Review: RME Fireface UC

If I had a dollar for every time some goon on a pro-audio forum issued a dictum committing USB 2.0 to the interface scrapheap I’d be a wealthy man. It’s amazing how often this ignorant declaration occurs, especially when you consider USB is capable of transfer rates of 480Mbps while standard Firewire options hit the wall at 400Mbps.

Unfortunately, in most cases the implementation of the universal serial bus architecture cripples the potential performance of the interface and restricts sustained throughput – critical for multichannel audio. Even Mac and Windows differ in their software utilisation of this transfer format, which – by its very name – is supposed to be universal. So why then would a company like RME, boasting a range of the most impressive Firewire products on the market, choose to invest significant time and energy developing a USB 2.0 device? For the answer we’ll have to delve deep into the fiery core.

WE SHALL OVERCOME

When the IEEE 1394 (Firewire) standard was first unleashed it was commonplace to find that your new external drive or video capture device didn’t work efficiently with the chipset inside your computer (if at all). We all took this in our stride of course – what else could we do? It simply became yet another example of an incident where compatibility issues occurred as each hardware upgrade rolled around. Over the years things settled down and some quasi standards were adopted by manufacturers in something of a mutually beneficial gentleman’s agreement.

There were, however, no guarantees, and a couple of years ago, in an ill-advised cost cutting exercise, Apple decided to replace all the Texas Instruments chipsets in its new models with a cheaper option. For many peripheral developers, including RME, the results were disastrous. The Fireface 400 and 800 simply didn’t work with this generation of machines without the use of an additional IEEE 1394 hub or repeater. I’m not sure if Matthias Carstens and his compatriots at RME were hanging out of the windows but I think it’s fair to assume they were mad as hell and decided not to take it anymore.

Their response is an innovative one, with significant implications for all involved in the development of external peripherals. RME found that, unlike Firewire, USB 2.0 didn’t require a third-party chip and could therefore be implemented using a comparatively affordable and programmable FPGA (RME’s Universal Core). In simple terms, the designers found a way to implement the USB 2.0 interface using a completely updatable software alternative to static and easily outdated chipsets. By taking charge, the developers at RME have opened the door to a world where firmware upgrades may eventually replace the wasteful material disposability to which we’ve all grown accustomed. With the Fireface UC this approach afforded some additional benefits.

As mentioned above, USB 2.0 is not always implemented in a way that enhances audio streaming, and an efficient device under Windows will not necessarily display the same performance under Mac OS. The programmability of the UC’s core affords RME the freedom to deal with both these issues. After extensive testing it was found that the most effective data transfer results under Windows were achieved using the interrupt transfer method while the Mac preferred the more standard isochronous mode. Interestingly, neither operating system supports the mode RME believes to be the optimal implementation for audio. As room was available onboard the FPGA there was simply no good reason not to employ both. The Fireface UC therefore contains discrete firmware versions for Mac and PC alongside backup versions. A simple button press prior to connection configures the UC for best possible performance on either platform. If future hardware developments cause instability, or OS updates allow for increased performance, changes to the core will be possible via downloadable software updates. The first of these has already been released in response to Windows 7 and Snow Leopard.

WHAT’S GOING ON?

For such a small box, the Fireface UC is packed with I/O options. The front panel houses the first four of eight analogue inputs, the selection and control knob and a low-impedance unbalanced stereo line output (channels 7/8), which doubles as a high-performance headphone amp. When listening to mastered or ‘loud’ recordings the headphone output delivers loads of volume to even low-impedance models. The representation is accurate with a solid bottom end and it was a joy to have an alternative to my multichannel headphone amp. I did find, however, that onstage in close proximity to guitar amps and a drum kit it was struggling to boost more dynamic backing material into the consistently audible range.

Inputs 1 and 2 are handled by Neutrik combo sockets allowing mic and line operation using XLR and TRS connectors. The mic pres are equipped with a maximum of 65dB of gain, adjustable in 1dB steps from both software and the rotary encoder. Phantom power is also individually switchable for each channel (from within software only). Inputs 3 and 4 offer a choice between balanced line operation and a Hi-Z unbalanced instrument input via ¼-inch jacks. Settings are again configurable from both the front panel control and the driver window, with 18dB of gain available to the instrument DI in 0.5dB increments. Both the mic pres and DI inputs provide open and detailed alternatives rivalling any of the outboard in my racks. While the pres lean towards reference style neutrality, the DI’s really shined on active bass and electric piano. Levels for all four inputs are digitally controlled but the changes are realised discretely in the analogue domain. Channel settings and routings can be stored within the unit to allow for stand-alone operation as a mic pre or even as an eight-channel mixer via the excellent TotalMix DSP software.

The back of the Fireface is chock-a-block with the remaining I/O. Ten ¼-inch TRS jacks accommodate the remaining four balanced line-level inputs and channels one to six of balanced output. Additional digital I/O is provided via coaxial S/PDIF and ADAT (eight channels @ 48k, four @ 96k, two @ 192k). Optical S/PDIF is also available via the ADAT connections. While all of these rear channels have their own corresponding mixer channels within TotalMix, only the six analogue outputs can be controlled from the front panel rotary encoder. The rear panel line up is rounded out by wordclock I/O and two MIDI ports, providing an underrated and exceedingly handy 32 channels of MIDI I/O.