DIGIGRID DESKTOP INTERFACES
The promise of audio over Ethernet was ‘I/O anywhere’. Waves’ Soundgrid meant you could take DSP along for the ride, too. Now — with power over Ethernet — DiGiGrid’s new line of portable interfaces allow you to capitalise on that DSP, while wangling I/O into any nook or cranny.
Review: Andrew Bencina
Up until now, the hardware/software DiGiGrid collaboration between Soundtracs (DiGiCo) and Waves has produced mostly staid-looking, black rack devices. Recently, the joint effort has hit a purple patch, and the new portable desktop interfaces are promising to spread Audio over Ethernet (AoE) out of the rack and into infinite studio space.
Included in the 24-bit/96k-compatible desktop range are the DiGiGrid [D], [M], [Q], and [S]. And no, we’re not using some weird editorial device there, the brackets come packaged with the name.
While the [D] looks like your standard four-in, six-out tabletop interface — just chunkier — the remaining three are housed within extremely solid nine centimetre minimalist metal cubes. All four units have rubberised bases but the cubes have an optional fluted base bracket, compatible with the threaded end of any standard microphone stand. This attractive and original option had me thinking outside of the square from the word go.
DiGiGrid’s underlying protocol, Waves’ SoundGrid, is uniquely powerful amongst the AoE crowd in that, in addition to routing audio over Ethernet, it can also share DSP resources [see Are You Being Served? below]. Though not quite as prevalent as Audinate’s Dante, SoundGrid has been adopted by multiple manufacturers. Beyond DiGiCo and Waves, it’s been wrapped into console interfaces and studio hardware from Yamaha, Allen & Heath, Apogee, Burl, Hear Technologies, Roland and Behringer. If you factor in DiGiGrid’s already established range of MADI and analogue I/O-wielding DigiLink interfaces, the number of potentially compatible hardware options is endless. Even Avid, rankled by Waves’ refusal to port plug-ins to AAX DSP, recommend using SoundGrid via MADI to plumb Waves plugs into its flagship S6L live console.
Still, aside from the Hear Back PRO personal monitoring system, the vast majority of these items constitute or are tied to key pieces of studio or live sound infrastructure. As such, they tend to suit a fixed location within a system. The new DiGiGrid devices encourage a greater shift towards genuinely portable, and modular, networked audio.
MAKING THE SWITCH
Installing the modules wasn’t plug and play but it was straight forward. Once installed, the SoundGrid Studio (SGS) application is used to configure the attached devices, channel routings and driver settings. A single device can either be connected directly to your computer’s gigabit Ethernet port or an appropriate network switch can be used to link multiple devices. While probably the least exciting member of this DiGiGrid quartet, the [S] is a Gigabit switch featuring a single upstream port (connected to the host computer or another switch) and four Power over Ethernet (PoE)-enabled downstream ports for SoundGrid devices.
[S] requires its own 48V DC power supply, featuring a smart yet simple twist-lock connector, but is subsequently able to power all of the other compact DiGiGrid modules over cable runs extending to 75m. Without it, or another suitable PoE switch, you’ll be stuck using 12V wall-warts throughout. Waves provides a short list of approved third-party non-PoE switches, all of which were available locally for between $100-160. Reputable PoE solutions seem to range from the top end of those prices upwards. Once everything’s connected an automated wizard can be selected from within SGS to create a generic configuration. While this process was easily navigated on both a Windows10 and OSX system, I noted some audio stability issues with SoundGrid on my older Windows machine.
Testing, with cooperative assistance from Waves support, revealed a presently undefined interaction between SoundGrid Studio and Windows storage drivers on my system; significantly raising system latency and compromising real-time audio. This is not an issue I’ve experienced in previous years of testing for AT, and apparently not one DiGiGrid had encountered before either. Thankfully, it was promptly resolved. It’s a useful reminder that digital audio systems rely on friendly communication between hardware and software layers from multiple providers. Trying before you buy is always worthwhile, if only to get a jump debugging your configuration.
GLEANING THE CUBES
If you’re adding to an existing SoundGrid network, setup will be a doddle. However, for newbies to SGS, most of the action takes place on the ‘Setup > Setting’ page and the ‘Device To Device’ page of the Patch tab. The SGS window is not fully scalable to fullscreen so depending on the number of attached devices you may find yourself scrolling around to find patch points. While this is a bugbear of mine — as is the absence of signal metering within the patcher to aid channel identification — the system’s ability to scale to hundreds of inputs means scrolling may become inevitable, and incorporating hundreds of live signal meters would significantly tax the CPU.
Patching is simply a matter of checking the appropriate boxes but it’s worth noting that while you can send one source to multiple destinations the patcher is not capable of bussing multiple sources to a single output. Also, unlike the rackmount DiGiGrid interfaces, at present none of the new devices support any form of remote control, which seems counterintuitive given the promise of networking remote I/O over the SoundGrid protocol. Seeing as the actual devices don’t sport endless rotary encoders or digitally-controlled preamps, this will be a lasting omission. DiGiGrid told us the analogue front end was a design decision to realise best audio performance, and that introducing digitally-controlled mic amps would take away some of the system’s simplicity. It’s a perfectly adequate explanation for home-oriented, single unit users. However, given DiGiCo’s high-end consoles and the rest of the DiGiGrid rack range relies on digitally-controlled preamps, it appears they’re audio quality is sufficient. While it seems DiGiGrid is positioning these boxes as entry points into the SoundGrid/DiGiGrid family, not including digitally-controlled preamps only shortchanges the system’s promise to expand via remote I/O points.
Without a SoundGrid server connected, the desktop devices do not support low latency pre DAW-created monitor mixes. The lack of onboard DSP makes them limited when used as standalone interfaces; particularly when you consider their retail pricing in a flooded market.
As some consolation SoundGrid does allow you to nest third-party Core Audio and ASIO drivers within their network, via SG Connect. I would generally insist that the differing devices share a Word Clock connection — with ASIO I found this to be prerequisite — but I was able to add some Lynx Aurora Thunderbolt-connected converters to a [D]-based SoundGrid system, when testing on an iMac. While the latency performance of the Aurora was significantly compromised I can imagine situations where a similar combination of devices could be invaluable (for example, sending and returning reverb chamber channels to unusual spaces using a single Ethernet cable).
D IS FOR DESKTOP
While [M] and [Q] are more limited, I’d describe [D] as the entry point of the DiGiGrid line. With two phantom-powered microphone inputs on XLR, two TRS line/instrument DI inputs, four TRS outputs and a headphone amplifier (with its own ‘big-knob’ volume control and metering), the [D] provides a practical solution to all of your basic recording requirements. A low cut and polarity inversion switch on the mic channels round out the standard features. As a tabletop device I’d love to have seen further connection options on the front face of the unit. The Master channel LED meters are configured pre-fader, differing from both the [M], [Q], and the [D]’s own headphone metering. If possible, this is one of a few options I’d like to see software switchable. With the master knob approximately positioned at 12 o’clock (no markings meant I was estimating somewhat) the output of the unit was equivalent to the loudest ‘+8’ setting on my RME converters. Not far beyond this the outputs began to distort, meaning a significant range of the knob is wasted.
The [M] is intended for musicians and is basically a baby two I/O version of the [D], with a single DI, mic channel and headphone amp. Notwithstanding the clear design focus, I’d have found a dual channel unit with switchable mic/DI channels to be more flexible and a great little stereo all-rounder.
Finally, the [Q] is a multi-input headphone amplifier (presumably with cue mixes in mind) with both ¼-inch and mini-jack outputs on the front panel. It’s either/or here, with the full-size socket disconnecting the smaller. On the rear are three of the four input connections (AES/EBU on XLR, analogue RCAs and the Ethernet port) with Bluetooth rounding out the source options. Input selection is achieved by pressing the Master knob with the otherwise-hidden selection labels backlit by coloured LEDs. Annoyingly, until a Bluetooth device is paired two LEDs flash incessantly, even when monitoring other sources. I think [Q] is a fantastic audio utility for any studio, irrespective of the level of SoundGrid investment, yet, as with [M], the inclusion of additional options — a pair of balanced outputs, perhaps — would only have expanded such utility.
While I might be a little frustrated by certain design choices I cannot seriously fault these boxes when it comes to audio performance. SoundGrid’s roundtrip latency is not quite up there with my records of the best performing interface/driver combinations but it falls well and truly within the next pack. Surprisingly, I even found that these units slightly outperformed the rackmount DiGiGrid IOS.
Using Radial splitters, I recorded a quick demo track via the [D], [M] and a range of DIs and preamps from Radial, Universal Audio, Phoenix Audio and Quad Eight. All gain levels were calibrated using test-tones and accurate metering so barely any post recording level matching was required. The [D] and [M] recordings are as close to identical as you could hope for and comparisons with the other channels was relatively favourable. This type of test always makes me question how I’ve allocated my studio budget and while you can clearly differentiate between the unmixed recordings, with the DiGiGrid DI channels in particular having a more emphatic top end, I wouldn’t hesitate to use either interface to access some new corners of the property.
All headphone outputs are amply powered and when compared with the other monitoring options on hand, preferences were more often determined by the combination of source material and headphones. In isolation these units may lack the flexibility and software features of comparably priced audio interfaces but within a broader SoundGrid network they’ll stand level with most other devices you’re likely to connect.
For a few logistical reasons, my introduction to SoundGrid has ended up being a little topsy turvy. These new modular DiGiGrid units expand and complement an existing audio network platform founded upon a strong reputation in DSP. If you’re looking for an AoE system that also distributes DSP, there’s no other game in town. However, while the DiGiGrid desktop devices are simple to use, sound great, and offer standalone functionalities that will be useful to many, they are perhaps more of a sweet reward for those already fully immersed in SoundGrid than a tasty lure for the uninitiated. One warning though, like one of Zumbo’s macarons, if you take the bait it may be difficult to stop at just one.
ARE YOU BEING SERVED?
Unlike other AoE platforms, Waves SoundGrid has been primarily sold on its capacity to offload plug-in processing and reduce latency when mixing live input signals. Running a modified Linux OS, referred to in-house as SoundGrid Processing, these devices are basically rackmount computers connected to your live console or studio machine via the SoundGrid network. Through either the MultiRack (console) or StudioRack (DAW) plug-in shells, it’s possible to run the extensive suite of Waves plug-ins and third-party plug-ins from developers including Plugin Alliance, Flux::, and, in the near future, Sonnox. The Waves Public API (WPAPI) format allows for continued growth in this selection. For those times when you need to escape the studio, StudioRack freely switches between native and SoundGrid processing while the eMotion ST mixer — run within the SoundGrid Studio application — opens up further possibilities for low-latency monitoring, input processing and even live mixing.
Waves SoundGrid branded servers have been around for a while now with three current flavours: the Impact (Intel Core i3-4150K, 4GB RAM); Server One (Intel Core i5-4590K, 8GB RAM); and Extreme (Intel Core i7-4790K, 8GB RAM) models. To give you a sense of the performance range, in a studio application running at 96k, the Impact will happily accommodate 138 instances of Waves’ SSL E-Channel (mono) strip while the Extreme can handle 512. At present 192k operation is not supported by the servers but it is scheduled for a future firmware update.
DiGiGrid models have added to the server range, combining SoundGrid processing with either audio I/O (DiGiGrid IOS — Intel i3) or Pro Tools DigiLink interfaces (DigiLink DLS — Intel i5), and their performance matches that of the equivalent SoundGrid models.
Unfortunately, I’ve only just begun to explore the possibilities afforded by a DiGiGrid IOS so a deeper discussion of SoundGrid processing will have to wait. However, in the context of this review, it is worth pointing out that any modular DiGiGrid system relies on the presence of a SoundGrid DSP server to fully maximise the potential of the platform. While the scalability of SoundGrid allows for this to be added at any time, the absence of a server reduces the low-latency monitor mixing options of the system considerably — especially when compared with the DSP mixing options included with current interfaces from companies like MOTU. DiGiGrid’s competition may not support any form of plug-in hosting (UA Apollo interfaces being the obvious exception), but basic mixer functions are universally available with all interfaces.