngdorf. Their marketing spiel:

To summarize, the basic concept is a 2-box synthetic speaker with the LF section remotely located in the room's front corners. Generally the MF/HF section is on or close to the wall (similar to e.g. Neumann/Genelec recommendations) but I'm not sure that's a requirement. The wrinkle is the LF-MF crossover is quite high for a remote LF section - I've read 200Hz to 400Hz in different contexts.

Has anyone tried this approach? I'm strongly considering experimenting with something like this for our formal living room, as it seems like it could potentially check the boxes of small perceived footprint but full range with smooth bass. The downside is time - if it doesn't work well I'll be stuck with it for a while due to limited time to fiddle with stuff and take good measurements. However, I'm also aware of comments by @hardisj regarding bass localization in his review of the 2.5-way JBL HDI3800, as well as by Dr. David Rich's similar observations in his AudioXpress review of an Infinity tower with a small midrange and a 500Hz specified crossover to a column of 6.5" woofers. Perhaps that can be mitigated with very steep filters?

 

[dannut]

Yes, the concept is sound. But I am also deaf to bass localization.

It works well in small/narrow rooms in tandem with room correction, that way the not-colocated sub-in-corner isn't so far from the main speaker. Tried that in early 2000ish, when DRC-FIR came out, maybe up to 250Hz?

Funny thing, that with RC you can do stuff that are no-no in 'regular' settings. E.g. speakers near walls, using thin absorbers only for high frequencies (>500-1000Hz), the rest is taken care by FIR based room correction. With DRC-FIR you have all the settings possible to fine tune the size of 'sweet spot'. By placing speakers near an offending boundary, you can go higher with the upper correction limit. High frequencies that are not possible to correct are easily taken care of by thin absorbtion. Yes - a totally backwards approach to room acoustics and system design.

Since then I discovered multichannel upmixing and multiple subs. Sadly none of the boxes offer customizable bass management and user uploadable/custom room correction.

Also check slides 25 and 26 from j_j: https://www.aes-media.org/sections/pnw/ppt/jj/room_correction.ppt

 

[dougi]

I tempted to try at the moment. I have a lyndorf pre, and a Sydney shop is reselling a set of lyngdorf dipole speakers and boundary woofers for only A$1500. These ones need an amp for the woofers though.

 

[jhaider]

It works well in small/narrow rooms in tandem with room correction, that way the not-colocated sub-in-corner isn't so far from the main speaker. Tried that in early 2000ish, when DRC-FIR came out, maybe up to 250Hz?

Hmm, the room in question here isn't that small or narrow. The width is basically 50' (room is on one end of a mostly open space on our ground floor), though the stub corners for the room are about 17' apart.

Funny thing, that with RC you can do stuff that are no-no in 'regular' settings. E.g. speakers near walls, using thin absorbers only for high frequencies (>500-1000Hz), the rest is taken care by FIR based room correction.

Or this, with no absorbers at all! And PEQ based correction does it, too.

Since then I discovered multichannel upmixing and multiple subs. Sadly none of the boxes offer customizable bass management and user uploadable/custom room correction.

Right. Here I'd be DIYing bass management and likely relying on Dirac Live to set the overall LF target curve ("preamp" is miniDSP SHD Studio). This system is a secondary one, and my first 2-channel (as opposed to surround or immersive) non-nearfield setup probably since 2001. The main system is 7.1.4-channel with multisubs.

 

[dannut]

Or this, with no absorbers at all! And PEQ based correction does it, too.

Absolutely, with modern good/narrowish dispersion there is no need for absorbtion and corner speaker placement.

Only exception to PEQ-s are non min-phase issues just above Schroeder. SBIR is a perfect example. You can have pretty robust cancellation with FIR/convolution based RC, that is not so good with IIR/PEQ. j_j-s slide goes in more detail. IIRC Windows Vista, of all things, had his implementation. I would think that DIRAC gets that correct also. It is especially needed with on-wall speaker placement, as opposed to in-wall. SBIR effects are at a high-300-500Hz region and at worst perceptually (for me..)

Neumann-Genelec (first was Allison?) method gets the low freq. near boundary and highs away from boundary (>250Hz ~ 80cm) or flush with boundary. That way there is no need for SBIR (room) correction.

PS - ugh, I cringe if I have to listen far field 2ch material, without a proper centre. You seem to be also an upmixing fan... Vinyl, with its massive crosstalk, makes it somewhat bearable. Also if you can, there is a 'shuffle' algorithm, that alleviates some of the problems with 2ch.

 

[Holmz]

I have Lyngdorf, and a couple of Theil subs to go into some boxes…
I’ll make some time in a couple of months to test it out.

 

[jhaider]

Well, I finally had a chance to take some in-room measurements this weekend. System as tested:

• Speakers: Revel Gem2 on factory pedestal stands, which make them about 7.5” thick. Gem2 is a 3-way speaker with an 8" woofer in a small closed box, and anechoic F3 in the 80Hz range.
• "Boundary subwoofers" - SB Acoustics SW26DAC76-8 10” woofers in itsy bitsy closed boxes (about 6L/0.25ft^3 volume, about 4.5" total thickness), powered by a Crown XLS2502 and with bass boost in the controller to extend nearfield response down to an F3 ~35Hz and an F6 of just over 20Hz. Basically the FR is sculpted anechoically close to Salon2 bass extension.

I did not include the two additional subwoofers in the system in these measurements, as they are intended to smooth the bass and perhaps to extend the low end down, but are not part of the boundary sub concept.

Here's what things look like. All measurements below are unsmoothed.

Let’s start with the some experiments with mains placement. I was curious if there were any differences between angled to the listening position (4.25” delta between inner corner and outer corner) and flat on the wall. I prefer the mids and treble angled, though they look better flat. In terms of bass it’s irrelevant. I was also curious about the effect on bass of moving them out as far as practicable in this room - outer edges about 28” from the wall.

I'm lazy so I only moved the left speaker. These measurements are single point, from where the center of my head would be, because I was only looking for differences rather than something to work with later. (In retrospect I should’ve either moved the right speaker, or used the right speaker for timing reference to capture the acoustic time-of-flight differences between on wall and out in space.)

As you can see, the differences between flat and angled to LP are de minimis in the bass. On wall also IMO looks a lot better than in the space - the suckouts from 80-320 Hz are narrower in bandwidth and there’s nice boundary gain from the wall. So all else equal on wall is probably better for these speakers no matter the subwoofer concept.

I was also curious if it made a difference whether the boundary subs were flat against the back wall or normal to the back wall. The answer is no. Flat on the wall is more attractive, so that’s how they’ll stay.

Take down the room peaks and I eyeball the in room F3 to be low-mid 30s and F6 right around 20Hz. They’ll have help down there, but I think that’s promising.

Here’s what the mains and boundary subs look like. These measurements are taken using the Geddes and Blind Sound Power method, i.e. a spatial average of 5 positions around the listening seat. The left side has a lot less bass, as expected. The right side has a real corner, whereas the left side "corner" is really just a short stub wall in an open concept floor plan.

I was initially fairly disappointed after seeing these data. I was hoping to that a crossover point would be obvious from the data, due to mains response "zigging" and boundary subs response "zagging." But that doesn’t happen.

However, here’s a quick and very dirty crossover, 150Hz 5th order. I’m dubious of the value of “time alignment” in the modal region, so I didn’t compensate for the ~2.5ms delay between mains and subs. (IMO in the modal region delay should be used to optimize the steady state response, not to hit an academic target.) These are single point measurements (unsmoothed) so only useful up to 150Hz or so; ignore the grass at the top of the graph. The second plot has a single PEQ band common to both subs: 75Hz, -11dB Q=2. I should’ve added one at ~48Hz as well, and the 75Hz PEQ gain should probably be closer to -8dB.

Considering I pulled a number out of a proverbial hat and made almost zero effort to optimize the splice or overall FR, I'll take it. At an eyeball, considering one more band of PEQ a shallower 75Hz PEQ, I seeing about ±6dB - before any real optimization or room correction. That may change as I work to smooth out the bass over a wider area. But it's certainly a start. I'll play with the initial splice in VituixCAD over time, and probably get something better. Then then run Dirac over all of it.

 

[dannut]

Not to take away anything from your (zero) effort, but what amazes me the most is how on-wall friendly the Gems are! Wide baffle, shallow depth speaker with a generous round-over. More speakers should be designed like this.

PS do you mind sharing the MDAT file with me(flat on-wall vs 28" out)? Would like to see in more detail how the front wall reflection (SBIR) manifests itself with this speaker.

 

[jhaider]

Not to take away anything from your (zero) effort, but what amazes me the most is how on-wall friendly the Gems are! Wide baffle, shallow depth speaker with a generous round-over. More speakers should be designed like this.

Agreed. With Gem2 at the end of its run (I wanted to buy new but Harman’s whole inventory was single black one, so I spent a while looking for used ones) I can't think of another US market "floor standing" on-wall speaker. (We didn't want to poke holes in the wall to run wiring, so something that looks integrated while fitting close to the wall was important.)

Even in the DIY space there aren't many either, especially if you take out the corner line arrays. @BenB's "Radical" design (see this post and this one, too) is probably the best thought-out of those.

If you broaden to on-walls, offhand there's Revel's S16 and Monitor Audio's Sound Frame 1. Sound Frame 1 is intriguing to me because it uses a novel coax - an inverted cone with tweeter at apex, in a rotatable housing. Oddly (or perhaps for good reason?) Monitor does not have serious marketing around it that I've seen.

FWIW I also considered JBL CBTs. I discussed with @Archaea a bit (he provided Amir's CBT 70J DUT). Ultimately the cost to get a pair and commission replacement cabinets that would aesthetically fit in our living room was just too high.

PS do you mind sharing the MDAT file with me(flat on-wall vs 28" out)? Would like to see in more detail how the front wall reflection (SBIR) manifests itself with this speaker.

Not at all. Here are those mdats for anyone to take a look. If you see something (or an error in procedure) I didn't, I welcome the input!

The other thing I'm pondering, if anyone has a thought: take out the bass-extending EQ on the "boundary subs" and do not filter them, so they're still contributing to ULF but at a much lower and declining level. Then fold in the two other subs so that they're filling in holes and contributing ULF. The thought came to me when I processed the spatial average of the Tymphany LAT-700 based sub built into the media cabinet.

I don't know that I've ever seen an in room sub response that good before any EQ. It also moves plenty of air: at ~.65L volume displacement, quite a bit more more than a typical 12" subwoofer. So with that and 3 10" cones (boundary subs + JL Audio E110) all playing in the lowest octaves I'm thinking I should ignore individual response entirely and just optimize the system for low variance. Then, use global EQ + Dirac to fix the summed FR in the end. I have no idea how Dirac (not DLBC, just the basic one) will work on what amounts to a Geddes-style subwoofer setup, though.

(NB, The E110 response is basically as expected. It's in a terrible place for a sub - in a stub "corner" but basically in the middle of an open concept floorpan. Nearfield it's the flattest and most extended sub unit in the room.)

 

[dannut]

... take out the bass-extending EQ on the "boundary subs" and do not filter them, so they're still contributing to ULF but at a much lower and declining level. Then fold in the two other subs so that they're filling in holes and contributing ULF. ...

Yes, folding in and overlap can be good things, contrary to popular (ignorant?) opinion. Both to modal problems and higher up in transition region. Whats holding back a 'regular' best-practice system (SFM bass + 80Hz crossed over best 'regular' speakers (Salon,Kef Ref, etc)) are precisely the problems in the 80-300Hz region. Up to ~500-700Hz with on-walls.
Allison-style folding in 'main' speakers mitigates floor and ceiling bounce dip, and other SBIR problems. And of-course use the 'flanking subs' also for modal control, by folding in 'main' subs. Real men* don't use crossovers. Only various highpasses and lowpasses, dictated by localization and output concerns. Bonus points for not over-specifying various components! Main speakers can be small, ditto for flanking subs.

I have no idea how Dirac (not DLBC, just the basic one) will work on what amounts to a Geddes-style subwoofer setup, though.

Yeah. You should see my rant in another thread. Just a little more flexibility with current bass-management schemes would go a long way mitigating small-room acoustical problems. And no, I would not want to use a Trinnov for a non-movie-theater style setting.

Here are those mdats for anyone to take a look.

Thanks for that. Will have to wait till after the weekend thou.


* as in person. Most other global languages trip me up with them every-noun-has-a-gender things. The world would be a better place if we adopted a global language with no gender specific third-person pronouns