• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Speaker Cabinet Design Considerations

Compression in the bass regions at higher SPL. Very disingenuous of Devialet to market it as being full range down to 14hz when you really wouldn't want to be playing it that low at loud volumes. It kind of irks me but I guess the marketing department behind the phantom range of speakers is completely separate entity from their engineering department.
Surprised.
I hadn't heard about claims of 14Hz. Certainly the DSP can't continue to correct for cabinet volume indefinitely but in real terms there is no obvious limit in my room. I have never had it get to a limit whilst listening and watching the (mains) power meter, even on daft film sound tracks LOUD.
My view is that criticism has been sour grapes from people incapable of making such a product, plus the obvious digital is no good plonkers.
I find little to complain about.
 
Compression in the bass regions at higher SPL. Very disingenuous of Devialet to market it as being full range down to 14hz when you really wouldn't want to be playing it that low at loud volumes. It kind of irks me but I guess the marketing department behind the phantom range of speakers is completely separate entity from their engineering department.

Also my skepticism of Devialet speakers is more to do with my opinion they are not even close to as good as the value proposition of DIY. Which is of course what this thread was about. I scoff at the idea that two phantoms at a price of $6000 is a good value proposition when you will get much better measured performance out of a pair of cheap bookshelves and two actual subwoofers in a substantial volume enclosure to get their SPL numbers.

That's nice you own a lot of speakers. I would too if I had the space. At one point I had 3 pairs in my small-ish room but I really just didn't have the space. You also are also more than 3 times my age, so you win on experience. -shrugs-
Physically small active 'speakers have used DSP to trade off loudness, i.e cone excursion, against frequency so as to reduce distortion and the risk of damage to the bass driver. I think it's a valid technique, PROVIDED that the driver isn't so small that the bass output is below Fletcher Munson levels, and so can't be heard.

That's where marketing comes in....yes, this 4" bass unit can go down to 20Hz, and yes, this loudspeaker can do 110dB SPL....just not both together........

S
 
Compression in the bass regions at higher SPL. Very disingenuous of Devialet to market it as being full range down to 14hz when you really wouldn't want to be playing it that low at loud volumes. It kind of irks me but I guess the marketing department behind the phantom range of speakers is completely separate entity from their engineering department.

Also my skepticism of Devialet speakers is more to do with my opinion they are not even close to as good as the value proposition of DIY. Which is of course what this thread was about. I scoff at the idea that two phantoms at a price of $6000 is a good value proposition when you will get much better measured performance out of a pair of cheap bookshelves and two actual subwoofers in a substantial volume enclosure to get their SPL numbers.

That's nice you own a lot of speakers. I would too if I had the space. At one point I had 3 pairs in my small-ish room but I really just didn't have the space. You also are also more than 3 times my age, so you win on experience. -shrugs-
Ha-ha if I am 3 times your age it is no surprise you still have so very much still to learn.
Enjoy the journey!
 
Unsure how much of an audible difference it makes, but when I look at a speaker cabinet, I'm looking for a a few things ideally:

-No parallel surfaces
-No sharp edges along the entire front
-A waveguide or a front that is spherical
-Not shaped like a box
-Not made from panels of regular wood

Not too many that follow all of these. a Devialet Phantom and Genelec 8260 come to mind, along with an "egg" shaped speaker some grammy winning artist uses, and maybe MBL speakers minus the lower woofer portion since there is no cabinet?

I'm curious what the tradeoffs are doing internal bracing (JBL) vs using a stronger, thinner material to maximize internal volume (Genelec.) Seems like the latter would be easier to work with due to the lack of internal braces. Not sure how crowded it is inside of a Genelec compared to something of the same volume with traditional cabinet design.

You have hit on the million dollar question. How can I tell which cabinet design is better? In this case, there are merits to each. Until recently, my approach has been to find great small speakers and supplement them with a subwoofer (or two). I have found this to be a better value than larger speakers. High quality monitors are less expensive and cabinet issues are more readily dealt with in smaller cabinets. As with anything, there are compromises...

My latest acquisition are Epique CBTs (not small, but are narrow and curved). They were pre-built, but are now only DIY. The cabinets are not wood, but MDF (so really a wood-based composite). MDF is used because it has a good tradeoff between damping, stiffness and cost. So would not count out all boxes too quickly. OTOH, if you really want to eliminate boxes, suggest you check out linkwitz lab or some of the open baffle designs. Happy hunting!
 
Olson's classic paper concerns the solutions of cabinet diffraction in the far-field limit. It is very difficult to get into the far-field space of any cabinet's diffraction pattern and still get a reasonable playback level. The box is simply too large. A consequence is that the textbook 6.02 gain difference from half-space at low frequencies is not about to be reproduced at a typical position a couple of meters away from the speaker. Olson's paper also assumes omni sources, ka<<<1.

AFAIK, re-radiation of sound from a loudspeaker cabinet is local in effect. That is, it mostly focuses into narrow bands. Generally, CLD would be preferred for increasing transmission losses over the extensional damping of the BBC style boxes. The walls are too thick, stiff, and heavy for that approach to be well-compatible. Most suitable elastomers usually have greater midrange losses as well.
 
Last edited:
Physically small active 'speakers have used DSP to trade off loudness, i.e cone excursion, against frequency so as to reduce distortion and the risk of damage to the bass driver. I think it's a valid technique, PROVIDED that the driver isn't so small that the bass output is below Fletcher Munson levels, and so can't be heard.

That's where marketing comes in....yes, this 4" bass unit can go down to 20Hz, and yes, this loudspeaker can do 110dB SPL....just not both together........

S
https://www.devialet.com/en-ca/phantom-premier-heart-bass-implosion/
"Working alongside SAM® and ADH®, HBI® acoustic architecture enables Phantom to reach the most extreme low-end frequencies ever achieved. Right on down to 14Hz. Travel beyond limits of the human ear.
Experience the physical impact of ultra-dense sound. Rushing through your body.


Big emotions no longer have to equal big speakers."

The wording here seems to indicate the Phantom's have useful low end extension down to 14hz, and it's of sufficient SPL that you could still feel it. I very much doubt that is truly the case. I actually hope I'm wrong. I will happily sell my JBL to get a pair of these if I am. I've always wanted full range speakers in a small package. But this goes against everything I've learned, that you need large cabinet enclosure to get low end extension, as well as my personal experiences owning smaller powered monitors, and how the bass sounds vs larger woofers.
 
The low frequency drivers look heavy in the Devialet Phantom. Are they simply using brute force to move them or am I missing something?
I'm pretty sure they are. No way those big plastic woofers don't weigh a ton

They design better amps then speakers. Phantoms are impressive feat of engineering but have host of issues that make them not much more than a gimmick. The SAM technology begs to be licensed out to the mass market instead of being relegated to a module on their highest end amplifiers.
Compression in the bass regions at higher SPL. Very disingenuous of Devialet to market it as being full range down to 14hz when you really wouldn't want to be playing it that low at loud volumes. It kind of irks me but I guess the marketing department behind the phantom range of speakers is completely separate entity from their engineering department.

I have a lot of issues with the Phantoms and their marketing too, but on acoustics, you have to elaborate. Almost every speaker with DSP either begins to compress bass at higher SPLs or is purposefully limited in overall. That's kind of the point - taking advantage of DSP in order to usefully extend the bass at typical listening levels.

But that aside, I think devialet's claims are justified. How high SPL are you expecting? I have not measured the larger speaker, but I remember playing 20Hz test tones on the base model of the larger Phantom and them churning them out with ample volume being no problem. The smaller Reactor 900 I did measure is nothing to scoff at. T=Here is their quasi-anechoic spin:
index.php


And here is compression per SPL with two speakers playing at a distance of 10 feet in a large room. Speakers are 1-2 feet from the front wall - a very normal listening setup. The purple line is 100 percent volume.

index.php


They claim 18Hz -6dB and that is what they achieve at realistic listening levels in a large room. Bass compression only begins around 95dB. 95 dB in a home is LOUD. Very few people regularly listen above 75 dB. That still gives me 20dB of headroom for dynamic peaks, which almost no music will actually use, and very few movies. Their biggest problem is that they have too much bass(and now way to EQ it in app). And if you really do need more power than this, that's what the biggest speakers are for.

Perhaps more importantly, I'm pretty sure there is nothing that is comparable in bass performance at its size at any price. You could argue for bookshelves plus subs, and in some scenarios that's advantageous. You could say their software sucks, which I largely agree with. But the acoustics are very much worth the praise for their price and arguably beyond.

(I feel like I've become an unwitting devialet champion of late, which is funny because outside of audio I've complained about other issues).
 
Last edited:
I'm not trying to hate on the Phantoms. But they just seem too good to be true. If its any consolation, I am seriously considering auditioning them sometime in the future. They are much cheaper then Kii 3 and Dutch and Dutch 8C, and I could pocket the money to use towards a larger HDTV which I've been wanting for a while now. If I'm impressed perhaps I can just move to a streaming setup. Moving away from your traditional hifi setup, passive loudspeakers, with seperate amp, pre, dac, and source sounds mightly appealing. My listening room/bedroom will feel a lot less cluttered and I could fit a nice TV console on my wall, and fit more of my video games and knicknacks to free up closet space for my clothes and shoes which clutter the kitchen... the sacrifices I have made for my hifi are numerous, and downsizing to an all active system is something I've wanted to do for over a year now.

I get that feeling, but being on ASR goes both ways. Sometimes claims are too good to be true and the data proves them wrong. Other times they're backed up by data. When I first heard the reactor, I was shocked at their output. When I measured them, I was probably more shocked - their performance is something most studio monitors could be jealous of.

The Phantom's biggest issue is around the software. I had them disconnect on me a few times, other times the volume on one speaker inexplicable became louder than the other. They are less stable than my a Sonos System or LSX. Not enough that I wouldn't still recommend them, but enough to have a reservation. I don't know if devialet has improved the software since I reviewed them though. But mainly, I just wish the app had EQ! Easy enough to do via a receiver, but it means many will call them bloated for having too much bass.

Edit: Though to be fair, Devialet's marketing tagline for the reactors is "unreasonable sound.":) So it makes sense one would be dubious.
 
Last edited:
Anyone know what damping material KEF is using in the LS50? Borrowing the idea of a damped brace seems promising. It almost certainly adds cost in manufacturing. I know about the usual prospects (sonic barrier, sorbethane sheets, etc.), but most of these are useless for serious dampening of bass frequencies.
 
They were pre-built, but are now only DIY. The cabinets are not wood, but MDF (so really a wood-based composite). MDF is used because it has a good tradeoff between damping, stiffness and cost. So would not count out all boxes too quickly. OTOH, if you really want to eliminate boxes, suggest you check out linkwitz lab or some of the open baffle designs. Happy hunting!

MDF I'd count as "not wood" as it doesn't have grains and there's a lot of glue and processing involved. It is a good material for cost considerations, and I would still count out boxes, as the less it's shaped like a box the better (rounding things and not having parallel walls and such.)

I'm okay with "Boxes" if they look like Genelec 8000 series monitors, a Devialet Phantom, or Cabasse pearl. I'm not aware of any inherent tradeoffs (besides cost) that come from not having parallel walls, not using wood, etc.

The reason I'd like to eliminate cost as a variable is because I'm curious what the ideal is... and I'm pretty sure it's not the JBL M2, which I'd like to hear, as I'm very skeptical of the engineering behind the entire design, minus the tweeter and waveguide. Everything else seems archaic and like it focused on cost too much for a $20,000 pair.
 
I'm not trying to hate on the Phantoms. But they just seem too good to be true. If its any consolation, I am seriously considering auditioning them sometime in the future. They are much cheaper then Kii 3 and Dutch and Dutch 8C, and I could pocket the money to use towards a larger HDTV which I've been wanting for a while now. If I'm impressed perhaps I can just move to a streaming setup. Moving away from your traditional hifi setup, passive loudspeakers, with seperate amp, pre, dac, and source sounds mightly appealing. My listening room/bedroom will feel a lot less cluttered and I could fit a nice TV console on my wall, and fit more of my video games and knicknacks to free up closet space for my clothes and shoes which currently clutter my kitchen... the sacrifices I have made for my hifi are numerous, and downsizing to an all active system is something I've wanted to do for over a year now.

The only reason they are too good to be true is because there are zero other companies even trying to compete with them concerning output, bass, and size. If more companies were concerned with the bleeding edge, we'd have dual opposing woofers from Focal and JBL that were in full range monitors that were shoe box size... But that's not what JBL and Focal care about.

They are much cheaper, and you could also spend the saved money on room treatment, and I highly recommend moving away from an "audiophile stack." The only thing you will be downsizing is the physical volume of the speaker... These things go stupidly loud.

If they could be repaired/I had Devialet care, and there were no software issues, I'd have ended my speaker search there. Instead, I had to spend twice as much for a used pair of Genelec 8260 monitos to get back into the same league as the Phantom Reactors when it comes to bass extension and vertical dispersion... Just can't stand the app issues on multi-thousand dollar speakers. Genelec has GLM software that is way more complex of a setup with cables and microphones and and external box, and literally "just works." It's a glass of cool water compared to the hell that was the Devialet app, or heaven forbid, their Dialogue box. They need a non-smart, no app version. Just have optical and RCA inputs and an extra wire or KLEER to link the speakers and call it a day.
 
The only reason they are too good to be true is because there are zero other companies even trying to compete with them concerning output, bass, and size. If more companies were concerned with the bleeding edge, we'd have dual opposing woofers from Focal and JBL that were in full range monitors that were shoe box size... But that's not what JBL and Focal care about.

They are much cheaper, and you could also spend the saved money on room treatment, and I highly recommend moving away from an "audiophile stack." The only thing you will be downsizing is the physical volume of the speaker... These things go stupidly loud.

If they could be repaired/I had Devialet care, and there were no software issues, I'd have ended my speaker search there. Instead, I had to spend twice as much for a used pair of Genelec 8260 monitos to get back into the same league as the Phantom Reactors when it comes to bass extension and vertical dispersion... Just can't stand the app issues on multi-thousand dollar speakers. Genelec has GLM software that is way more complex of a setup with cables and microphones and and external box, and literally "just works." It's a glass of cool water compared to the hell that was the Devialet app, or heaven forbid, their Dialogue box. They need a non-smart, no app version. Just have optical and RCA inputs and an extra wire or KLEER to link the speakers and call it a day.

Yep, my thoughts precisely. If there were a way to wire the speakers to each other directly to avoid the Wi-Fi issues, almost all of my complaints would be gone. Literally, they could come out with a studio monitor version of the same speaker with proper I/O and EQ switches and make bank in the pro industry.

Edit: Devialet does offer a "Phantom Reactor Custom" version of the Phantom Reactor for integrators.
 
Last edited:
MDF I'd count as "not wood" as it doesn't have grains and there's a lot of glue and processing involved. It is a good material for cost considerations, and I would still count out boxes, as the less it's shaped like a box the better (rounding things and not having parallel walls and such.)

I'm okay with "Boxes" if they look like Genelec 8000 series monitors, a Devialet Phantom, or Cabasse pearl. I'm not aware of any inherent tradeoffs (besides cost) that come from not having parallel walls, not using wood, etc.

The reason I'd like to eliminate cost as a variable is because I'm curious what the ideal is... and I'm pretty sure it's not the JBL M2, which I'd like to hear, as I'm very skeptical of the engineering behind the entire design, minus the tweeter and waveguide. Everything else seems archaic and like it focused on cost too much for a $20,000 pair.

Forgive me if I missed, but are you mainly listening to music, what type of music, what size room, any other constraints (spouse, neighbors, speaker size, location in room, etc.)?
 
Could we keep the posts on-topic please?
 
Forgive me if I missed, but are you mainly listening to music, what type of music, what size room, any other constraints (spouse, neighbors, speaker size, location in room, etc.)?

I'm just talking about what an ideal design would be and the merits or drawback I don't know about, and questioning the design decisions that I don't understand when it comes to cabinets on high end speakers.
 
Compression in the bass regions at higher SPL. Very disingenuous of Devialet to market it as being full range down to 14hz when you really wouldn't want to be playing it that low at loud volumes.
They did at one point list extension at a few output levels, then I assume marketing removed it.
It's fun to play music with very deep bass loud on them and watch the drivers go crazy, but it's quite distracting when listening to music properly, being side mounted and white they are very visible when flapping about.
 
I'd like to see a lot more on measurements that correlate to audibility for construction issues, and then work back to how designers solved the problem. I'm sure this is an area where good engineers with access to good measurement tools can solve the problem cheaply, where others can only throw money at the problem.
 
I'll copy my post from the other thread.

This is a Magico cabinet, but I've done the same thing as DIY with a friend. Layers glued together. We used inexpensive wood or even thicker layers of MDF. You can with care and time do this with a jig saw. When doing with MDF we also angle each cut and could curve the enclosure top to bottom as well as the two dimensions shown here. We also built some layered like this putting threaded rod through top to bottom. You let the bottom be feet, and clamped it all together with lots of force using a metal top and bottom plate to spread the load over wood. Cover the top of the rod with a decorative wood piece.

This was heavy, and mostly impractical as a commercial design. It made for one heck of a stiff and damped cabinet. We were crazy to do something this labor intensive, but that is one advantage a DIY has. Now with less expensive CNC setups you could program it and knock it out in layers pretty quickly.

For lighter designs we'd use 2x3 and 2x4 to build a lattice frame, cover with thin wood. We'd attach the wood with permatex and screws. Based upon the idea the soft adhesive and many screw joints would damp and dissipate energy. We'd didn't have any way to really design it for best results, but results were pretty good. I guess closer to Harbeth (and BBC) ideas.

1580853074768.png


1580853085170.png


Also the cabinets my friend and I did years ago were shaped like the truncated pyramid from an earlier post for low diffraction. We had the offset version of this for tweeter and midrange.
1580853259761.png
 
I get that feeling, but being on ASR goes both ways. Sometimes claims are too good to be true and the data proves them wrong. Other times they're backed up by data. When I first heard the reactor, I was shocked at their output. When I measured them, I was probably more shocked - their performance is something most studio monitors could be jealous of.

Really sounds like @amirm should get hold of this seemingly magical speaker and see how it performs in the NFS…
 
Really sounds like @amirm should get hold of this seemingly magical speaker and see how it performs in the NFS…

Maybe he should ask the company. They have nothing to hide unless my measurements are totally off:). I'm actually really curious about how it does below 100Hz and in sound power, as a limitation of the nearfield bass splice is you can't really measure at angles. Also had some trouble figuring out how to do baffle step compensation for the chassis but I'm confident the results are accurate enough. Kind of wish I'd spent more time with them to test a few other things I'm curious about.

Keeping this on topic with cabinets, there's something interesting that happened with testing the phantom reactors and their capsule-shaped cabinets.

See that little scoop around 3K? When I first measured the speaker, it was quite a bit deeper, which seemed at odds with my listening, but was more akin to soundstage network's measurements of the bigger phantom, which shows a much bigger dip in this region. I always thought this was weird since I didn't remember the original phantoms being recessed in this region, but we all know how unreliable audio memory is.

I noticed the Phantom's were quite sensitive to placement on the stand I use for measurements. Normally I align the front of the speakers with the front baffle but because the Phantom's baffle is so curved, it seems the wavefront kind of wraps around the baffle and causes a reflection in the gap between the high frequency unit and the stand.

Snag_3a90fbbc.png
So instead I had to push the speaker all the way forward such that the driver didn't "see" the stand, and that largely rectified the dip, though some of it remains. Can't recall any other speaker being so sensitive to positioning on its stand.

I can't help but wonder if something like this happened with the soundstage network measurements because it seems so at odds with the rest of the frequency response. I wish I'd saved those measurements to demonstrate. I imagine this is a big part of the reason the reactor's stand is a tripod - minimizing any stand reflections. So PSA - if you listen to these, push them all the way forward on whatever surface they're resting on.
 
Back
Top Bottom