Arturia’s first foray into the world of audio interfaces squeezes a huge amount of connectivity into a compact desktop box.
French manufacturers Arturia are probably best known for their synths and drum machines, but they have plenty of expertise in developing computer software and hardware. Over the years, they’ve made numerous products that are designed to integrate with Macs and PCs, but the AudioFuse is the first fully fledged audio interface to emerge from their Grenoble HQ.
Although it’s a desktop interface, and a relatively compact one at that, the AudioFuse is incredibly feature-rich. It’s as though Arturia’s designers asked everyone they knew to submit feature requests, then decided to implement all of them. So, for example, not only does it incorporate conventional mic, line and instrument inputs, but you can also connect a turntable, while one of the line outputs also does duty as a re-amp out for guitar recording. Monitor control likewise goes well beyond the ordinary, with speaker switching, talkback, mono and dim functions all accessible at the press of a button. Oh, and the AudioFuse also acts as a three-port USB hub. Yet all this functionality has been packed into something about half the size of your average house brick.
Like a half brick, the AudioFuse is squarish, rather heavy for its size, and comes in several different colours. Unlike a half brick, it’s festooned with buttons, white LEDs and dials, and has a snug fitting lid to protect it from damage in transit. It’s solidly built, but although Arturia describe it as having “a great-looking exterior”, I thought it a bit too dumpy and crowded to be considered beautiful. The AudioFuse’s aluminium case seems to do duty as a heat sink, and gets uncomfortably warm after extended use.
Arturia have followed UA’s lead in operating their LEDs in three modes: off to tell you that a function is unavailable, dimmed to indicate that it is available but inactive, and bright white to show that something is both available and switched on. It looks as though they’ve implemented the ‘dimmed’ setting by strobing the LEDs, which then flicker in a mildly distracting fashion.
Unusually, the AudioFuse has a micro USB socket for connection to the computer. Adjacent to this is a socket for the supplied PSU; neither the power nor the USB socket uses a locking connector, but they mate pretty securely. The unit can be bus-powered, and to this end, comes with a custom USB cable that has two plugs at the computer end, allowing it to draw power from two USB sockets. Even so, there isn’t quite enough juice available on the bus to run it at full tilt, so you have to select one of three compromises involving reduced output headroom or disabling inputs.
The complete AudioFuse I/O count includes four analogue inputs and four outputs, plus two headphone outs. These are complemented by an optical digital input and output, a pair of RCA phonos which can be used for either S/PDIF or word-clock in and out, and MIDI in and out. For space reasons, the MIDI sockets are on mini-jacks, with short adaptor cables terminating in five-pin DIN connectors supplied. A nice touch is the inclusion of a USB hub offering three full-sized Type A sockets, which are perfect for connecting iLok dongles and the like.
Arturia’s marketing material for the AudioFuse places a lot of emphasis on their ‘DiscretePRO’ mic preamps, which are featured on the first two analogue inputs. Each has its own analogue gain control, along with switchable phantom power, polarity inversion and pad. Input sources are connected using ‘combi’ mic/XLR sockets on the front panel of the unit, which automatically sense whether an XLR or a jack is plugged in. They are also routed through quarter-inch analogue inserts on the rear panel.
The AudioFuse comes with an ‘Achievement Certificate’ containing Audio Precision measurements of each preamp’s equivalent input noise, frequency response and gain range. The first of these is given as -131dBu, which would be even more impressive if they hadn’t chosen to quote an A-weighted measurement. The frequency-response plot suggests that the -3dB points lie below 5Hz and above 100kHz, and that the preamps are flat to ±0.06dB within 20Hz-20kHz. The gain range runs from 2dB-75dB, which should be more than adequate for every conceivable purpose.
The second pair of analogue inputs are duplicated on quarter-inch jacks and phonos. The latter are intended for attaching a turntable, while the jacks operate at line level, though one of them can also be switched to act as an additional high-impedance input for connecting electric guitars and the like. There are no physical gain controls for these inputs, but level adjustment is available in the unit’s control panel software.
On the output side, there are two pairs of quarter-inch jacks labelled Speaker A and Speaker B, which are switched from the top panel. It’s also possible to treat these as two independent output pairs, within certain routing restrictions we’ll look at shortly. The S/PDIF output can duplicate either of these, which could be useful if you have monitors with a digital input. Unusually, not only are there two headphone outputs, but each has both a full-sized and a mini-jack socket. Each also has its own mono button, which acts independently of the one affecting the main output; on the review unit, these did not apply the correct attenuation to the summed signal, but this should have been fixed in a firmware update by the time you read this.
Like many other USB audio interfaces, the AudioFuse is class-compliant, and Arturia claim as a benefit that there are: “No fiddly driver installations — we’ve made the decision to use generic drivers and worldwide USB2 to ensure you’re always ready to record, unlike proprietary drivers that need to be developed and maintained after each operating system update.” Which, to my mind, is a bit like a sports-car maker explaining that they’ve chosen to use Ford Fiesta engines instead of designing their own, because there are lots of spares available in case they ever go wrong.
Personally, I’ve never found audio interface drivers more fiddly to install than anything else, and I’m perfectly happy to do so if custom drivers will give me better performance. The idea that not having to do this is an advantage seems especially tenuous given that everyone will need to install the Arturia Control Center utility — and that on Windows, the Control Center installer includes an ASIO driver, which is necessary to use the device with most audio software! On the plus side, its class compliance does mean that the AudioFuse is compatible with Linux, Android and iOS as well as Mac OS and Windows.
The Control Center presents a single pane, the central section of which reproduces the top panel of the AudioFuse. Controls that appear in both hardware and software are mirrored, so pressing a physical button on the AudioFuse makes its virtual counterpart light up, and vice versa. There are a few AudioFuse controls that operate in the analogue domain, such as channel 1/2’s gain controls, and these are therefore omitted from the Control Center. To balance things out, the Control Center sports a number of software-only controls for which space could not be found on the AudioFuse itself. These include clocking and digital input settings, talkback mic gain and routing, and gain controls for inputs 3/4.
Should you not be able to achieve sufficiently good low-latency performance to monitor in real time through your recording software, the Control Center is also the key to accessing the AudioFuse’s built-in direct monitoring. And this is where things can get complicated, or at least, where they sometimes felt that way to me.
The AudioFuse presents up to 16 inputs to your recording software: four analogue inputs, stereo S/PDIF, up to eight ADAT channels depending on sample rate, and another two-channel input which carries the mono talkback input on both channels. Your recording software can also address the ADAT or S/PDIF digital outputs directly, but when it comes to the analogue outputs, there is an extra layer of abstraction. DAW software ‘sees’ three stereo outputs, but these are not hard-wired to particular outputs: they are, in effect, buses, which can be routed in various different ways depending on the AudioFuse’s configuration. By default, the Main bus appears at whichever of the Speaker A and B outputs is selected, while the Cue 1 and Cue 2 buses are routed to the two headphone outputs. However, any of these outputs can be switched to pick up any of the other buses instead, and it’s also possible to set things up so that the Speaker B output duplicates the second headphone output rather than acting as an alternate speaker output, thus giving you a total of four independent line outs.
So far, this is all reasonably logical, but the way in which these interact with the direct monitoring is unconventional. Direct monitoring is implemented using a pretty basic digital mixer, which gives you fader, pan, solo, mute and channel linking controls for all the analogue and digital inputs (apart from the talkback mic). These sources are mixed down to a single stereo output, which can then be routed to any or all of the Main, Cue 1 or Cue 2 monitoring buses.
The odd part is that the stereo output from this mixer is fed back into the monitoring path in the analogue domain, using a single balance control. One knob thus controls the global balance between directly monitored signals and those emanating from the computer, and does so for all the outputs, no matter which of the cue buses they’re picking up. In other words, assuming you are using the AudioFuse in its default configuration, the main output and both headphone outputs can each be addressed separately by your DAW, but only a single mix of the direct signals is available, and each output receives this at the same level, or not at all.
An effective implementation of direct monitoring always involves a trade-off between versatility and simplicity. Desktop interfaces are all about the latter, so it’s no surprise that Arturia’s approach tends that way, but I’m not totally convinced they’ve got the balance spot on. The global balance control and the additional ‘cue bus’ layer of abstraction between computer and speakers compromise the AudioFuse’s ease of use somewhat, but they don’t add greatly to its flexibility: there’s only one direct monitor mix, and only having the one balance control between that and the computer playback is quite limiting. Moreover, although it doesn’t bother me that the direct monitor mixer lacks fancy stuff like in-built effects, EQ and dynamics, it would be nice if channel linking affected mute, solo and pan controls as well as just the faders. I also found that any channels I muted in the mixer invariably unmuted themselves after a few minutes.
These slightly counter-intuitive routing arrangements mean that the AudioFuse isn’t quite as easy to get to grips with as some of the desktop interfaces I’ve tried. However, once you get a feel for what goes where, you begin to appreciate how much Arturia have packed into this little box, and how much of it is directly controllable from the top panel. With its built-in speaker switching, mono buttons and talkback, it offers as much monitor control as most of us are ever likely to need. The ability to use separate headphone feeds for performer and engineer is important in a lot of home and location recording scenarios, so the AudioFuse also scores heavily over most other desktop interfaces by enabling this.
Arturia make great claims for its sound quality, and although it doesn’t quite match rivals such as the UA Apollo Twin MkII in the specs department, it’s definitely no slouch. The preamps are a cut above those found in bog-standard computer interfaces, with a wider gain range and smooth, non-bunching gain potentiometers, while the quarter-inch inputs bypass the preamp stage, thus ensuring the cleanest possible input from line-level sources. The phono preamp is a really nice touch for anyone who might want to sample from vinyl, or indeed just digitise their record collection, and has a separate grounding point for turntables that need one. The re-amping functionality is likewise both practical and useful.
From a software point of view, I admire Arturia’s brave attempt to spin their use of generic drivers as a positive, but I don’t really agree with it. The third-party Windows driver they’ve licensed offers unstellar low-latency performance without really yielding the benefits that Arturia promise. It still has to be installed, and as far as I can see, it’s still a proprietary driver, just not one developed or controlled by Arturia! They have a stronger case on Mac OS, given that Apple’s Core Audio USB driver is both pretty good and truly generic, but it’s still not as fast as custom drivers from manufacturers like RME and MOTU, and nor does it rival the performance of typical Thunderbolt interfaces. It’s all perfectly usable, but given the company’s heritage as developers of soft synths, I could have wished that low-latency driver performance was further up Arturia’s list of priorities.
That said, the AudioFuse proved perfectly stable and reliable in use, and although the limitations of its direct monitoring are occasionally frustrating, it’s more than capable of taking most simple recording jobs. However, it’s the AudioFuse’s hardware design that will sell it. It’s truly portable, but really does cater for almost all the possible input sources you’re ever likely to encounter, with the option of adding extra preamps over ADAT or S/PDIF should you need to up the channel count. Competition in the world of desktop interfaces is becoming increasingly fierce, but the AudioFuse is perhaps the most versatile such interface yet, with several genuinely unique features. If you need a Swiss Army knife of the audio world, the AudioFuse has more blades than most.
Arturia claim as a positive the fact that the AudioFuse uses generic drivers familiar from many other interfaces, including the PreSonus Studio 192, Antelope Zen Tour, JoeCo Blue Box Recorder and numerous others. On Apple computers, that means the Core Audio driver for class-compliant devices built into Mac OS. Like every other interface I’ve tested that uses this driver, the lowest 32-sample buffer size yields a reported round-trip latency of 5.6ms in Reaper at 44.1kHz sample rate, and a loopback test confirmed that the actual latency is only 10 samples or so greater than the reported value.
ASIO is not a Microsoft standard, so the AudioFuse uses the same third-party OEM driver that all the above mentioned devices employ. This is very popular among manufacturers, but I don’t share their enthusiasm for it: it’s opaque and confusing to set up, and offers fairly lacklustre performance. With this driver, the system latency is determined not only by the buffer size setting but also by a second menu which offers a series of modes from ‘Minimum Latency’ to ‘Extra Safe’. Many of the possible combinations of buffer size and mode result in a warning that you shouldn’t have selected them, which makes you wonder why they are available in the first place.
The lowest possible latency is achieved by selecting Minimum Latency and the smallest 64-sample buffer size. At 44.1kHz, Oblique Audio’s RTL Utility reported the round-trip latency as being 7.3ms. However, if — like mine — your PC won’t run the AudioFuse at that setting, the next ‘approved’ pairing is 128 samples and Low Latency, at which point the round-trip latency leaps to 12.2ms. This isn’t particularly impressive compared with the results that manufacturers such as RME, Focusrite and MOTU are achieving with their custom USB drivers.