Devialet show a world beyond the black box at RMAF ’14


rmaf_2014Chrome dreams. If there’s a more future-facing hi-fi company than France’s Devialet, I’ve yet to encounter them. Despite being the ultimate fingerprint magnet their all-in-one devices are a sight to behold. These French fellas know that aesthetic attraction is fundamental in getting the consumer to step closer, to ask questions and perhaps purchase. I’ve witnessed their DAC/amplifier units at numerous shows and always come away feeling like I’ve been wowed for the very first time.

At their Paris headquarters, Devialet now employ eighty people of which forty-five are engineers. Then there’s the recently launched showroom in Shanghai. These guys aren’t messing about – they are growing fast.

So what’s going on inside their chromed pizza boxes? Seventy patents is what.

ADH technology sees pure Class A generate the voltage signal and Class D provides the current and never the twain shall meet. The musical signal reportedly travels along one of the world’s shortest signal paths – two inches – and is never exposed to Class D.

Along the long wall of their RMAF demo room a pair of Devialet 400 configured as monos can be found driving Focal floorstanders. So visually impressive is the Devialet look I omitted to take note of the loudspeaker specifics. Oops. A heavily bearded Christopher Brightwell tells me that there’s a guy in Europe who runs B&W Nautilus with eight (8!) Devialet monos.


Then there’s Devialet’s SAM technology – ‘Speaker Active Matching’ – which our hosts are at pains to point out is emphatically not digital re-EQ-ing. It works like this: people can vote to have their loudspeaker model of choice shipped to Devialet HQ where seventy metrics are taken and a virtual model created in a DSP environment. The speaker’s amplitude and phase are then corrected but only below 150Hz – like our friends at DEQX, Devialet know that messing with the midrange and high frequencies is a fool’s errand.

Eighty loudspeaker models have been SAM-d to date with Devialet checking off ten more as each week passes. Brightwell says the most commonly voted loudspeaker manufacturers are B&W, KEF and Sonus Faber whose ginormous Amati Futura will be SAM-d very soon!

Along the short wall the Devialet chaps have the Athom GT1.0. These French standmounts have already been through measurement-and-correction wringer. SAM takes their bass response from 45Hz to 22Hz – zut alors! A factoid that is readily discernible during in the playback demo; with SAM engaged via remote control the GT1.0 sound ballsier with better dynamic punch. Know that the delta isn’t always as dramatic. The Magico S1 reportedly don’t benefit from SAM processing as Devialet would have hoped.


Lest you assume that each Devialet box is just for digital audiophiles, allow me to correct you (as I was corrected in Denver). The Devialet software configuration system can pull all sorts of trickery. Each of the analogue inputs can be configured to accept a phono signal – MM and high output MC on the 120, MM and MC on the 200 and up – after which loading and gain can be selected. In a neat twist, Devialet units can be instructed to automatically switch of WiFi as soon as they see an analogue input signal. Compelling stuff.

Curious? Check out the software simulator on the Devialet website.

With everything packed into a single box – DAC, amplifier and a matching remote control that also oozes stylishness – you can expect to see further coverage of Devialet on these pages as we move towards the end of the year.

Further information: Devialet

Written by John H. Darko

John lives in the NOW + HERE = NOWHERE. He derives an income from the ad revenues of DAR. John is also an occasional staff writer for Stereophile, 6moons and TONEAudio.

Twitter: DarkoAudio
Instagram: DarkoAudio
Facebook: DAR


Leave a Reply
  1. I also admire Devialet’s attention to quality and detail in their electronics’ design, but their visual aesthetics make my teeth start rotting. And (since I lack a 6-figure income) I am not very taken to their prices, either. This is, of course, merely my personal opinion. Summarizing: I treat their products as an interesting curiosity – nothing more than that, not even drool worthy…

  2. I’ve yet to hear a Devialet (I try to stay away from expensive stuff to protect my wallet/savings), but they are truly cool. There are many products with an excellent sound, and many with exciting new features, but the way that Devialet provides both, at a not-completely-unreasonable price, is probably unique.

  3. Devialet’s 11k $aus setup at the Melbourne AVShow this weekend was smashing. Beautiful effortless midrange, not quite the treble geschpritz of some other setups, but given you get the kitchen sink for their price I felt it was a compelling alternative to ‘rack’-based multi-component systems.

    • 100% agree with you Stefan. Not only that but standmount speakers, more appropriate for a hotel room, also contributed to that space being one of my favs of the weekend.

  4. I’ve been looking into this and I’m pretty certain Devialet’s technique is quite different to DEQX’s.

    It appears that Devialet is doing something similar to linear predictive coding (although I suspect it isn’t actually LPC), where there is a mathematical model of the speakers with parameters determined from measuring speakers. Each audio sample is passed through the model and a new value calculated at the end which compensates for the effects of the crossover and driver.

    In contrast DEQX appears to be determining convolution filters which are then multiplied with input samples to produce the output.

    It sounds similar but Devialet is working more in the time domain whereas DEQX is working in the frequency domain. It would be interesting to see if DEQX can produce the time domain corrected graphs that Devialet produce. DEQX seem to be focussing on frequency response and group delay as a whole rather than the specific signal.

    I could be wrong in my interpretation of what they are doing, but they do appear to be fundamentally different approaches and the only similarity is that they both focus on the lower frequencies. My suspicion is that the Devialet is more advanced but may not cover as many aspects as DEQX (such as room correction).

    To put this another way Devialet are using a technique that corrects the time domain signal, as a result a lot of other aspects are corrected such as frequency response and group delay, and potentially other benefits. Whereas the DEQX technique specifically tries to fix the frequency response and group delay without specifically aiming for an accurate time domain response (although you would tend to think that would improve).