By mitchco •Devialet Phantom Gold Loudspeaker Review
The Devialet Phantom Gold loudspeaker is one of the most technically advanced “all in one” wireless speakers I have ever heard or seen. I initially listened to them at the Vancouver Audio Show and was surprised by the sound quality coming out of these oblong spheres. When I got home from the show, I looked them up and read some interesting technical specifications that seem to defy the laws of physics. How can this small package reproduce such low frequencies? My curiosity got the better of me and I asked Chris if he could get a pair for review.
When they arrived, I was surprised again how small and heavy they were. Roughly 10 inches in diameter, 13.5 inches in length and 25 lbs. Yet the Phantom Gold contains an on-board computer, multiple speaker drivers, 4500 watts of peak power amplification, digital to analog converter, and multiple connectivity options.
I must say, I was floored by the sound coming out of these speakers. The dispersion characteristics of this loudspeaker is outstanding, largely due to the shape of the speaker eliminating enclosure diffraction effects. Frequency response sounded full to my ears, even reproducing the lowest octave, as if a subwoofer was somehow hooked into the system. An astonishing technical feat in such a small package.
I am following same format as established in my review of the Dynaudio Focus 600 XD. First a technology overview of the speakers, including setup, followed by subjective listening impressions and quantified with objective measurements. I am not going to repeat the details of the review approach here, but if interested, one can read how I employ industry guidelines for subjective listening and objective measurements to establish a reference baseline in which to compare the speaker under review to.
Phantom Gold Technology Overview
With 108 patents, this speaker is a tour de force of advanced technical design and engineering. Analog Digital Hybrid technology combines Class A and Class D amplification. The Phantom is built around two hermetic woofers that function under high pressure to reproduce levels down to 14 Hz. The Active Cosperical Engine designates the spherical design of the Phantom, inspired by the “thrusting sphere of Olson”, the perfect acoustical shape. This is referencing Harry F. Olson’s landmark paper on “Direct Radiator Loudspeakers Enclosures” from 1951. The idea is that a sphere is an ideal shape to reproduce sound and spread its linear energy in all directions, without any detrimental sound diffraction from the surface of the loudspeaker enclosure.
Speaker driver compliment is one titanium tweeter, aluminum midrange driver and two aluminum bass drivers. 4500 watts of peak power per Phantom with 20 Hz to 20 kHz ±2 dB frequency response and a -6 dB bandwidth of 14 Hz to 27 kHz. 108 dB SPL at 1 meter maximum output. System on a chip (SoC) Cyclone V ARM Cortex-A9 MPCore 800MHz dual-core processor, physical IP, and FPGA with 512MB DDR3 memory. Devialet DAC embedded with 24 bit/192 kHz operation. With the purchase, comes free firmware upgrades.
Multiple Connectivity Options
The Gold’s can be connected via Airplay, Bluetooth (A2D and AVRCP profiles, aptX, AAC, SBC audio codecs), Spotify Connect, proprietary network Wi-Fi Dual-band (a/b/g/n 2.4 GHz & 5 GHz, Ethernet RJ-45 10/100/1000Mbps, CPL Homeplug AV2 and Toslink optical input.
I used Phantom Dialog to connect my Lynx Hilo via Toslink to the Dialog, then wireless to the Golds. In addition, I used the Spark app to setup and connect to the Phantoms on my Win 10 computer, but I did not use Spark’s music management capabilities.
The Spark app setup is an interesting process as it involves activating the loudspeakers via touch. Check it out:
The speakers emit a quiet ethereal sound when going through the steps. When one puts their hand on top of the speaker, another sound emits along with physical movement from the woofers showing activation. An interesting tactile experience that somehow one has brought the speakers to life by merely touching it.
By clicking continue, one is presented with the ability to configure one to many Phantom’s in different locations. In my case, it is a stereo set up in my living room:
Clicking on continue provides a dialog box that indicates setup complete and the next step is to configure music sources:
I configured Spark for local music from my PC.
Here the Dialog is setup so that the music source is from my computer and the optical output from the Lynx Hilo is connected to the optical input of the Dialog, which then transmit the signal wirelessly to the Phantoms. I skipped over creating an account and finally one lands on the Spark apps main page where one can select Sources, Devices and manage music with a Player:
That’s it. I was able to launch JRiver Media Center and start playing music. I did not evaluate other sources or devices or any other options as the main focus is the sound quality of the Phantoms. However, it should be clear that one can hook up virtually any source and be playing music in about 5 minutes. Kudo’s to Devialet for a smooth, and interesting way, to set up speakers.
Subjective Listening Impressions
The introduction is intended to establish a vocabulary for correlating subjective descriptions to an objective frequency response range. However, I described that in detail in the first review of the Dynaudio Focus 600 XD, and not going to repeat here. I have followed the exact same procedure, listening level, etc., as described in the previous article. However, I feel it is worthwhile to put up Bob Katz’s frequency response chart with subjective descriptions that correlate to an objective frequency response range:
Phantom Gold Listening Impressions
The Gold’s are set up in an equilateral triangle in the exact same spot as my reference speakers and where I reviewed the 600 XD’s:
If one looks back at the 600 XD review, I have similar picture from this angle to compare.
The Phantom Gold’s sound unlike any speaker I have heard. A fundamental contributor to the speaker’s dispersion characteristics, is its enclosure shape. The tweeter and midrange are housed in a spherical shape. As mentioned at the front of the article, inspired by the “thrusting sphere of Olson”, the perfect acoustical shape. Check out Harry F. Olson’s landmark paper on “Direct Radiator Loudspeakers Enclosures” from 1951.
If you get a chance, audition a pair of these speakers in a stereo setup. The dispersion is simply amazing. No matter where I move across my 6 foot couch, I do not notice any tonal change in frequency response, and the image is completely stable. In fact, I can go far off center in the room and the high frequency response still sounds the same – really amazing characteristic of this loudspeaker.
While waveguides (i.e. a type of horn) are an alternative to achieving constant or controlled directivity, it is over a limited coverage range. For example, my JBL 4722 cinema speakers employs a constant directivity waveguide with a 90 degree horizontal by 50 degree vertical dispersion pattern. That’s just enough to illuminate my couch area, but moving beyond that, the high frequency rolls off, in a predictable and smooth way.
Speakers like BeoLab 90 and Kii Three use multiple drivers and DSP to control directivity and the 600 XD’s I reviewed, use DSP to control speaker resonances and directivity through the cross over range. Speakers like the MBLs produce a full omnidirectional output. For further reading on the importance of loudspeaker directivity or polar response, I refer folks to read Dr. Earl Geddes excellent article on, “Directivity in Loudspeaker Systems.”
The Gold’s offer an interesting design to achieve a wide dispersion pattern using a spherically designed enclosure, yet with a full range sound, in such a small package. I was surprised how much and how low the bass went on these speakers. Make no mistake, they are full range speakers. The bass drivers are unusual to say the least, but solid sounding, with no cabinet resonances. I pushed them quite hard on content I know have reasonably deep bass and still sounded good near peak reference level (i.e. 105 dB SPL at the listening position). While the sound was clear, the bass drivers were visibly vibrating and looked the loudspeaker was going to take off.
Are they giant killers? Perhaps a bit unfair to compare to my high efficiency cinema speakers, but still amazing levels of low bass energy come from these speakers. The frequency response is remarkably similar to the 600 XD’s, just delivered in a much different fashion. The sonic signature or tone quality has a bump in the low end and a flat high frequency response.
We will get into the details in the objective measurement section, but I do want to say, while initially the little bit of extra bass and flat extended frequency response, at reference level, has a tendency towards the “boom tiss” type of tone quality. On rock music, especially if it has been overly dynamically compressed, has a tendency towards too much high frequency energy coming at my ears. Initially it is like candy to one’s ears, but over time, my preference would have been towards less of a bass bump and more of a sloping (i.e. tilt) roll-off frequency response on the top. To be sure, that is my subjective preference and it is splitting hairs as the audible difference is subtle.
To my ears, operating the loudspeakers within their recommended range, I did not hear any distortion. The bass response does sound different than a traditional transducer in a wood box. I don’t have any words to describe it really, just different. Again, I recommend auditioning the loudspeakers to hear the difference with your own ears.
The midrange and top end sounds smooth and transparent. At the listening position, some 9 feet away, I do not hear any audible hum or hiss coming from the Gold’s. Moving my ear up close the speaker, I do hear audible hiss coming from the tweeter, but not objectionable at the listening position.
For such a small package, I am amazed at its omnidirectional dispersion characteristics and full range frequency response.
While this review is intended to be mostly about the sound quality of the loudspeakers, their subjective design really elicits people’s opinions to be one extreme or the other. People either are blown away by the aesthetic design, or… not so much. My wife and daughter said that if they were offered in black, would be an improvement. Now, part of the issue here folks is that my wife and daughter are used to having industrial type speaker designs in my listening room since forever, as I like big waveguides and woofers in black painted “washer size” cabinets. Pretty lucky that the WAF is so high in my household
As far as the speaker’s coherency (i.e. timing) is concerned, they sounded coherent, but in a slightly different way than my time aligned loudspeakers. Of course, I now have hindsight after measuring the speakers, but I must say the speakers timing and imaging sounded very good to my ears.
Listening Impressions Summary
Full range, smooth frequency response, both on and off axis. The off-axis frequency response is simply amazing, unlike anything I have heard to date. Same with the bass response, very unbox like. Amazing full range sound quality coming from such a small package.
Objective Measurements - Introduction
I had written a large section on introducing objective measurements in my previous review of the 600 XD and not going to repeat here. If one has questions about the measurement approach, please read the objective measurements intro section in the 600 XD review first, as that may answer one’s question.
However, I do want to leave a frequency response chart here as it will become useful to compare objective measures to subjective preferences.
This is from, “The Measurement and Calibration of Sound Reproducing Systems”, which is a free, open access Audio Engineering Society Paper, by Dr. Floyd Toole. In Dr. Toole’s paper, turning to page 17:
My subjective preference is the “trained listeners” curve. I calibrated my reference sound reproduction system using DSP, so that it is a similar frequency response as measured at the listening position and listening area. That target curve sounds neutral to my ears. Note the untrained listeners curve has a tendency to emphasize the low end and a slightly elevated top end, lending to a “boom, tiss” type of tone quality. I call it the “candy curve”, meaning like eating candy, tastes really good for a while, but too much and…
Objective Measurements – Phantom Golds
Here is the frequency response, as measured at the listening position in my room:
For an in-room response, a very smooth response, even with the room ripples. The different low end response between left and right speakers, are due to the fact that the left speaker (i.e. red trace) is more towards to the left hand corner of my room as my stereo is offset to the left of center in my room. That can be seen via the photo, earlier in the article. Other than that, both left and right speaker track each other very well, with a dead on match from 2 kHz onwards.
I have included my preferred target frequency response, which is flat to about 1 kHz, with a straight line to about -6 dB at 20 kHz. This, to my ears, is a neutral sounding response. As can be seen, the Gold’s have some bottom end lift and no high frequency roll off or “tilt” down on the top. Leading a bit to a “boom tiss” sound, but the boom isn’t really a boom as there is no cabinet resonance and the tiss is not really that as it is so smooth sounding. However, I listen to mostly overly compressed rock music, which has a tendency to put too much high frequency energy towards my ears already. Folks that listen to classical music on these speakers may not have that preference, preferring the flat to 20 kHz tone quality instead. Again, it all comes down to preference, and these speakers are certainly in the range as compared to the various subjective listening curves in the Toole chart.
Also note, low frequency extension to 14 Hz. There was significant output at 10Hz, when I started my measurements, which sweep from 10 Hz to 24 kHz. Almost defying the law of physics from such a small package.
Off axis response? Check this out. What you are seeing are two sets of left and right speaker measurements. One set is with the measurement mic, moved 3 feet off to the left of center, and the other set, 3 feet off to right of center. That’s 6 feet between the two mic positions and provides a good “in-room” indicator of the tone quality across my 3 seat, 6 foot couch at the listening area:
Quite a feat actually. Especially the top end, look how smooth and extended it is. No wonder I hear very little variation in tone quality across a wide listening area. From 5 kHz to 20 kHz is amazingly flat response with very little roll-off. While there may be a tad too much high frequency energy for my tastes, it is remarkably smooth, even off-axis.
Here is a frequency response comparison of the Phantom versus the 600 XD’s that I reviewed previously. Everything is identical relative to the software and equipment used, right down to the speakers and measurement mic in the exact same locations, with the only difference being the loudspeakers themselves:
The red and green traces are the left and right Phantom Gold’s respectively. The brown and blue traces are the left and right Dynaudio 600 XD’s respectively.
Note how close the frequency responses are. The Gold’s have a wee bit more bottom end and tad more top end in the 10 kHz range. Otherwise, they are virtually identical tonally, and almost following the room ripples identically. Once could argue that if these two speakers were measured in an anechoic chamber, they are likely to measure almost identical, yet the two speakers are physically very different from each other. Something to ponder…
Time Coherence – Step Response
What about the speakers timing response? This certainly is an area of controversy as there is still today, no definitive listening tests to show if time alignment impacts the listening experience. My personal preference, and listening tests, indicate there is an impact. I prefer a speaker that is time aligned. For now, I feel if folks read the 600 XD time alignment section, time coherence is well explained there.
In the case of the Phantoms, the tweeter and midrange appear to be time aligned, with the woofer just slightly behind:
I know it is a bit difficult to interpret what is shown here. The brown “step” response is calculated as the “ideal” step response given a speaker where all frequencies are arriving at ones ears at the same time. In the case of the Phantom Golds, we are seeing the tweeter and midrange arriving first and the woofer just slightly behind. While the Phantom Golds are not “textbook” time aligned, they do follow the general shape of the ideal step response. Again, it is tricky to determine if it is audible or not. This to me is an area that requires more controlled listening tests to come to a similar understanding that we now know about preferred frequency responses ala Dr.Tooles paper referenced in the previous section.
For some preliminary technical reading on speaker time alignment, see Rod’s article, “Phase, Time and Distortion in Loudspeakers.”
I am using John Mulcahy’s REW program below where I measure the Phantom’s and comparing them to my reference speakers, which are custom built (and since sold), 3-way floor standers using digital XO, linearizing and time aligning the drivers, and then applying some overall amplitude and excess phase correction at the listening area, which for me is a 6’ x 2’ grid area where my couch is. I have calibrated my speakers using DSP, so that they have the target frequency response that I prefer, along with time alignment. Again, everything else software and gear wise is the same, the only change is the loudspeakers themselves.
First the frequency response:
My reference speakers are the blue and purple traces and the Phantoms are the red and green traces. As one can see, the Phantoms response from about 80 Hz to 5 kHz are remarkably similar in frequency response to my reference and certainly my preference. A little too much bottom for my taste, but not in a bad way. Like I say, you owe to yourself to hear the remarkable bass that is un-box like (technical term) in sound quality. 5 kHz and above is a departure where the Golds are a bit brighter than my reference/preference.
The hard part to communicate while looking at the measurements is how big of a deal is it? I will say it is certainly audible. One can look at the frequency range and use a digital eq, (like in JRiver DSP Studio for example) and dial in eq on your own speakers that boost the low end and top by the relative difference amounts as shown in the chart above. While listening at reference level (i.e. ~ 83 dB SPL C weighting at the listening positon), switch in and out the eq, in real-time, which will give you a rough idea of the tone change. It is not a lot, but certainly audible.
This is the first 5 milliseconds of sound arriving at the microphone at the listening position. The reference is pretty close to the calculated ideal step response. With the Phantom, the drivers are not quite time aligned as the woofer pulse is slightly behind the midrange and tweeter. Or at least that is the best I can discern from my measurements, as I have not seen this type of step response before.
Having played for years with speaker time alignment, I have come to the same conclusion as Rod Elliot in his article, “Phase, Time and Distortion in Loudspeakers, “For what it's worth, I originally started this article not to praise, but to debunk the theory that time alignment is the only way a speaker should ever be designed. Having done the research, run tests, and written the article, I confess that I must agree with many (perhaps even most) of the points made by the time alignment proponents. My overall opinion, based on the research for this article (primarily tests and simulations), is that time alignment is a very good thing, and perhaps all speakers should be designed this way.”
I must say I am really impressed with the Phantom Golds. No subwoofer required to get some real low end frequency response reproduction at reference level, and beyond. The spherical shape does wonders for a linearly dispersed sound, no matter where one listens in the room. This has all sorts of practical applications where one is not forced to sit in “the listening window” to get the best sound quality.
There certainly was enough power on tap to drive to peak reference level (i.e. 105 dB SPL) without sounding overly stressed. As mentioned before, the bass character is different than a boxed loudspeaker, hard to describe, but worthy of an audition if one is tired of box loudspeakers.
It would be great if there were a few frequency response “profiles” built into the Phantom and that allowed one to select to better match to ones preference. Or the ability to shelf eq the low end to bring the overall level down just a bit. And the ability to “tilt” down the high frequency response (not a shelving eq) so when listening to overly compressed rock and pop music, the high frequency energy coming at ones ears is reduced. That would be my preference.
A small nit that some may or may not notice, but there is an overall system delay to the sound of the Phantoms as a result of the DSP processing. So, if one is watching a movie the sound slightly lags behind the video. I noticed this on lip sync, but after a while it was close enough that it did not bother me. I told my wife about this and it did not bother her at all.
As far as the shape or color of the Phantom’s themselves, that’s a subjective call. I find the shape interesting, but I would have preferred the speakers in black and gold, instead of white and gold. But its unique shape and technical design, that sets its sound quality apart from other traditional box speakers. I still can’t get over how linearly dispersed the sound was. That really sets these speakers apart from the majority of others.
If the thought of a small, full range speaker, that does not require a sub, and with wide dispersion characteristics are of interest to you, then the Phantoms are well worth an audition.
- Product - Devialet Phantom Gold $2,990 each, Dialog $329
- Phantom Gold Product Page - Link
- Dialog Product Page - Link
Where To Buy:
Mitch “Mitchco” Barnett. I love music and audio. I grew up with music around me, as my mom was a piano player (swing) and my dad was an audiophile (jazz). My hobby is building speakers, amps, preamps, etc., and I still DIY today. I mixed live sound for a variety of bands, which led to an opportunity to work full-time in a 24-track recording studio. Over 10 years, I recorded, mixed, and sometimes produced over 30 albums, +100 jingles, and several audio for video post productions in several recording studios in Western Canada.
I wrote this book to provide the audio enthusiast with an easy-to-follow step-by-step guide for designing a custom digital filter that corrects the frequency and timing response of your loudspeakers in your listening environment so that the music arriving at your ears matches as closely as possible to the content on the recording. Accurate Sound Reproduction using DSP. Click on Look Inside to review the table of contents and read the first few chapters for free.
Edited by The Computer Audiophile0