samedi 28 avril 2012

Test - Stello U3 (English version)

April Music, founded in 1998, is a South Korean company editing 3 brands: Aura, Eximus and Stello. A highly skilled staff, often musician themselves, is surrounding Simon Lee, the founder. He is taking a particular care at manufacturing his pieces of electronic to produce highly musical elements. Passionate, well known for having a particularly sharp ear, he is offering his customer a wide range of electronics. Today, we are going to put our attention on a very useful accessory to compensate the lack of an optimized USB input on a converter.

Computer audiophile music has this unique interest in the fact that it allows combining of a whole variety of elements, either at transport level, either at computer or streamer level; It is bringing a quality of output which could not be hoped for people with limited budgets a few years ago. At another level, streamers are providing us with an integrated solution (transport + dac) and at the opposite, computers often shared for domestic usage are not so well optimized for audiophile music.

With the massive arrival of DACs on the consumer market, in parallel, one has developed accessories to transform computers into high performing “state of the art” transports. As of today, the USB connection from manufacturers is all in all equivalent using proprietary drivers or not, it is resulting of the price drop and popularization of USB chips; On the other hand, the margin of progression at operating system level, playback software and S/PDIF conversion section remains important. This is naturally why USB accessories manufacurers including Stello turn towards this direction.


Stello U3 is manufactured as a rounded squared metal box, rigid, rather heavy which straightly inspires confidence and will show up as very resistant to:

- Interferences as of metal made;
- and to vibrations as it is very rigid.

The bottom is totally covered with a thick anti-skating rubber, an excellent choice to insure device stability. Until set, it does not move anymore, even connected to thick and heavy connectors/cables.

Connectors have a particular good treatment that deliver a prosperous yet warm image through their gold plated RCA and XLR Neutrik plugs. Providing with 110 Ohms symmetric link on top of regular S/PDIF 75 Ohms is an excellent idea because of device market positioning. Numerous high level converters are often equipped with an optimized symmetric section, offering them computer connection at least cost and therefore sustaining audiophile investment.

On this board where components are not on top of each other, a lot can be found. A 13MHz clock is dedicated to XMOS processor management, where two other external clocks are used to generate 44,1 KHz and 48 KHz frequency multipliers. As on some other products of the same type, the reduced size of the XMOS component, its powerful 500 MIPS (millions of instructions per second) and its power consumption optimization (an average of 112mA) allows to plan a good reserve of current to draw for the rest of the components. In addition, a red LED on the face shows up when the device is powered and locked by the computer. Taking the product positioning in consideration, we would have appreciated to have an optional battery-powered supply.

So, Stello offers a product which is perfectly in line with the new manufacturing standard of consumer market into providing USB components out of XMOS. On Mac an Linux system, the box does not require any specific pilot to work, relying on the embedded operating system one. On a Microsoft Windows platform, things remain unchanged as Audio Class 2.0 is not implemented and the Redmond firm does not plan, neither communicates any roadmap concerning this support on Windows 8. So, it stills require a driver, which by the way integrates DS, KS, WASAPI support but also an ASIO implementation.

Our tests have been performed using the Kernel Streaming mode. In the next section, we will explain how to make a device setup for this mode

Let’s take note that XMOS does not develop its own asynchronous driver, but rely on a partner, a German company called Thesycon. This well know driver which does not really change from one product to another, offers a very poor integration at operating system level. Recalling me about 90’s drivers, settings have to be adjusted in a startup group application that remains in the tray and thus consume resources.

Thesycon control panel allows adjusting of fundamental parameters such as streaming buffer size for asynchronous transfer. You can adjust it from 1 to 32 milliseconds. I recommend to directly setting it to 1 ms; if it works fine this way, no need to make any step to adjust this value anymore and generate unnecessary jitter:

Let’s take note that ASIO buffer size is also available from the same screen, it can be set to a minimum of 2 ms only (from 176 to 768 samples)

All the other parameters are grayed, except volume adjustment that I advise you not to change. Changing volume would introduce additional jitter and might waste bit perfect flow by doing on the flow data changes.

In a nutshell, Stello U3 device does not derogate to April Music’s rules, “built to last”. Immediately, the box impresses, we now need to check whether this very positive feeling on its built transforms into facts on its musical performance.

System used for the review: 

Especially built for digital music playback:
  • Apple Extreme switc with Full Duplex GB ports
  • Ethernet Cat 6 shielded and braided cables, plugs are also shielded
  • Laptop with an external switched power supply based on Core 2 Duo with 6 MB RAM, SSD disk of 80 GB. PC is liked to network at 1GB/s Full Duplex.
  • Operating System is Windows 7 Ultimate x64 SP1, tuned for audiophileoutput conforms to the PC Audiophile.
  • A Phonic FireFly DAC powered on battery
  • A wireworld platinum Digital S/PDIF cable.
  • A NAS of 2 TB linked to the network at 100MB/s Full Duplex
  • An iPhone 4 to be used as remote control paired to wi-fi network on the Apple Extreme access point
  • Software to play tracks is foobar linked with JPlay V4 (Beach + Hibernation/Fullscale) or JPlay mi
  • ni in mode Hibernation/Fullscale
The rest of the output chain is at iso perimeter to compare the elements:
YBA Diamond, Pre/Amplifier based on DIY Vecteur I6.2, ProAc D28.


First contact is very positive, we are definitely in presence of a well performing transport system. Transparency is perfect evidence, like the feeling of being totally free of color on the medium range. The soundstage is very well defined, one of the best I have ever tested with such a device. Voices are splendid, rich in details and clarity, but never too dry. Eric Clapton on Next time you see her (slow hand 24/96KHz) offers us a world of nuance in the course of his scratched voice, full of veracity.

Let’s take note that, as usual, it is only a question of taste to like this kind of musical signature or not. You easily forget about the Stello box, music takes over technical aspects, dynamic is really present, forte pressure and jump from one range to another seemed to me very realistic.

On HD tracks, Stello transport is without any concession, let’s say brilliant! Listening to Monty Alexander on Calypso Blues (24/96KHz) delivers a very realistic percussions and a rich and surprisingly varied piano, it was easy to guess piano brand at first notes! We are now very close, as long as the S/PDIF cable is also of high grade, to the best transports I could ever hear and especially close to famous Linn Akurate DS2 streamer.

Less fleshly and ethereal than the hiFace EVO style, one will willingly raise volume up without any “digital hardness” feeling in the higher range. On well-engineered recordings, like those from Diana Krall, music is especially fluid and breathes. On From this moment on (24/96KHz) out of the same name record, swing from brasses charms us as well as the deep voice from the singer. Romeo and Juliet from the Russian national orchestra (Alexander Vedernikov - 24/88.2KHz) perfectly simulate the big orchestral mass without any excess or tension.

In order not to derogate to the rule, we have also used our usual test tracks. We have especially appreciated to listen to the classical Take Five where it was really easy to hear with very accurate precision the chords move of the bass. 

Another good point, applause and crowd atmosphere were both very well transcribed; On Sade live, one will also appreciate the accurate precision of the scratched voice accents from the singer. Plane separation is naturally also excellent; Space between instruments and voices is very precise. Apparent bandwidth seems to be perfectly coherent, without any feeling of holes or lack at the start or end of the audio range. Extinctions of notes generate from time to time a frustrating feeling, not because of the system, but because of engineering recording lacking of precision. Lower part of the bass range is clean, never giving a feeling of being slow. 

In front of Internet radio flow, the Stello U3 box does not reacts a very exceptional way. I mean that you can feel compression too much. Listening to high definition channels (256 or 320Kbits/s) brings a better feeling, but seems to be a bit tiring to listen for hours. It might be interesting to swich EQ on at this time!

In terms of listening enjoyment, this box offers an equal pleasure on all kind of music styles, transparency and respect of timbres appeared to me in perfect line with requirements to listen to classical or jazz concerts as well as rock. In a nutshell, I would say that voice performances are brilliants. Cecilia Bartoli delivers here a strikingly true Sposa relying on a truthfully chord background. Lastly, Katie Melua is close to us, at a few inches on Piece by piece bathed in a firm and coating bass line.


As I like to say, using devices such as U3 remains very touchy at a pure technical point of view. It requires pairing the right pieces of electronics to get the expected result; today, we are in front of a very stringent product, but so stringent that a mistake of taste is not to exclude. A few more fleshly products or more colored could outstrip it in the heart of a few audiophiles attached to a specific sound signature!

As a conclusion I think that U3 deserves its own place on a stringent and robust pantheon of audiophile computer world. Even if at the technical level, there are no special very high level components compared to competition, assembly and topology seem to reach target, this is what, at the end of the day, counts.

Savor without any restriction!

Manufacturer link :

Link on French Audiophile forum :

6moons (anglais) :

Spécifications : 

compatible with asynchronous USB Audio Class 2.0 mode
bit perfect audiop stream up to 192kHz
2 external clocks at low jitter

22.5792MHz/24.576mHz to generate multipliers of 44.1kHz/48kHz
44.1, 88.2, 176.4, 48, 96, 192kHz/24bits Audio
1 USB 2.0, Type B connector
galvanic isolation and up to 192kHz
1 RCA  75 Ohms, gold plated
1 Neutrik XLR AES/EBU 110 Ohms, gold plated

LED on front panel

Os Support: Windows XP, Windows Vista, Windows 7Apple Mac OS X natif

Dimension :
100 X 36 X 100mm (WHD)

Banc d'essai - Stello U3 (version Française)

April Music, fondée en 1998, est une entreprise Coréenne du Sud éditant 3 marques, Aura, Eximus et Stello. Une équipe de haut vol d'artisans souvent musiciens entoure son fondateur, Simon Lee, qui porte une attention toute particulière à la fabrication en se focalisant dans le but de produire des appareils d'une musicalité d'exception. Passionné réputé pour avoir une oreille particulièrement affûtée, il met notamment à disposition de ses clients une gamme très complète d'électroniques. Nous allons aujourd'hui nous intéresser à un accessoire fort utile pour compléter les convertisseurs qui ne disposent pas d'une entrée USB optimisée.

La musique dématérialisée à cet intérêt qu'elle permet une multitude de combinaisons de matériels tant au niveau des sources de transport comme ordinateurs ou streamers, tout en apportant un niveau de restitution que l'on pouvait qualifier d'inespéré pour les petits budgets il y a encore quelques années. Le streamer apportant une solution intégrée (transport + dac), l'ordinateur quant à lui, n'étant qu'un ordinateur souvent domestique, n'est souvent pas directement optimisé pour une utilisation audiophile.

Avec l'arrivée massive des DACs sur le marché s'est développé un marché parallèle des accessoires de couplage aux ordinateurs offrant une lecture des fichiers numérisée optimale. Si le couplage USB est à peu près uniforme de nos jours avec ou sans pilote propriétaire suite à la baisse et la vulgarisation des puces de support USB 2.0, la marge de travail reste grande sur la partie système d'exploitation, logiciel de lecture et section de conversion S/PDIF. C'est naturellement dans cette direction que se tournent les éléments actuels et notamment ceux fabriqués par Stello.


Le Stello U3 se présente sous la forme d'un boîtier métallique aux coins arrondis, rigide et plutôt assez lourd qui d'emblée inspire à la fois confiance et se montrera fort résistant :

- aux interférences du fait de sa nature;

- et aux vibrations du fait de sa grande rigidité.

Le socle est totalement recouvert d'un patin en caoutchouc anti-dérapant, excellent choix pour assurer sa stabilité. Une fois posé, il ne bouge plus même sous le poids de connecteurs/câbles assez rigides.

Les connecteurs sont particulièrement soignés donnant une image cossue, RCA dorée et prise XLR d'origine Neutrik. La fourniture de liaison symétrique en 110 Ohms en complément du support S/PDIF en 75 Ohms est une excellente idée en raison du positionnement de l'appareil. Nombre de convertisseurs haut de gammes sont souvent pourvus de sections très optimisées en symétrique et leur offrir un couplage de source par ordinateur à moindre coût pérennise donc l'investissement.

Sur cette carte à l’espace relativement aéré sont installés de nombreux composants. L'horloge de 13MHz est dédiée à la gestion du composant XMOS, les deux horloges supplémentaires servent à gérer les multiples des fréquences de 44,1KHz et 48KHz. Comme sur d’autres produits du même type, on notera que la taille réduite du composant XMOS, sa puissance de 500 MIPS (Millions d'Instructions Par Secondes) et l'optimisation de sa consommation (en moyenne 112mA d'après les spécifications) permettent d'envisager une réserve de courant disponible pour le reste des étages et des filtres confortable sur un port USB courant. De ce fait, cerise sur le gâteau, la faible consommation du système alimenté par la prise USB de l'ordinateur est complétée par une LED rouge signalant l’alimentation. En raison du positionnement, on aurait aimé en complément la possibilité d’une dérivation d’alimentation sur batterie.

Stello propose donc un produit parfaitement en ligne avec le nouveau standard manufacturier du marché du fait de l’utilisation de composant de couplage USB XMOS. Sur environnement Mac et Linux, son fonctionnement ne nécessitent aucun pilote spécifique. Sur plate-forme Microsoft Windows, USB Audio Class 2.0 ne fait l'objet d'aucun pilote standard du système d'exploitation de la firme du Redmond à présent, ni prévu pour sa version 8. Un pilote fourni par Thesycon, partenaire XMOS offre sous Windows le support DS, KS et WASAPI, ainsi qu’une implémentation ASIO.

Nos essais se sont concentrés sur l’utilisation du mode Kernel Streaming. Nous allons donc revenir sur le paramétrage et l’intégration au système.

Le pilote Thesycon est connu et d’un produit à l’autre ne présente que peu de différence, son intégration à Windows reste donc quelque peu spartiate. Les réglages ne sont pas disponibles par le panneau de contrôle du son, mais via un panneau de configuration sous forme d'une application dédiée. Un mauvais point à Thesycon à une époque où l'interface utilisateur revêt une importance capitale.

Depuis le menu démarrer, dans le groupe de démarrage de Windows, le panneau de contrôle a été ajouté, vous retrouverez donc l'icône Thesycon dans la barre de tâche à l'avenir.

Le panneau de contrôle livré permet d'ajuster quelques éléments fondamentaux dont notamment le tampon interne (streaming buffer size) pour le transfert asynchrone. Une échelle de 1 à 32 ms est disponible. Je recommande de régler la valeur à la plus petite valeur pour démarrer : 1 ms. Si tout fonctionne correctement, inutile d'introduire de la latence complémentaire inutile et nuisible à la gigue (jitter) :

Notons que le réglage de la latence en mode ASIO est également réglable sur le même écran :

L’ensemble des autres paramètres sont tous grisés, sauf le réglage du volume que je conseille de ne pas toucher. Toucher le niveau de volume signifie intégrer de la latence incontrôlable par traitement du signal à la volée.

A titre de résumé, le Stello U3 ne déroge donc pas à la règle de l'entreprise April Music, "construit pour durer". D'emblée, l'appareil impressionne, reste à vérifier si l'impression fort positive sur la construction se traduit également musicalement parlant.

Le matériel utilisé :
Spécifiquement prévu pour une restitution dématérialisée, il est composé :

  • D'un commutateur (switch) Apple Extreme avec ports Gb Full Duplex
  • De câbles Ethernet Catégorie 6 blindés et écrantés courants du commerce, les prises sont également blindées
  • D’un pc portable alimenté par une alimentation externe, Core 2 Duo, 6 Mo de RAM et disque SSD de 80 Go. Le PC est connecté au réseau en Ethernet à 1Gb/s Full Duplex.
  • Le système d'exploitation est Windows 7 Ultimate x64 SP1, l'optimisation apportée au système est conforme aux recommandations sur le PC Audiophile.
  • D’un DAC Phonic FireFly moyen de gamme alimenté sur batterie
  • Un cable Digital S/PDIF Wireworld Platinum
  • D'un NAS de 2 To connecté à 100Mb/s Full Duplex
  • D'un iPhone 4 servant de télécommande avec l'application Remote connecté en wifi sur l'Apple Extreme
  • Le logiciel de lecture est foobar2000 couplé à JPlay V4.1 (Beach + Hibernation/Fullscale) ou JPlay mini en mode Hibernation/Fullscale
Le reste de la chaîne de reproduction reste à iso périmètre pour les comparaisons, à savoir des câblages YBA Diamond, Préamplification et Amplification à base Vecteur sur mesure, Enceintes ProAc D28.

L'écoute :

Les premières impressions sont fort positives, on est en présence d’un appareil de transport performant. La transparence est d’une parfaite évidence, tout comme l’absence d’excès de coloration dans le medium. La scène sonore est parfaitement définie et d’une rigueur digne des meilleurs produits que j’ai pu tester. Les voix sont splendides, claires et riches en détail, sans présenter une impression d’excès de détails offrant une impression d’aseptisation. Eric Clapton sur Next time you see her de l’album Slow Hand en 24/96KHz nous offre un éventail de nuances au détour de sa voix éraillée empli de véracité. 
Notons enfin que, comme à l’habitude, il s’agit d’une question de goût pour apprécier ce type de restitution claire. Le produit se laisse volontiers oublier L’impression de dynamique me semble bien présente, la tension des forte et des sauts de tessiture sont assez réalistes. 
Sur les plages haute résolution le transport est sans concessions, bref excellent. A l’écoute de Monty Alexander sur Calypso Blues en 24/96KHz les percussions sont fort réalistes et le piano riche en nuance, on reconnait volontiers la marque de l’instrument à la première note ! On se rapproche bien volontiers, pourvu que le câble S/PDIF soit également excellent, à un mouchoir de poche des meilleurs transports que j’ai pu entendre et notamment de l’excellent streamer intégré Linn Akurate DS2. 

Moins charnel et plus éthéré dans le style que le hiFace EVO, on montera aisément le son sans sensation de « crispation numérique » ou dureté dans le haut du spectre. Sur des enregistrements soignés, comme ceux de Diana Krall par exemple, la musique est particulièrement fluide et aérée. Sur From this moment on en 24/96 de l’album du même nom, le swing des cuivre envoute complété par la voix grave de la chanteuse. Le Roméo et Juliette de l’Orchestre national Russe (Alexander Vedernikov) en 24/88.2KHz restitue fort bien l’ensemble de la grande masse orchestrale sans excès ni crispation.

Afin de ne pas déroger à la règle nous avons également utilisé nos plages de test habituelles. Nous avons particulièrement apprécié l’écoute du classique Take Five où l’on pouvait distinguer les amplitudes des cordes de la contrebasse avec une précision et une justesse diabolique.

Encore un bon point, les applaudissements et ambiances de foule sont fort bien reproduits, sur le live de Sade on appréciera également la précision très complète sur les accents éraillés de la chanteuse. La séparation des plans est naturellement de premier ordre et l’espacement des instruments et voix fort précis. La bande passante suggestive semble parfaitement cohérente, sans sensation de manque ou de trou dans les extrêmes. Les extinctions de notes provoquent de temps à autre quelques frustrations non pas à cause du système, mais de l’enregistrement manquant de précision. Le bas du spectre est plutôt bien dégraissé sans excès, sans traînage.

Confronté au flux compressé des radios internet courantes, le Stello U3 ne m’a pas fait une impression exceptionnelle, on ressent naturellement la compression très facilement. L’écoute des stations dites haute définition (256 ou 320Kb/s) s’avère apporter une écoute moins frustrante mais fatigante à la longue du fait de la grande clarté, je pense qu’une égalisation active s’impose pour redonner un peu de couleur.

En terme de plaisir d’écoute, cet accessoire procure un égal plaisir sur tous types de musiques, la transparence et le respect des timbres m’étant apparue autant rigoureuse sur des concerts classiques, jazz ou rock. Enfin, à titre de résumé les performances vocales sont d’exception. Cecilia Bartoli nous livre ici un Sposa criant de vérité sur fond de cordes empruntes de justesse. Enfin, Katie Melua est à quelques centimètres sur Piece by Piece baignée dans une ligne de basse ferme et enrobante.

Conclusion :

Comme j’aime à le répéter l’utilisation d’accessoires tels que le U3 reste assez délicate si l’on se place d’un point de vue purement technique. Il convient d’associer les bonnes pièces pour obtenir le résultat souhaité, aujourd’hui nous sommes en présence d’un produit rigoureux mais justement si rigoureux qu’une faute de goût n’est pas à exclure. Quelques produits plus charnels ou colorés pourraient bien lui damer le pion dans le cœur de certains audiophiles attachés à un style de reproduction !

En conclusion je pense que le U3 a définitivement pour moi sa place dans une configuration dématérialisée aboutie et rigoureuse. Même si au détour de l’analyse de la partie technique on ne remarque pas de composants sortant du lot en terme de qualité par rapport à ceux des concurrents, l’assemblage et la topologie semblent porter ses fruit et c’est au final ce qui compte.

A consommer sans modération !

Lien du site constructeur :

Le fil de discussion sur le forum des audiophiles :

Article 6moons (anglais) :

Spécifications : 

compatible USB Audio Class 2.0 en mode asynchrone
Flux de données audio bit perfect jusqu'à 192kHz
2 horloges externes à très basse gigue

Fréquences d'échantillonnages :
22.5792MHz/24.576mHz pour les multiples de 44.1kHz/48kHz

Résolution 44.1, 88.2, 176.4, 48, 96, 192kHz/24bits Audio

1 USB 2.0, Type B connector

isolées galvaniquement et supportant jusqu'à 192kHz
connecteur RCA plaqué or 75 Ohms, 
1 connecteur Neutrik XLR plaqué or AES/EBU 110 Ohms

Indicateur de fonctionnement :
LED sur panneau avant

Support des plate-formes :Windows XP, Windows Vista, Windows 7Apple Mac OS X natif

Dimention :
100 X 36 X 100mm (WHD)

mercredi 18 avril 2012

M2Tech hiFace Two (english version)

This week I offer you to get back on M2Tech, this italian company I am following the activity for 2 years. M2Tech is an inovative company who has a good worlwide coverage in manu countries through its distribution network. 

This new version of the USB stick is having the same external shape than the previous one; a USB 2.0 Type A male plug, a (black this time) plastic body and a gold plated RCA female plug. If the external aspect does not reveal any special information, what about the inside?

A little bit of story:

So far… when M2Tech offered the first compact adapters with a USB plug and a very good integrated S/PDIF section, we had to admit that it was a winning strategy. The rest of the market walked quickly in their footprints; nevertheless the other actors did not apply a so good penetrating go to market strategy. 

A few moments after, some of them got the benefits of embracing the challenge like notably John Kenny who today offers a very nice packaged version, as well as Teddy Pardo who offers external power supplies for the EVO version. In parallel XMOS was growing, driven forward by large industry actors equities, into producing highly integrated processors specialized with some in USB Audio Class 2.0. The underlying technical and financial stakes consequences were that XMOs soon began to play a key role in offering a wide and powerful range of components (SU1, L2, L1 and G) scaled in terms of power, features ingeniously shaped to match business market need. 


M2Tech today aligns on the new manufacturer standards provided by XMOS, most probably to answer to a need of optimizing their R&D costs and develop a standard USB connection. 

On a Mac or Linux system, the stick does not require any specific pilot to work, relying on the embedded operating system one. On a Microsoft Windows platform, things remain unchanged as Audio Class 2.0 is not implemented and the Redmond firm does not plan, neither communicates any roadmap concerning this support on Windows 8. So, it stills require a driver, which by the way integrates DS, KS, WASAPI support but also (it is new) an ASIO implementation. 

At the end of the day, it does not matter, whichever pilot DS or ASIO is used, what matters for us is performance and quality of pilot programming, especially talking about Kernel Streaming mode. Today we will not test any M2Tech driver, but an OEM of XMOS asynchronous driver developed by their partner, a German company called Thesycon. This element seems to me sufficiently interesting so that we make a stop on it, especially on the settings to adjust on the Windows platform to get the best in class result in playback: 

The Windows integration is very poor from my point of view. Recalling me about 90’s drivers, settings have to be adjusted in a startup group application that remains in the tray and thus consume resources… With the native M2Tech drivers of version 1, the integration was better and settings were seamlessly integrated in the sound control panel. 

Let’s give a bad point to Thesycon for this very nasty control panel integration. 

Installing the M2Tech hiFace 2 Thesycon pilot: 

Run the installation file, then press on Next

If like I did, you try to install the pilot before you plugged the stick, then you will have an error message displayed without any possibility to get back with a button such as

Previous. You will have to close the application and run it again.

If your hiFace was plugged in, it will be detected and the setup will offer to install the files in a standard destination folder, click on Next:

Files will be copied to the destination folder:

Then, at the end of the installation process, check that the following message is being displayed: Preinstallation was successfull, then click on Next:

Click on Finish, installation is done:

From the startup menu, in the startup group, you will notice that there is a new item. You will find it and have a new system tray icon for the USB 2.0 driver.

Thesycon control panel allows adjusting of fundamental parameters such as internal memory buffer for asynchronous transfer. A 6 values scale is available. I recommend to directly setting it to the lowest value : Minimum latency; if it works fine this way, no need to make any step down to adjust this value anymore and generate unnecessary jitter.

Another important parameter that has to be adjusted is streaming. By default the pilot selects Power Saving mode, you have to select Always On.

Finally, let’s finish on a good point to balance the previous bad point. M2Tech hiFace EVO or Young driver does not conflict with hiFace 2’s one. Both are working fine in parallel on the same system.

We also noticed that the streaming section also offers under certain conditions and certain XMOS processors to manage 5.1 audio flows as well.


Having a look at the elements delivered by the manufacturer, we discover that, in addition to the USB section, first, a new output stage has been used and second, switching regulators have been replaced by linear versions. M2Tech offers us a product where the name is close to the previous one, where the look is also close, but inside, everything is totally different! It is thereofore perfectly legitimate to ask ourselves question on the fact that if comparing those two devices has really a sense or not? In my opinion no and listening has to be done as it was a totally new product.

Face view (XMOS US1002L1, 22.5792 clock, 24.576 clock, 13MHz clock)

On this very reduced space are installed plenty of components. The 13MHz clock close to the USB input is dedicated at the XMOS processor; the two others close to the RCA plug are used to generate frequency multipliers of regular 44,1KHz and 48KHz frequencies. It should be noted that the reduced size of the XMOS component, its powerful 500 MIPS (millions of instructions per second) and its power consumption optimization (an average of 112mA) allows to plan a good reserve of current to draw for the rest of the key, taking in consideration  what a regular USB port is allowed to offer.

Back view (atmel 1108)

At last, it should be noted that the manufacturer states that this new architecture prevents clicks and especially during gapless playback. This apparently insignificant detail is, in fact important because our speakers can suffer from this, like our ears! My tests confirm that this new version  after long hours of playback do not output any clicks.

System used for the review: 

Especially built for digital music playback:
  • Apple Extreme switc with Full Duplex GB ports
  • Ethernet Cat 6 shielded and braided cables, plugs are also shielded
  • Laptop with an external switched power supply based on Core 2 Duo with 6 MB RAM, SSD disk of 80 GB. PC is liked to network at 1GB/s Full Duplex.
  • Operating System is Windows 7 Ultimate x64 SP1, tuned for audiophileoutput conforms to the PC Audiophile.
  • A Phonic FireFly DAC powered on battery and linked to the hiFace using a regular S/PDIF digital cable.
  • A NAS of 2 TB  linked to the network at 100MB/s Full Duplex
  • An iPhone 4 to be used as remote control paired to wi-fi network on the Apple Extreme access point
  • Software to play tracks is foobar linked with JPlay V4 (Beach + Hibernation/Fullscale) or JPlay mini in mode Hibernation/Fullscale
The rest of the output chain is at iso perimeter to compare the elements:
YBA Diamond, Pre/Amplifier based on DIY Vecteur I6.2, ProAc D28.


An immediate review based on return of listening was difficult for me as the hiFace was new and never burned in. It was necessary for it to run for a few hours so that the electronic components get ready. For this purpose, I use a loop of tracks based on Densen disk for 10 hours, then ran it for an additional 8 hours on a specific playlist. Nevertheless, at first sight, even without burning it in, we can guess the overall benefit and defects of this key.

At start, the main feeling is transparency, with a little bit more color that it should be on the medium range of my test equipment. Voices are splendid, crystal clear, without offering any excess of details such as we can observe on monitoring oriented products. Feeling is at relaxing, softness, without any tension or digital hardness especially on high definition tracks. Regarding this last point, without reaching the brio of more expensive products such as EVO or Stello U3, I find that hiFace two is a success and, let's say, most probably more achieved than the first version.

Because we shall speculate at comparing, tonal color seems to be less achieved, music less fleshly in a general manner than on EVO; at the opposite, compared to hiFace One, musical message seems to be “better”, being at the same time different. So let’s ask ourselves if this is all about sound signature? I think so; I think that hiFace two does not reproduce music with the same tones, signature than the previous version 1.

Another good point is applause reproduction. It is very well reproduced and really feeling like natural. Planes separation is also excellent. Apparent bandwidth seems to be perfectly coherent, without any feeling of holes or lack at the start or end of the audio range. Extinctions of notes do not give any feeling of lack as well. When the hiFace two is used on too high quality electronics, we can see that the lowest part of the bass range outputs as a bit slowed, something that happens to be logical to me. As the signature of music appears as soft, without any excess of expression, attenuation of lowest bass is aligned with it. Brass section seems to be lacking of natural on classical music, woodwinds are better such as chords as forte are reproduced with adequate generosity. I think that one will more especially appreciate hiFace Two with modern music such as electronic, lounge or studio recordings.

In front of Internet radio flow, hiFace two reacts a very traditional way. I mean that you can feel compression and excess of color that some of the channels are adding to please teenagers. Listening to high definition channels (256 or 320Kbits/s) brings a feeling close to the Lounge way: good point!

In terms of listening experience, at first place, this accessory offers a very good surprise on electronic tracks, pop, rock, where its signal processing depicts an obvious efficiency. Vocal performance is unique, Cecilia Bartoli makes us cry on Sposa, as Katie Melua makes us vibe on her « cudling music » ; we willingly forget about microdetails offered by products of the upper classes. Target is then perfectly reached ; Similarly, with Cranberries or Sade performing live, both are offering a wide stage filled of true artistic emotion, isn’t it the main point? Let’s add that smoothness is really adapted to a long listening experience of many hours without any feeling of tiredness.

In a second step, I stayed for a long time examining classical music and acoustic jazz recordings. These two are requiring a very special brio to allow the audiophile emotion to be entirely fulfilled. If in the overall the result is excellent and will match to numerous middle class systems, music appeared as limited to pretend at coupling with high end systems to my ear.
In that regard, I think that one should choose the EVO box from the same manufacturer for upper class electronics. Even if in the overall, the points and counterpoints are well reproduced, the acoustic performance of Capuçon brothers (two part inventions) was letting a slightly misled, let’s risk, artificial soulless feeling on violin and cello. Let’s point out that on the first version of hiFace the results were the same, nevertheless at the advantage of the new version two, which brings a more coherent global picture, more details without caricature and especially without tension on the higher range part.


At a strictly technical point of view, usage of accessories such as hiFace two remains always touchy. It is necessary to pair the right pieces of electronics to get the expected result. This is in this way, that I think, that this new USB connection will be of good marriage with middle class product such as the one of 1K€ range. I think more precisely about Marantz, Atoll, Rega or Cambridge. These ones are not especially equipped with best of breed USB sections even if some have XMOS chips inside. For a little bit more than a hundred euros, the M2Tech solution has a real credibility. 

In terms of challengers, we can also find seducing opportunities with other products. Every one of them bring its own signature, such as Musical Fidelity VLink 192 or Audiophileo 1. Naturally, these products are much more expensive (close to 1.5), but the sound they bring could seduce. 

Should we thank M2Tech for this new type of swiss knife? I think yes, for many reasons: 

At first, it is a "tour de force" to offer a piece of electronics at this very competitive price in its range. At less than 150€, M2Tech offers a concentrate of technology able to transport music with a remarkable fidelity. It is taking advantage of actual technologies at the maximum, not only external DACs without XMOS chipsets will benefit of it, but also those that do not have a so good s/pdif converter section.
In second, M2Tech offers in this way a real evolution of its “mobile” range. If the hiFace One does not have to blush of its performance in the overall, taking the actual technology market improvement situation in consideration, hiFace two appears like a successful implementation regarding what the audiophile consumer expect in terms of digital / mobile music. Size, weight and simplicity of the product target at a young and mobile population, equipped of multi-usages laptops. Time where the driver will automatically Tweet current track information is not so far, isn’t it? 

Last thing, around product, once again, we demonstrate that despite the brio of the device, despite the care brought to the implementation of USB 2.0 Audio Class 2.0, quality of the S/PDIF cable seems to be as sensitive as before. One must be very careful to use a good digital cable as USB link is not enough by itself. 


Input 1 x USB A type male
Output 1 x RCA or BNC female

I/O Standard
Input USB 2.0 Audio Format
Output S/PDIF Stereo Digital Audio Format

Sampling Frequency
44.1kHz, 48kHz, 88.2kHz, 96kHz, 176.4khZ, 192kHz

Resolution 16 up to 24 bit

Dimensions 10.2(d) x 2.2(h) x 2(w) cm

Power Supply 5V DC from USB bus

Temperature from 0°C to 70°C

Weight 50gr approx

Link to manufacturer site: