Can this unique combination of software and hardware really recreate the sound of some of the most revered microphones in the world? We put the VMS up against the originals to find out!
Most hardware components of the traditional recording studio are now available in digitally modelled or ‘virtual’ plug-in software versions. Pretty much every classic compressor, EQ, analogue tape recorder or effects unit in the history of recorded music is now seemingly available to use in our DAWs with the minimum of fuss. Even a studio’s live room can be recreated to some extent, thanks to increasingly advanced room-modelling software and convolution-based reverb. Microphones, however, have been among the toughest nuts to crack. It has generally been considered extremely difficult, if not impossible, to truly model how different microphones respond to any given source, from any angle, in any space.
It’s perhaps not surprising, then, that the release of the Slate Digital Virtual Microphone System — or VMS for short — has created something of a stir in the industry. Slate Digital’s CEO Steven Slate has helped generate some of the hype, of course, and his highly personalised, ultra-confident marketing style has made him something of a polarising figure in the world of pro-audio. Look beyond the hype, however, and you’ll see that many people, including me, are regular and enthusiastic users of his ever-expanding range of mix processing tools. Slate are also a manufacturer who aspire to innovate: they were one of the first plug-in companies to embrace an affordable pay-monthly system for their entire suite of plug-ins, and their Raven touchscreen DAW controllers offer a new way of working with our computer-based setups.
The Slate Digital Virtual Microphone System claims to precisely model many of the classic large-diaphragm valve capacitor microphones, and some well-know preamps, of the last 70 years. Is the VMS a new frontier for pro audio, then? Or does it represent a step too far for today’s digital modelling technology? When you consider the fact that the whole system comes in at less than £$1000, and that some of the microphones in question will set you back more than £$10,000 each on the second-hand market, it is certainly not only an intriguing proposition but a bold one.
The Slate VMS comes as a package of both hardware and software. Included is a large-diaphragm capacitor microphone, a stand-alone hardware preamp, and the all-important modelling software, which is downloaded from the Slate Digital web site. The software functions as part of their Mix Rack package and allows you to audition the different microphone and preamp models, both in real time and, significantly, post recording.
The microphone and the preamp are designed to provide a relatively characterless, transparent signal, which the software can then process to add the character of the microphones. This included hardware is key as, unlike a purely software-based system — such as the old Antares mic modelling plug-in — the hardware ensures a known neutral source characteristic with which the software can react.
With crude but obvious pseudonyms for the mics in question, the software is available initially with three microphone options: the FG-47 (Neumann U47), the FG-251 (Telefunken ELA M251) and the FG-800 (Sony C800g). Slate have since released an additional expansion pack, which adds a number of additional tube-mic options including the Neumann U67 and the AKG C12. Once you have auditioned and selected your mic of choice, you have the option of increasing the ‘intensity’ of a particular model’s character: the software slider ranges from 100 to 150 percent, where 100 percent is the closest emulation, and higher values increase the level of modelled harmonics and valve saturation.
Also included within the software are two virtual preamp options, emulating the Neve 1073 and the Telefunken V76. As well as using these along with the microphone models, you can also record direct into the line input on the Slate preamp, to use these preamp models for a keyboard or direct bass recording, for example.
If the VMS does, or even comes pretty close to doing what it claims to do, then the system could represent excellent value for money, and would bring the sonic characteristics of many of the most coveted microphones in recording history to a whole new group of users. Perhaps most intriguingly, it also has the potential to remove a layer of decision-making in the recording process and potentially add a brand-new one to the post-production or mixing stage. Deciding which microphone to use in a given situation is a big part of a recording session, and the ability to audition microphones after the event has many implications — both good and bad.
Could this kind of technology be another nail in the coffin for the conventional recording studio? Most of us take advantage of numerous kinds of digital modelling software these days, and whilst many users acknowledge that they perhaps still don’t sound quite the same as the real thing, they’re content that it’s close enough, or sounds good in its own right, and happily embrace the convenience and affordability that software emulations give them.
Rather than just trying it out on a few sessions at my studio, the team at SOS decided that this review would be a great excuse to let a group of us out into the wider world and pitch the VMS against some of the actual microphones that it claims to model. With help from the fantastic team at FX Rentals in London, we were able to get access to four of the classic mics that the VMS attempts to model: the Neumann U47 and U67, the AKG C12 and the Sony C800g. I agreed to host the session at my Cambridge-based Half-ton Studios, and in addition to my own ears, those of SOS editors Hugh Robjohns, Sam Inglis and Chris Korff were present, as well as regular SOS author Mike Senior.
While it’s difficult to get too scientific about these things, we came up with a simple system by which we felt we could consistently audition the respective microphones against their emulations. As different mics could be changed in the software on the Slate VMS, we knew we could keep the Slate microphone in a fixed position in the live room. We would then position each of the vintage mics, in turn, with the capsules as close as practically possible to the Slate’s, and record a short vocal performance. We decided we needed to test both a male and female vocal and, while soundchecking the mic with some spoken word, we decided that a further voiceover-style recording would be a worthwhile addition. We decided to focus the review very much on the microphone side of things, but since I have a Neve 1073 preamp at my studio, we used that for all of the ‘real’ microphone recordings, and applied the Slate 1073 emulation to all the modelled mic recordings.
Before beginning the session proper, we all had a good look at and feel of what’s included in the hardware side of the system. The microphone itself is called the ML-1; it has a good feeling of weight and quality to it, and whilst the all-black styling won’t appeal to everyone, it does give it a neutral feel that is not trying to look like any other well-known microphone. Without knowing any details of the ML-1’s manufacturing, it looks to me very much like a generic large-diaphragm mic that you would perhaps expect to cost around £300 to £400$400 to $500. The included shockmount feels solid and was easy to use, and the fact it comes with a decent quality hard case is a plus.
The included preamp, the VMS-One, was OK in terms of build quality, and I found it pleasant enough to work with during the session. Designed to sit on a desktop, the VMS-One has all the basic controls you might expect from a stand-alone preamp: phantom power, input pad, polarity switching, mic/line selection and a large gain control knob.
The system setup is designed so that you route the ML-1 microphone into the VMS-One preamp in a conventional manner. Your signal would then go in to a line-level input on your mixer or interface and then on to your DAW of choice. Reading, and watching, the promotional material for the system, it seems very much intended that you perform ‘through’ the software and use the emulations in real time. This concerned me ever so slightly, as while the software itself doesn’t produce any latency, you are at the mercy of the limitations of your particular system. In my experience, even small amounts of audible delay can be very distracting to performers, and while there are many setups these days that are capable of operating at very low latency, I suspect there will be some users who will struggle to be able to use the VMS in real time without issues.
For Mac-based recording systems, Slate Digital recommend using the VMS with a low-latency Thunderbolt interface running at 96kHz, with a buffer size around 64 samples, to achieve a latency lower than 2ms. For PC, they recommend a low-latency USB interface (“RME make the best, in our opinion”), or PCIe-based options such as those made by Lynx. In my own studio, I have a Lynx Aurora interface with a PC running Pro Tools 12, at a fairly low buffer size of 256 samples. With conventional routing, I could still hear latency ay work when I walked through to the live room and spoke into the mic with the headphones on. You could of course just monitor through the mic and then play with the software afterwards, but once you have found a model that suits your voice, this might become frustrating.
With a lot to get through on the day, we decided for the sake of our exercise to not monitor with the mic emulations in real time. We were, of course, fortunate enough to have the real microphones set up, to both monitor with, and to feed to the respective singers’ headphones.
The first classic mic out of its flight case was the legendary Neumann U47. With the microphone safely positioned next to the Slate ML-1, the first vocalist up was Jules Harding, who we had somehow convinced to be our voiceover artist. I always find spoken word to be an excellent illustration of a microphone’s basic character — helped greatly of course by the fact that it stands alone, without any music to mask the subtle nuances.
I’ve used a U47 a few times before, and in its natural state it doesn’t suit every voice, by any means. When it does, though, its rich bottom end, and general all-round smoothness, can be a dream to work with. Jules had a naturally deep and resonant voice anyhow, so it almost felt a bit too ‘full’ until we got him to back off the mic slightly to reduce some of the U47’s generous proximity effect.
With a short 30-second piece captured — and with a gaggle of ever-so-slightly sceptical engineers hovering in the control room — we were able to make our first comparison with the Slate Digital VMS. I left the ‘intensity’ control flat for the moment and switched between the modelled mic and the real one. I would describe the first impressions as pleasantly surprised, rather than jaw-dropping, but it’s important to note that no two U47s (or any other mics of that age) are ever going to sound the same. The term ‘it’s in the ball-park’ was certainly appropriate, though, and the Virtual Microphone System was seemingly doing a pretty good job of capturing the right ‘flavour’. It’s notable that the emulated version was a little brighter than our U47, and this would prove to be a recurring theme.
Cranking up the intensity made things more ‘syrupy’, however, and it also helped fill out the bottom end to more closely match the real thing. It was quite obvious even at this early stage that there wasn’t a huge amount of difference between them, and that the VMS deserved to be taken seriously. Briefly flicking between the ‘flat’ mic and the emulations, it also seemed clear that the software was able to produce dramatically different sounds to the unprocessed ML-1 microphone, with barely any noticeable artifacts.
Moving onto our singers, our male example was SOS’s very own Sam Inglis. As Sam often records his voice in combination with him playing acoustic guitar, we decided that this could be a useful test of the system’s ability to model the off-axis sound of the microphones, as benign off-axis sound is a much-appreciated feature of many high-end microphones. Listening to the real thing first, the U47 wasn’t necessarily a natural choice for Sam’s voice, as it sounded almost too rich and full. It did a great job of softening any harshness and sibilance issues, however, and I was keen to see if the VMS was able to do a similar job. It did, up to a point, and we were impressed with just how close it was coming to sounding like our particular vintage model. Again, playing with the intensity control in the software enabled us to bring the bottom end more in line with the real thing.
However, in the instrumental sections of Sam’s song where only the off-axis guitar sound was captured, there seemed much more of a difference in tone between the two mics. This does not mean the off-axis sound was bad, just that it was further from the real thing than the on-axis sound.
For our female voice, Grace Kuhl had very kindly agreed to be our test subject for the afternoon and, importantly, she would be singing over a full-sounding pre-recorded backing track. Why is this important? Well, if the VMS can come reasonably close to accurately modelling these mics, then maybe any differences that might remain would be indistinguishable in the context of a full track.
Like many female singers, Grace has a voice that suits a bit of ‘airiness’, and I didn’t think our real U47 sounded all that great on her voice, producing a tone that was a bit dull. This could easily be solved with a little EQ, of course, but in this context I preferred the slight extra brightness of the VMS version. Listening on a pair of headphones, I could detect a few small issues with the lower mid-range balance on the VMS version, and when I added some high-frequency shelving EQ — as I might in a mix — the real U47 opened up in a beautiful way that I felt left the modelled version slightly behind. We are talking small differences here, though, and the results with the Slate model were still impressive.