Integrating subwoofers with planar speakers

[klettermann]

A further comment on sub / panel integration...

I have listened for substantial periods of time to various of the Martin Logan "Hybrid" panels, with the built in subwoofer.

Although I tried valiantly to like them - I never found the bass to be "of a piece" with the other frequencies.... there was a definite disconnect in the timbre/texture of the bass vs the midrange... which always bothered me.

Although we have a lot more tools for SW integration - the fact that a vaunted speaker designer like ML could not integrate properly within the speaker, has always made me think that there is definitely something to the belief that panels integrating with subs is a difficult conundrum....

Years ago, Gradient made a "panel" sub, to mate to the Quad ESL63 panel speakers (the SW-63) - they were reviewed a number of times, and consistent reports of good integration resulted... They are very rare here in Australia, and I have never had a chance to audition a set - so I can make no personal experience based comments.

I'd argue that the "design flaw" is attaching the subs to the main panels in the first place. This negates any possibility of optimizing sub placement making any kind of real system optimization a nonstarter. Hence the problem isn't integration with the rest of the speaker. If you're free to move subs around you can work wonders. Otherwise, not so much.

 

[DVDdoug]

The other aspect, is that so many audiophiles have grown up with box resonances,

A good won't have audible resonances no matter the design.

Pro studios use "box speakers" and nobody complains about pro monitors being "boxy".

Of course it's going to sound different from a dipole. Every speaker sounds different and dipoles are "more different".

are you playing this in mono? it doesn't sound right.... yes, the magic of the Quad midrange was missing....

Here is a quote from Amir. It's a slightly different subject. He's actually talking about evaluating speakers in mono.

A dipole speaker surely generates spatial effects that are not real and were never heard in the studio. Combine two such speakers and you are just dealing with fantasies as opposed to mimicking what someone may have setup.

and indeed it did not sound as good.

Again, every speaker sounds different,and better or worse.

But it just goes to show, without the months of acclimatisation to the panels... the other speakers would have sounded "fine".

It's a matter of taste. But you aren't hearing the music "accurately as intended" or as it was heard during mixing/mastering. I'm not saying that's a bad thing... I'm guilty too. Personally, I use a "hall" or "theater" setting on my AVR for some delayed reverb in the rear and the "feel" of a larger space. I've had my speakers in a dance hall and to me they sound much better in larger space.

 

[Willem]

But you aren't hearing the music "accurately as intended" or as it was heard during mixing/mastering.

And yet, Philips classical recordings, known for their superb representation of the acoustical environment, were often monitored with Quad electrostats. Legendary independent sound engineer Tony Faulkner similarly uses them.

 

[mglobe]

I appreciate all the comments, and hope we can keep on the topic of integration of subs with planars, vs the pros and cons of conventional vs planar speakers.

 

[dlaloum]

Don't take offense now. I'm using myself as an example to show what I mean. One of the worst things I know is being able to pinpoint the sound to the speakers.
I want them to "disappear" so to speak.

In any case. I find that my pinpointing ability varies from day to day. Whether it's due to concentration difficulties, musical material OR an imaginary notion that I can pinpoint and that's why I do it, I don't know. It could also be a combination of all of the above.

With that said. Do you think you hear box sound, box resonances because you expect it to be there? Or are there really, objectively, resonances that are audible that you experience? In any case, as I said, don't take offense. It was just a thought.

In several cases, I walked into showrooms that had both panels and boxes deployed.... and on asking for confirmation, they confirmed it was the "boxes" playing - so yes, in a number of cases, I could tell "blind"

Interestingly I find speakers that are NOT panels but also NOT boxes (such as the Gallo Nucleus series) to be neutral... ie: not subject to those resonances

For WAF reasons I ultimately switched from my panels to the Gallo Nucleus Reference 3.2 .... one of the few non panel speakers within my budget that sounded good. (those Box identifiers are a bugbear for me!)

And no not all boxes have them (the identifying resonances), and there have been some box speakers I have really enjoyed... The Boston Acoustics A400 comes to mind.

But in walking the showrooms and demo rooms, somewhere over 70% of the "box" speakers tend to leave me cold...

Theorising - I have a feeling it is the resonances being transmitted via the side panels... often those I like have shallow depth (an unusual shape nowadays) - which limits the resonant surface area...

 

[dlaloum]

I'd argue that the "design flaw" is attaching the subs to the main panels in the first place. This negates any possibility of optimizing sub placement making any kind of real system optimization a nonstarter. Hence the problem isn't integration with the rest of the speaker. If you're free to move subs around you can work wonders. Otherwise, not so much.

That may well be the case... - but then you are positing that any full range speaker is fundamentally flawed, and all speakers should be bandwidth limited with subwoofers relocated for optimum room behaviour...

But that then asks a different question - there is localisation of the higher bass frequencies... which isn't a big issue in HT, as the LFE channel is just effects, and not required to be localisable... but there is localised material in the main channels - which then needs to be crossed over to sub(s) - and if crossed over too high (what is too high?) - you lose any bass localisation.

Definitely the 120Hz down to 80Hz frequencies are in that category... I have a feeling it may well extend down to 60Hz...

By the time you get down to 40Hz I am pretty sure it is no longer localisable....

So for high quality music purposes - I would suggest the main panel should hopefully stretch down to 60Hz, before crossing over to the sub(s)

Looking forward to Dirac ART making all this a lot easier!!

 

[Chrispy]

Not that the speakers are that difficult, rather the way/gear you implement such integration IMO.

 

[DavidMcRoy]

If you substitute "slow vs. fast" with "distorted vs. clean" it makes sense.

 

[Chrispy]

If you substitute "slow vs. fast" with "distorted vs. clean" it makes sense.

Not really, still lots of interpretation variances as to what the hell someone means by a vague statement like that....

 

[MAB]

If you substitute "slow vs. fast" with "distorted vs. clean" it makes sense.

I thought 'fast' meant 'weak'. Lots of those 'fast' bass speaker have very high bass distortion.
But the description never really made sense to me, so any interpretation is in play.

 

[Sir Sanders Zingmore]

I'd argue that the "design flaw" is attaching the subs to the main panels in the first place. This negates any possibility of optimizing sub placement making any kind of real system optimization a nonstarter. Hence the problem isn't integration with the rest of the speaker. If you're free to move subs around you can work wonders. Otherwise, not so much.

This is a very valid argument that is surely not only applicable to panels. Freedom of placement for subs should allow for better results with speakers if any type.

Having said that, I imagine that in many cases the sub isn’t really a “pure” sub in that it probably crosses to the panel at a frequency that means it can be localised. That’s certainly the case with my Sanders speakers. They cross to the woofer at ~170hz which I think is quite a bit lower than most other “hybrid” electrostats. They are also fully active with crossovers done in the digital domain which allows for steep slopes and time alignment. To my ears, the integration is seamless.

 

[DanielT]

In several cases, I walked into showrooms that had both panels and boxes deployed.... and on asking for confirmation, they confirmed it was the "boxes" playing - so yes, in a number of cases, I could tell "blind"

Interestingly I find speakers that are NOT panels but also NOT boxes (such as the Gallo Nucleus series) to be neutral... ie: not subject to those resonances

For WAF reasons I ultimately switched from my panels to the Gallo Nucleus Reference 3.2 .... one of the few non panel speakers within my budget that sounded good. (those Box identifiers are a bugbear for me!)

And no not all boxes have them (the identifying resonances), and there have been some box speakers I have really enjoyed... The Boston Acoustics A400 comes to mind.

But in walking the showrooms and demo rooms, somewhere over 70% of the "box" speakers tend to leave me cold...

Theorising - I have a feeling it is the resonances being transmitted via the side panels... often those I like have shallow depth (an unusual shape nowadays) - which limits the resonant surface area...

I suspect that if I also went blind into a demo room where boxes vs panels were playing that I would spot the difference between the speakers. But I could have done that if it was two traditional box speakers with fairly different FR.

Let's say a pair of panels vs box speakers with the same FR and still be able to spot each blindly. My guess then is that it is due to dispersion and that reflexes that bounce differently in the room in different ways due to its different speaker designs give rise to the audible sound difference.

Last I would guess a difference in distortion/resonances that was the reason for picking out each blind.

You might have a point in that you can get so used to resonance-free speakers so that when you listen to non-resonance-free, say box speakers, you don't need much resonance before it's detected. The question is mostly how much resonance is needed before it becomes audibly annoying? Refrigerator-sized speakers with almost paper-thin walls without any damping material in them, sure I can imagine that. But small bookshelf speakers with "normally thick" walls, with appropriate damping material in them. I find it hard to believe. Maybe you can detect resonances from such speakers but I'm doubtful. That is assuming they don't have a poorly designed port that in itself gives rise to resonances.

For example, these solidly built, no ringing, GR speakers what I can guess have no audible resonances. However, as Ageve points out: 8% (-22 dB) 2nd harmonic distortion at 4 kHz. Atrociously high distortion that can definitely be audible. It could perhaps be taken for resonances?, which it is not:

D

'Erin reviews GR-Research NX-Bravo'

Apr 3, 2025

Short snippets from the review:

- on axis is really bad, but it evens out by 15 degrees off axis
- super wide dispersion up to around 4k, then narrows quickly
- very high distortion from the tweeter (1% H4/H5 around 2k at 86dB/1m)
- super clean decay behavior


Overall, it's... Fine, I guess. Hard to justify at the price point of $1525 base considering there are better options out there from other manufacturers e.g. Kef or Asciilab.

• dfuller
• Replies: 243
• Forum: Speaker Reviews, Measurements and Discussion

• 

Round speakers like Gallo that you mention and speakers with narrow side walls. The reflections that the constructions give around the speaker itself vs. traditional box. I can absolutely imagine that it gives rise to a difference in reflections and thus sound.
A round speaker, with a coax/broadband driver with even on and off axes can probably be an excellent speaker. Or there is potential for an excellent speaker with a spherical shape speaker design.
Round subwoofers on the other hand, which thereby reduce baffle support, I am more hesitant about.

 

[mglobe]

That may well be the case... - but then you are positing that any full range speaker is fundamentally flawed, and all speakers should be bandwidth limited with subwoofers relocated for optimum room behaviour...

But that then asks a different question - there is localisation of the higher bass frequencies... which isn't a big issue in HT, as the LFE channel is just effects, and not required to be localisable... but there is localised material in the main channels - which then needs to be crossed over to sub(s) - and if crossed over too high (what is too high?) - you lose any bass localisation.

Definitely the 120Hz down to 80Hz frequencies are in that category... I have a feeling it may well extend down to 60Hz...

By the time you get down to 40Hz I am pretty sure it is no longer localisable....

So for high quality music purposes - I would suggest the main panel should hopefully stretch down to 60Hz, before crossing over to the sub(s)

Looking forward to Dirac ART making all this a lot easier!!

To some extent I agree with the theory that full range speakers are flawed due to the inability to adjust for room influence compared to the flexibility of a main-sub combination. FWIW, I've got my Reference 1 Meta's crossed at 60 with my subs, which are run in stereo mode. This allows me to get whatever localization affects are there. I'm fairly convinced that you can sense basic left/right directionality. All of this now goes through a Trinnov NOVA BTW. Gave up on waiting for ART.

 

[NTK]

Dipoles couple to room modes totally differently from monopoles. Dipoles excite velocity modes (at the pressure mode nulls). Since velocity modes are directional, the orientations of dipoles with respects to the room modes is of critical importance. Monopoles excite pressure modes, and only their locations (or more specifically, the locations of their acoustic centers) matter.

Therefore, putting a monopole subwoofer right next to or under a panel speaker will sound differently from using the panel speaker alone, as the bass response will be different due to different room mode coupling. EQ based on in-room measurements is required for proper integration.

 

[Nicoff]

I owned Magnepan speakers in the past (IIIa, 20.1, 20.7) and was able to fully integrate them with multiple subs (2-3 subs)
Two things to consider: 1) Placement of the subs is very important, and 2) using devices like mini DSP or even an A/V preamp to optimize level, crossover, and timing.

 

[mglobe]

Dipoles couple to room modes totally differently from monopoles. Dipoles excite velocity modes (at the pressure mode nulls). Since velocity modes are directional, the orientations of dipoles with respects to the room modes is of critical importance. Monopoles excite pressure modes, and only their locations (or more specifically, the locations of their acoustic centers) matter.

Therefore, putting a monopole subwoofer right next to or under a panel speaker will sound differently from using the panel speaker alone, as the bass response will be different due to different room mode coupling. EQ based on in-room measurements is required for proper integration.

I have no doubt that you are correct, but I'm struggling to understand how to put this in the context of optimizing an integration. Since the mains and subs would be crossed over, it seems like there would be a very narrow range where the difference in the locations of modes and nulls between the two sets of speakers would matter to the listener. That's of course assuming the speaker locations are reasonably good. Is there anything you could do with the software you used in your directivity pattern thread https://www.audiosciencereview.com/...ectivity-patterns-couple-to-room-modes.45518/ that would help us understand?

 

[Sukram]

My Journey with Magnepan Room Correction and system optimization

System Optimization
Since the 1970s, when I first heard the Magnepan Tympani, I've been a fan of dipole speakers. I had many KEF and Focal speakers, and later I returned to dipole speakers: the Martin Logan Stylus, the Magnepan 1.7, and finally the Magnepan 3.7i. The best power amplifier for the Magnepan and all speakers is the Purifi power amplifier period. I wouldn't trade it for a $95,000 Pass Lab. Additionnally I added a Rythmik F12 XLR3 subwoofer to satisfy my bass needs.
I started my journey of room correction with a Magnepan with a NAD C658 preamp and a full-featured Dirac room adjustment system. What happened:

Magnepan Room Correction Problems with Dirac
Using Dirac Live with Magnepan 3.7i speakers presented challenges. The time-alignment across the full frequency range made the Maggies sound boxy and unnatural. Dirac’s global phase correction interfered with the dipole radiation and the natural 'air' of the speakers. Even if I restricted the frequency range to 20 -200Hz, Dirac would time align my room reflections to 20 kHz. The result was technically correct but musically unsatisfying.
By the way the NAD C658 is a very good versatile piece of equipment, which is made for the Purifi power amp.

Advantages of Using the dbx VENU360 Instead of Dirac
Switching to the dbx VENU360 offered advantages over Dirac when paired with Magnepans. The dbx focuses on amplitude-based AutoEQ without imposing full-band time alignment. This preserves the natural openness and imaging of the Maggies while smoothing room-induced peaks and dips. Additionally, the dbx allows precise crossover and subwoofer integration (time alignment) without over-processing.

Going Analog into the dbx
Initially, the system was fed into the dbx via analog inputs (NAD’s XLR’s). This worked well, but it meant going through an extra A/D conversion before processing, which limited the potential clarity and added unnecessary conversion stages.

Switching to AES Digital Input with Wiim Pro Plus
The introduction of the Wiim Pro Plus streamer changed the chain significantly. Using the AES/EBU input on the dbx connecting the WiiM pro plus digital coax out over a 75-to-110 Ohm impedance converter, which allowed a direct digital handoff. This eliminated the redundant A/D step, ensuring cleaner signal delivery into the dbx.

Wiim Pro Plus – Best Software in Class (personal opinion)
Beyond the hardware, the Wiim Pro Plus stands out for its software ecosystem. The app is intuitive,
stable, and offers excess to all streaming apps. Compared to many competitors,

Restricting Output to 24/48 to Match the dbx
Although capable of 24/192 output, I restricted the Wiim to 24/48. This matches the dbx VENU360’s internal DSP rate, preventing unnecessary down-sampling inside the processor. In this setup, the Wiim delivers a bit-perfect 24/48 stream directly clocked to the dbx, achieving maximum transparency.

Clocking and Jitter Considerations
With Sample Rate Conversion disabled on the dbx, the processor locks directly to the WiiM’s low-jitter clock. This ensures stable, accurate timing. Feeding 24/192 into the dbx would have forced a resample down to 48 kHz internally and used the lower quality DBX clock. By restricting the Wiim to 24/48, the system avoids this redundancy.

High Headroom for Volume Control
The Wiim Pro Plus upsamples internally to 32-bit precision for volume control. Feeding the dbx a 24-bit, 48 kHz stream ensures over 140 dB of dynamic range, so the Wiim volume control provides plenty of headroom without audible resolution loss. This makes digital volume control transparent and safe, even at moderate listening levels.

Final Take
I calibrated the system with the dbx venue 360 connected to the dbx RTA microphone. I used the DBX AutoEQ. The calculated DBX room correction can then be fine-tuned manually. I used the microphone positions for the right and left ear. This results in a very close seating position, but with the best results for me. Finally, a few thoughts on high-resolution DACs or improving with better streamer transport, etc. With room noise of ~30dB and the influence of the room and speakers, investing in expensive equipment is not worth it. I checked out expensive DACs, but in the end, I sold them; I didn't hear any noticeable difference.
With the WiiM pro plus -> DBX Venue 360 -> Purifi -> 3.7i + Rythmik system, I'm happy with my Maggies. Finally.

 

[Farenheit]

My Journey with Magnepan Room Correction and system optimization

System Optimization
Since the 1970s, when I first heard the Magnepan Tympani, I've been a fan of dipole speakers. I had many KEF and Focal speakers, and later I returned to dipole speakers: the Martin Logan Stylus, the Magnepan 1.7, and finally the Magnepan 3.7i. The best power amplifier for the Magnepan and all speakers is the Purifi power amplifier period. I wouldn't trade it for a $95,000 Pass Lab. Additionnally I added a Rythmik F12 XLR3 subwoofer to satisfy my bass needs.
I started my journey of room correction with a Magnepan with a NAD C658 preamp and a full-featured Dirac room adjustment system. What happened:

Magnepan Room Correction Problems with Dirac
Using Dirac Live with Magnepan 3.7i speakers presented challenges. The time-alignment across the full frequency range made the Maggies sound boxy and unnatural. Dirac’s global phase correction interfered with the dipole radiation and the natural 'air' of the speakers. Even if I restricted the frequency range to 20 -200Hz, Dirac would time align my room reflections to 20 kHz. The result was technically correct but musically unsatisfying.
By the way the NAD C658 is a very good versatile piece of equipment, which is made for the Purifi power amp.

Advantages of Using the dbx VENU360 Instead of Dirac
Switching to the dbx VENU360 offered advantages over Dirac when paired with Magnepans. The dbx focuses on amplitude-based AutoEQ without imposing full-band time alignment. This preserves the natural openness and imaging of the Maggies while smoothing room-induced peaks and dips. Additionally, the dbx allows precise crossover and subwoofer integration (time alignment) without over-processing.

Going Analog into the dbx
Initially, the system was fed into the dbx via analog inputs (NAD’s XLR’s). This worked well, but it meant going through an extra A/D conversion before processing, which limited the potential clarity and added unnecessary conversion stages.

Switching to AES Digital Input with Wiim Pro Plus
The introduction of the Wiim Pro Plus streamer changed the chain significantly. Using the AES/EBU input on the dbx connecting the WiiM pro plus digital coax out over a 75-to-110 Ohm impedance converter, which allowed a direct digital handoff. This eliminated the redundant A/D step, ensuring cleaner signal delivery into the dbx.

Wiim Pro Plus – Best Software in Class (personal opinion)
Beyond the hardware, the Wiim Pro Plus stands out for its software ecosystem. The app is intuitive,
stable, and offers excess to all streaming apps. Compared to many competitors,

Restricting Output to 24/48 to Match the dbx
Although capable of 24/192 output, I restricted the Wiim to 24/48. This matches the dbx VENU360’s internal DSP rate, preventing unnecessary down-sampling inside the processor. In this setup, the Wiim delivers a bit-perfect 24/48 stream directly clocked to the dbx, achieving maximum transparency.

Clocking and Jitter Considerations
With Sample Rate Conversion disabled on the dbx, the processor locks directly to the WiiM’s low-jitter clock. This ensures stable, accurate timing. Feeding 24/192 into the dbx would have forced a resample down to 48 kHz internally and used the lower quality DBX clock. By restricting the Wiim to 24/48, the system avoids this redundancy.

High Headroom for Volume Control
The Wiim Pro Plus upsamples internally to 32-bit precision for volume control. Feeding the dbx a 24-bit, 48 kHz stream ensures over 140 dB of dynamic range, so the Wiim volume control provides plenty of headroom without audible resolution loss. This makes digital volume control transparent and safe, even at moderate listening levels.

Final Take
I calibrated the system with the dbx venue 360 connected to the dbx RTA microphone. I used the DBX AutoEQ. The calculated DBX room correction can then be fine-tuned manually. I used the microphone positions for the right and left ear. This results in a very close seating position, but with the best results for me. Finally, a few thoughts on high-resolution DACs or improving with better streamer transport, etc. With room noise of ~30dB and the influence of the room and speakers, investing in expensive equipment is not worth it. I checked out expensive DACs, but in the end, I sold them; I didn't hear any noticeable difference.
With the WiiM pro plus -> DBX Venue 360 -> Purifi -> 3.7i + Rythmik system, I'm happy with my Maggies. Finally.

I have the same setup as you, except I use QUAD ESL 988 speakers and different subwoofers and amplifiers. My question is: why did you select 48 kHz instead of 96 kHz, when both the DBX and the WIIM support both options?.

Electrostatic greetings!

 

[klettermann]

I owned Magnepan speakers in the past (IIIa, 20.1, 20.7) and was able to fully integrate them with multiple subs (2-3 subs)
Two things to consider: 1) Placement of the subs is very important, and 2) using devices like mini DSP or even an A/V preamp to optimize level, crossover, and timing.

^^^ This. Fully agree. x steps:
1. Optimize subs - delay and EQing
2. Fix panel EQ
3. Integrate panels <> subs (xo, tining)
Done. Sit back and enjoy. My system never sounded better and endlessly blows my mind.

My biggest problem was not knowing when to quit fiddling with DSP and just enjoy it. Cheers,

 

[levimax]

To some extent I agree with the theory that full range speakers are flawed due to the inability to adjust for room influence compared to the flexibility of a main-sub combination.

After years of messing around with subs, including all the fancy measurement and DSP tools, I have come to the opposite conclusion. Main / subs are inherently flawed in that you are making the customer be a speaker designer without the training, tools, known driver spacing, or known driver parameters, and most importantly hopeless time domain measurements due to the limits of LF measurements in a small room. Sure subs will usually sound better than no subs, but I find full range speakers, despite having a more uneven LF measured response, with some light handed DSP, to be preferred.