Do your virtual acoustic drum parts never seem to work in the context of a mix? Here's how to fashion more natural and compelling results...
Software instruments that can emulate live drumming are 10 a penny these days, and many come with libraries of MIDI parts that make programming a convincing performance a breeze. Most such instruments also provide multi-channel outputs, a bit like the multiple mic channels of a real drum kit recording, but in practice you'll rarely get the best results at mixdown if you try to treat a virtual drummer just like a real one.
The first thing to realise is that, with any software drummer, the mixing process should always start within the instrument interface itself — if any individual kit component doesn't sound quite the way you want it to, then reach for the software's own instrument controls before you start getting any other mix processing involved. Your snare, for example, will sound a lot more natural if you increase its sustain using the software's built-in envelope controls, rather than trying to compress the snot out of it with down-stream plug-ins. Similarly, adjusting the design or pitch of a virtual drum makes a tonal difference that's impossible to replicate with EQ — and if you tune the drum to the key of the track it'll often blend better straight away, without the need for artificial reverb.
Beyond changing the sounds themselves, you'll usually get some flexibility to rebalance those sounds within each of the software instrument's audio output channels. In the simplest case, there'll be a selection of output channels for individual kit components (kick, snare, rack tom, floor tom, hi-hat, ride, crash and so forth), each providing a 'complete' dry mixed sound for that instrument, and those will be accompanied by some kind of 'ambience' or 'room' track you can use to give the mix a roomier sound if you like. In this kind of situation, your highest priority should be to decide how loud each kit component is within the simulated room, because the louder a given kit component is in the room sound, the roomier it'll seem in the mix relative to the other components. However, it's not uncommon to find that the software instrument's simulated room sound is more appealing on some kit components than others within your specific mix context, in which case you can make your life a lot easier by muting certain kit components in the software instrument's room channels, and then adding effects to the individual component channels within your DAW instead.
The more sophisticated virtual drummer instruments, however, don't output each kit component 'ready mixed', but rather emulate the signals you'd expect to get directly from a miked-up acoustic kit: in other words, multiple virtual close-mic signals per kit piece, together with overhead mics and room mics that pick up the whole kit. While this does afford you much more scope for sonic adjustment, for many readers that often translates in practice as 'enough rope to hang yourself with' — especially for those lacking enough experience mixing real acoustic kit recordings.
My first shortcut to decent results in this case is to ignore all the close mics at the outset, and just listen to the virtual overheads. Ideally, these should sound as close as possible to the kit sound you have in mind. I can't tell you how often I fade up the overheads of a virtual drum instrument and it sounds nothing like you'd expect a natural kit to sound — perhaps the snare's very low in level, or the hi-hat's overwhelmingly loud, or one of the toms is entirely missing. Unless your overheads sound fairly natural, you'll be facing an uphill struggle trying to get the full drum kit mix to sound any good. If the snare is too low in level, it'll lack width and size; if the hi-hat's too loud, you'll probably balance the overheads too low, so everything else won't really glue together; and if you're only hearing the tom through its close mic, it'll always sound a bit 'stuck on' rather than blending with the kit as a whole. There is one common exception to this rule of thumb, though: it's very common for kick drums to be presented unnaturally dry in modern rock mixes, so you may want to remove that drum entirely from the overheads mix. (That said, legendary rock mixer Andy Wallace once told me that he actually likes the sound of kick-drum ambience, so this is something you might want to consider while listening to some of your own favourite productions.)
My second piece of advice is to switch off any simulated 'spill' on the virtual close mics. You know, where you can decide to hear a bit of the cymbals and toms on the snare close mic, for instance. The reasoning behind this facility seems sensible on the face of it: real kits have spill on their close mics, so why not emulate that? But almost invariably I find that this simulated spill simply doesn't respond in the same way that real acoustic spill does. You see, with a real drum recording you can make spill work in your favour, manipulating the phase relationships between the different mic signals so that the spill supplements and enhances the sound of each kit component. But whenever I try to do the same thing with the virtual spill signals in software instruments, the result always ends up sounding a bit hollow and phasey, and generally less satisfying than simply switching off all that spill and working with the artificially spill-free close-mic signals instead!
Consider switching off any compression effects that are built into the instrument patch, because they can be a bit overblown — presumably to impress prospective customers...
While you're in the software instrument's internal mixer, I'd also consider switching off any compression effects that are built into the instrument patch, because they can be a bit overblown — presumably to impress prospective customers who are likely surfing presets in isolation! The problem is that many styles that rely on live drums also benefit from assertive master-bus compression for cohesion and a sense of excitement, but if you've already squashed the drums flat within your software instrument, you won't have enough drum peaks left for the master-bus compressor to react to. I've also discovered a tendency for people to over-compress software instrument close mics in their DAW's mixer, forgetting that it's both easier and more effective to adjust the dynamics of the performance without any compression side-effects, simply by editing the Velocity values in the MIDI trigger data.
Another internal feature I'm normally inclined to switch off in software drum instruments is any built-in reverb processing. Again, it's not that reverb isn't frequently useful for mixing drum kits; it's just that you'll get better cohesion in your mix if you use send effects within your DAW which can be applied to other instruments in addition to the drum kit. That way, lots of instruments will share some acoustic commonality with the drums, and you'll get more of a sense that everything was performed at once in the same room.
If you've done your work with the software instrument properly, mixing the virtual drum kit channels really shouldn't be too difficult. Indeed, if it is difficult, that's a good indication that you need to tweak the instrument's internal parameters further! Because spill isn't really a concern, you obviate the need for most of the gating and filtering you might require for a real acoustic recording. Similarly, you can avoid dynamics troubleshooting by refining your MIDI trigger data, and there should be much less need for in-depth reverb treatments once you've balanced your overhead and room signals sensibly — just a touch of global reverb to bind the kit together with the rest of the band, perhaps. As such, it should probably go without saying that mix templates designed for live drums aren't likely to be much use for virtual drum instruments.
One thing that is always worth confirming, though, is that the phase-relationships between the different close-mic signals and the overhead/room mics are optimised. You'd hope that library developers would check for this kind of thing, but I have come across situations where reversing the polarity of a mic or two has made a useful improvement to the combined sound. And don't be afraid to cut away some of the drum frequencies once you start trying to incorporate the kit sound into your mix. Remember that instrument presets are often designed to impress in isolation, which means they're unlikely to leave enough mix real-estate for the rest of your band.
For me, the bulk of the work when mixing virtual instruments usually involves trying to add two things: cohesion and interaction. The first of these is the sense that the ensemble belongs together and that the instruments don't sound like a bunch of disconnected overdubs. As I've already mentioned, a touch of global reverb can help here, and this doesn't need to have a long tail to work. I usually keep that kind of reverb quite short, in fact, so that it's not really detectable at all — until you switch it off and the mix doesn't blend any more! Subtle master-bus saturation or tape emulation can help here too (particularly if that adds a subliminal background noise layer to the whole production), but do keep a careful ear open for unwanted side-effects, such as any loss of low-end solidity on the kick, or any softening of snare drum transients.
Compression can also help improve the sense of cohesion if applied to the drum subgroup/bus, but if you apply it to your master bus, then it can introduce a sense of interaction too; when the kick-drum hits, the levels of the other instruments in the production are momentarily ducked, for instance. But the problem with master-bus compression is that all it sees is the signal level, not the musical value of each part. So don't forget to take advantage of fader automation — because it uses a human gain control element (ie. you), it can introduce interaction between the drums and the other parts in a more musical way, fading up interesting hi-hat fills, for instance, or emphasising the kick drums and cymbal hits at the starts of important sections.
I hope this article has given you some pointers to improve your mixing of software drums. Although they sometimes don't respond in exactly the same way as real drum recordings, they really can sound fabulous once you learn to manage their eccentricities.
Back in the '80s, people often wired up a Yamaha NS10 woofer to an XLR and used it to pick up the low-end of the kick drum, and Yamaha later used a similar principle to develop a product called the SubKick. Because the speaker cone is so heavy, it hardly responds at all to high frequencies, while the low frequencies tend to ring on longer than you'd expect on account of the resonance characteristics of the driver itself. And, of course, if you mix the resultant low-end booming with your other kick-drum mics, it can make the low end sound huge.
I seldom use this kind of signal much in the mix. This is partly because it leaves no headroom in the lower octaves of the spectrum for a full-sounding bass (something that's usually more important for underpinning the harmonies in mainstream song-led productions). But in addition, if you add that kind of low end to up–tempo tracks, you'll just make the kick-drum patterns feel sluggish and lacking in punch. In my experience, more traditional kick-drum mics will almost always provide more useful and controlled low-end power.