How the heck does Devialet do it?

[Piranesi]

Clever adaptive DSP for bass and flatness and curved baffle shape helps with dispersion too.

Such compact subwoofers and well behaving directivity loudspeakers exist also from other companies, admittedly not so often though combined in a single package.

A complete system with a DAC, DSP, amplifier, and speakers would cost more for comparable performance however. There's cost benefits to be had by integrating them into a single package. And that's even before considering the much greater space needed, as well as the all-important WAF.

 

[solderdude]

So my question is, what gives? What kind of black magic is Devialet using to get bass like that in a device that tiny,

The woofer isn't similar to 'normal' woofers and the amount of power pumped into those 'motors' is huge.

 

[BDE]

Following some some details:
White-Paper-Phantom-en.pdf (audiofile.nl)
Caixas Devialet Phantom tocam melhor que falantes vinte vezes maior - AutoSom.net
On the second page you see that the woofer is special with his piston it looks more like an servo drive.

 

[oivavoi]

First of all, I love these little boom-boxes! All of the Phantom models have way too much bass relative to the rest of the spectrum, but that's usually ok for low to mid level listening. At higher volumes the small bass drivers is being gradually high-passed higher and higher in frequency to avoid over-excursion, evident by the compression tests that napilopez ran in that review.

This means that for all intents and purposes those little speakers are not full-range. They can't play 20 hz at an audible level.

Another issue with small speakers with low sensitivity is that they require enormous amounts of power to play loudish. This means thermal compression and probably audible energy storage in the drivers that will make them sound thick/heavy/slow compared to high sensitivity drivers.

While the Phantoms are the results of remarkable engineering, there's physicals laws in place that demands that any gain here must be paid for there.

Small boxes and deep extension means low sensitivity that requires huge power. Huge power in passively cooled closed system means thermal issues which leads to compression and distortion.
Small drivers taking huge power to move great distances will have a tougher time to slow down and start over again fast enough to compete with larger drivers with bigger and stronger motors.

In essence, small speakers start with so many crutches at the start of the race that they can't really compete all the way to the end. While they can certainly sound extremely good, which these things do, they're not capable of high-end performance due to the assholeness of physics.

Interesting thread. I think @Absolute nails it here IMO. Yes, it's impressive, but there are some unavoidable physical tradeoffs. To what degrees are those things audible? Difficult to say.

I've listened to the biggest Phantoms three times at two different dealers. Really wanted to like them and bring them home, which I ended up not doing. To me they sounded very impressive for the size, but just a little bit dull or "thick". That may be due to the eq profile, which is too bass heavy for my taste, or it may indeed be about energy storage, distortion or compression in the drivers. Or both. I have no idea. I would have been interested in participating in a blind shootout between the Phantoms and a regular high end box speaker which was equalized similarly, and see if I then had a preference for one over the other.

 

[Frank Dernie]

I put a edit in, thats news to me.

This is probably longer than you want to watch but shows how easy it is to get into! And how beautifully made and expensively tooled, so cheap compared to Kii and D&D IMO for the engineering involved.

 

[Frank Dernie]

means the Phantoms are about 10-12 dB hot below 80 hz

So Amir will love them

 

[NTK]

I wonder if they employ psychoacoustic tricks such as ones described in this paper.
https://www.researchgate.net/public...ual_Bass_System_with_a_Phase-Vocoder_Approach

Abstract (with my emphasis):

A bass enhancement technique based on a phase-vocoder approach is presented. Instead of direct bass boosting, the proposed method creates a bass impression by exploiting the psychoacoustic properties of humans. This technique is most useful in audio reproduction using​

small loudspeakers that have no low-frequency capability, where direct boosting will likely result in nonlinear distortions. In light of psychoacoustics, the bass effect is synthesized by augmenting the original signals with high-frequency harmonics. Unlike conventional methods that rely on nonlinear processing, the proposed method performs the required frequency transformation by using a phase-vocoder approach. Apart from frequency transformation, another key element of the proposed technique is the magnitude adjustment of the generated​

harmonics. The underlying principle for magnitude adjustment is based on a polynomial model of equal-loudness contours. The method is implemented on a digital signal processor with the aid of multi-rate signal processing. To validate the proposed technique, objective and​

subjective experiments are conducted for PC multimedia loudspeakers and handset loudspeakers. The subjective listening experiment followed the procedure of multi-stimuli with the hidden reference and anchor (MUSHRA), and the data were analyzed by using the multi-analysis of variance (MANOVA) method. As indicated by the results, the proposed technique proved effective in rendering bass impression with acceptable audio quality.​

 

[test1223]

I wonder if they employ psychoacoustic tricks such as ones described in this paper.
https://www.researchgate.net/public...ual_Bass_System_with_a_Phase-Vocoder_Approach

Abstract (with my emphasis):

A bass enhancement technique based on a phase-vocoder approach is presented. Instead of direct bass boosting, the proposed method creates a bass impression by exploiting the psychoacoustic properties of humans. This technique is most useful in audio reproduction using​

small loudspeakers that have no low-frequency capability, where direct boosting will likely result in nonlinear distortions. In light of psychoacoustics, the bass effect is synthesized by augmenting the original signals with high-frequency harmonics. Unlike conventional methods that rely on nonlinear processing, the proposed method performs the required frequency transformation by using a phase-vocoder approach. Apart from frequency transformation, another key element of the proposed technique is the magnitude adjustment of the generated​

harmonics. The underlying principle for magnitude adjustment is based on a polynomial model of equal-loudness contours. The method is implemented on a digital signal processor with the aid of multi-rate signal processing. To validate the proposed technique, objective and​

subjective experiments are conducted for PC multimedia loudspeakers and handset loudspeakers. The subjective listening experiment followed the procedure of multi-stimuli with the hidden reference and anchor (MUSHRA), and the data were analyzed by using the multi-analysis of variance (MANOVA) method. As indicated by the results, the proposed technique proved effective in rendering bass impression with acceptable audio quality.​

What makes you think they did something like this?

 

[Frank Dernie]

I wonder if they employ psychoacoustic tricks such as ones described in this paper.
https://www.researchgate.net/public...ual_Bass_System_with_a_Phase-Vocoder_Approach

Abstract (with my emphasis):

A bass enhancement technique based on a phase-vocoder approach is presented. Instead of direct bass boosting, the proposed method creates a bass impression by exploiting the psychoacoustic properties of humans. This technique is most useful in audio reproduction using​

small loudspeakers that have no low-frequency capability, where direct boosting will likely result in nonlinear distortions. In light of psychoacoustics, the bass effect is synthesized by augmenting the original signals with high-frequency harmonics. Unlike conventional methods that rely on nonlinear processing, the proposed method performs the required frequency transformation by using a phase-vocoder approach. Apart from frequency transformation, another key element of the proposed technique is the magnitude adjustment of the generated​

harmonics. The underlying principle for magnitude adjustment is based on a polynomial model of equal-loudness contours. The method is implemented on a digital signal processor with the aid of multi-rate signal processing. To validate the proposed technique, objective and​

subjective experiments are conducted for PC multimedia loudspeakers and handset loudspeakers. The subjective listening experiment followed the procedure of multi-stimuli with the hidden reference and anchor (MUSHRA), and the data were analyzed by using the multi-analysis of variance (MANOVA) method. As indicated by the results, the proposed technique proved effective in rendering bass impression with acceptable audio quality.​

It doesn't look like it from the plots @Absolute posted, it shows signal at the actual low frequencies. I think the sort of speakers that exploit this sort of effect are the really tiny ones and whilst the Phantoms are small for their frequency response they are not small compared to the typical computer speakers for which this technique would be useful.
A treble, mid and 2 effectively 8" bass units is, after all the same driver complement as the Dutch & Dutch 8c but with more power, so certainly no surprise in @Absolute plot or my findings with mine.

 

[pjug]

A complete system with a DAC, DSP, amplifier, and speakers would cost more for comparable performance however. There's cost benefits to be had by integrating them into a single package. And that's even before considering the much greater space needed, as well as the all-important WAF.

This is true, and the engineering is impressive. But in my opinion the WAF is low, at least from photos. Me being my own W for the sake of discussion.

 

[boXem]

This is probably longer than you want to watch but shows how easy it is to get into! And how beautifully made and expensively tooled, so cheap compared to Kii and D&D IMO for the engineering involved.

One has production volumes with 4 to 5 zeros, the two others are craftmen.

 

[Frank Dernie]

This is true, and the engineering is impressive. But in my opinion the WAF is low, at least from photos. Me being my own W for the sake of discussion.

I am not a massive fan of the styling but it is really tiny compared to a conventional speaker with similar extension and that makes quite a difference to how readily one can find a non obtrusive place in the room for it.

 

[Frank Dernie]

One has production volumes with 4 to 5 zeros, the two others are craftmen.

Really?
I hadn't realised the Phantom production volume was that big, though it would probably need to be to pay for the tooling which will have been hugely expensive.
The most expensive Phantom is a fraction of the price of Kii and Dutch & Dutch as well, of course.

 

[boXem]

Really?
I hadn't realised the Phantom production volume was that big, though it would probably need to be to pay for the tooling which will have been hugely expensive.
The most expensive Phantom is a fraction of the price of Kii and Dutch & Dutch as well, of course.

I don't have the production volumes for the Phantom, in my mind it sells much more that the Expert. One of my suppliers told me that they quoted one component for the Expert with yearly volumes of 40k .

Edit: not only they are much cheaper for the customer, but dealer margins are much higher.

 

[Frank Dernie]

Edit: not only they are much cheaper for the customer, but dealer margins are much higher.

That is clever engineering then!
As an engineer I look at a Phantom and a Kii and would expect the Phantom to cost at least double to make, maybe more but maybe their investment in tooling makes the components inexpensive, like with cars. One of the senior engineers of one of the big car makers told me they spend millions on tooling so the BOM is cheap and spend more per car on marketing than on making it!

 

[richard12511]

being flat anechoically to 20Hz (in fact level below 100-200 Hz) leads to a bass heavy in room balance.

While this is true, Napier's measurements show that the Phantoms are not actually flat in the bass. They soom to be bass boosted by a few db above flat anechoic. For example, 60-80Hz looks boosted ~3-4dB above ~1kHz. My guess is that they're doing this intentionally to counteract our ears' bass insensitivity at quieter than reference listening levels. While they're certainly good enough for critical listening, they seem to be targeting the lifestyle crowd more than traditional speakers. Ime, that style of listening is usually at softer levels to allow people to talk easily over the music.

 

[Purité Audio]

Mitchco’ s review/measurements,
https://audiophilestyle.com/ca/reviews/devialet-phantom-gold-loudspeaker-review-r654/

Keith

 

[theyellowspecial]

Ugly as sin but great measurements.

 

[Chrispy]

I just don't "get" the desire for these. Seems more flash than substance.

 

[GDK]

I just don't "get" the desire for these. Seems more flash than substance.

If you are looking for a speaker that is (a) wireless, for domestic reasons, and (b) powerful, because sometimes you get the place all to yourself, it is a fairly short list. I like my pair, but I don’t think that they are the be all and end all. However, they are perfect for what I needed.