Computers have made music production fundamentally easy. One no longer has to spend hours miking instruments for recording, for example, and simply needs only to buy the sampled sounds he or she desires. However, with this ease of use comes some unexpected complications which may not be apparent at the onset, especially when dealing with something as seemingly innocuous as a sampled drum kit.
In this tutorial we will look at how to avoid some common pitfalls inherent to virtual drum programming and how to properly 'mic' a virtual drum kit for it to be of ultimate usability in your projects. I will be using Logic and Ultrabeat, but any DAW and drum sampler will work.
The first thing I did was to instantiate a multiple output instance of Ultrabeat, add some natural sounding samples, import a jazz loop, and separate the voices onto individual tracks. The next thing I did was go into the mixer and create individual channels for each of the drum voices.
I set the kick and snare to 'no output' and sent each of them to two separate auxiliary channels. I then set up a summing channel for the kick and one for the snare and set up two additional busses on the output of each voice.
As it stands, we have four major elements in conflict with each other, level, pan position, frequency and space. There are no soft or loud drums, everything is homogeneously loud and difficult to listen to. Each voice is basically centered with no pan separation. The lows and especially low-mids of each voice run into each other. Each of the individual voices were recorded in a different space and therefore do not have a cohesive sonic spatial signature (ambiance).
To address these considerations, we are going to have to play with the drum levels, adopt a panning strategy, do a lot of high pass filtering, and dry out the samples as much as possible.
The Kick & Snare
You may be wondering why I have sent the kick to two auxiliary tracks. The answer to that question is that in a real drum recording setup, the kick (and snare) would be double miked. One microphone is set inside the drum to pick up the attack of the beater against the head (the highs) and another is set outside to pick up the bass.
To recreate this, I have inserted EQs on each channel and low cut the internal channel while high cutting the external channel. I put some overdrive on the internal channel to accentuate the highs and parallel compressed and gated the lows to give it punch.
To emulate microphone preamp saturation, I used PSP's mix saturator on the summing channel and found a subtle setting. I then used an enveloper to fix the in and out channels to a similar amplitude envelope (and to cut out much of the spatial reflections recorded on the sample) and did some final EQing. When using this trick it is always imperative to do some type of dynamic and/or frequency processing on the summing channel to marry the two separate inputs.
In a recording situation, two microphones are used on the snare as well; one on the top to capture the low-mid attack, and another on the bottom to capture the crisp strainer sound. I set up the snare in much the same way as the kick. I band passed the top channel to cut out any unnecessary lows and to dampen the highs. I then parallel compressed it to give it a decent amount of punch.
For the bottom channel, I cut most of the lows and added some overdrive to make the strainer sizzle. I added the noise gate next to cut out much of the resonance of the sample as that is largely a characteristic of the space it was recorded in. I then enveloped, saturated and EQed the summed sound.
The original kick and snare.
The miked kick and snare.
For the low tom, high tom, hi-hat, ride and crash I immediately filtered the lows and low-mids in order to clean up the competing frequencies. I then tried using envelopers on each track to cut of the resonating tails, but found a noise gate was really the necessary tool for the job in all instances save the hi-hat. For sake of moving on, I did not bother trying to emulate an analogue recording chain with saturators or the like. More info on that can be found here.
I then went through and did a quick level mix and some slight panning. The level mix was pretty straightforward and done relative to the kick drum. The panning is of some interest in that I adopted a panning strategy that emulates how a person in an audience would experience a real drum performance. I kept the kick mono, panned the snare slightly to the right and placed the hat and ride even further outside. The high tom is just left of the kick, with the low tom and crash further to the left.
None of the panning is extreme, with most voices within the 15 degree range. This keeps the drum kit together and realistically wide while giving each drum its own distinct space.
The processed kit.
Overheads And Room
Although we've succeeded in getting rid of the conflicting lows and ambiances, we have been left with some fairly dry and tailless sounds that stop abruptly and sound artificial. To help I've dedicated the first of the two remaining auxiliary tracks (the blue channels from earlier) to the overhead mic. To do this, I have inserted Logic's convolution reverb plug-in, set it to a medium sized space and tightened up the decay time and spread to emulate a closely placed ambient mic.
In a true drum recording set up, two overheads are generally used to achieve a consistent sound from both the left and right sides of the drum kit. We only need the one stereo channel here since the amount of signal sent to the overhead will be based on the bus send level of the individual voices and not the left/right placement of the microphone relative to the sound source.
For the room mic, I have duplicated the settings of the overhead (it is pretty important that our microphones are placed in the same room!) and opened up the decay and spread to capture the late reflections of the modeled room. I've also increased the pre-delay to emulate the more distant mic placement. Once again, only one stereo channel is needed as the levels are dependent on the bus level and not left/right placement.
It is important to send all the drums to each reverb, since both microphones would pick up at least minimal levels of each voice in a recording environment. When doing so, keep in mind the front to back relationship of each drum within the kit to achieve a more realistic effect (e.g. the kick is further from the overhead and room mics than the ride and should have less level sent to the reverbs, etc.).
As a final touch, I rolled off some of the highs from both reverbs to give a greater sense of depth and sound absorption. The room mic was cut a bit more severely as that microphone would be placed furthest from the drum kit.
The final product.