Electronic Euphoria

Feb 1, 1999 12:00 AM, By Jeff Casey

           

Building a powerhouse mix with electronic instruments.

What Goes Where?

When working entirely with sequenced MIDI parts, you have the option of mixing tracks virtually without ever recording them to a multitrack. The benefits of mixing directly from the sound modules are obvious. For one thing, your signal path is shorter, so your chances of collecting sonic garbage drop substantially. In addition, sticking with MIDI tracks leaves more audio tracks available. But most important, it means you’re not committed to anything—you can edit and automate synth sounds and sequences in ways that would be difficult to do with audio tracks, and you can do it at the last minute if necessary. (If you’re performing virtual mixing with a digital audio workstation, make sure your system has the ability to mix live inputs and that you have enough I/O available.)

FIG. 2: In this example, the kick drum and one drum loop are right in the center of the mix, as in a live band. Because the synth bass has a heavy low-frequency content, I have panned it in the center, which also helps lock it in with the kick. The lead vocal and guitar solo are in the center and set hot. They don’t occur at the same time, and they have different spectral content than the kick drum, drum loop, and bass, so each instrument will be distinct.

However, many professional engineers like to print MIDI sequences to multitrack audio media (especially analog two-inch tape). Sometimes it’s unrealistic for the personal-studio owner to go this route, but it has several advantages. In a small studio with limited resources, for example, recording to audio multitrack allows you to apply outboard effects to individual tracks and submixes, freeing up your limited supply of effects processors for reuse at mixdown. In addition, submixing to tape or disk can simplify your final mixdown process. If you print to analog tape, of course, the tape recorder operates as a signal processor in the sense that it can add a desirable sonic quality.

Whether you mix virtually or print your sequencer parts to tape, it’s generally better to clean up the signals at the sound modules rather than at the mixer. Selective filtering and compression can be used to remove unwanted frequencies or dynamics before a sound leaves the module. However, I usually try to save elaborate effects processing for the mixing environment, where I can employ dedicated units (unless I’m mixing on a DAW, where DSP is a precious commodity—in which case I might do some processing at the modules and some at the mixer). It’s a balancing act, but it’s better to have too many options than too few.

If your mixer doesn’t offer dynamic automation, you can use MIDI Control Change 7 (Volume) and 10 (Pan) messages to automate your sequenced tracks. The drawback to this is that a mixer channel will remain open even when no signal is present, which is not the case when you manually ride the faders. With an analog mixer, you’ll get some hiss; to fix this, you can use noise gates or expanders at the channel inserts.

In general, it’s good practice to maintain as many dedicated instrument channels as possible. Granted, sometimes you may have no choice but to submix several parts to a stereo output, for instance. Just avoid unnecessary submixing; the more signals you have to work with at mixdown, the better.

Acoustics and Electronics

There are three basic kinds of electronic sounds: emulations or samples of acoustic or electric instruments; completely artificial sounds; and emulations or samples of real-world, nonmusical sounds (a dog barking, for example). The mix engineer needs to approach each type differently.

When working with acoustic and electric instrument sounds, the goal is usually to replicate an accurate image of each instrument, positioning it in a realistic place on a virtual soundstage and making sure its frequency content is similar to what it would be in the real world. You’ll then give the recorded instrument a volume and depth on that stage by using level control, reverb, and sometimes other processing, such as delay. (This is not a hard-and-fast rule, obviously; there are no rules in the creative arts.)

The same philosophy usually holds true if the “acoustic” or “electric” instrument happens to be a sample or synth patch. For example, even though the Roland JV-1080’s grand-piano patches are electronic samples, I’d probably still EQ them like real pianos and put them in realistic positions in the stereo image, unless I was trying to achieve a weird result. However, the methods you use ultimately depend on the style of music you’re producing: for alternative and urban styles, perhaps the piano would need to be equalized like a guitar and spread across the entire stage.

When working with completely synthetic sounds, a different set of rules applies. Trying to place these instruments in a realistic spot on a “stage”—an acoustic environment where they wouldn’t normally be heard—is pointless. In addition, there are no real-world templates of synthetic sounds on which to base EQ settings; I mean, what is a Telefunken or a Space Warp Pad supposed to sound like, anyway? The same is true of nonmusical samples (unless, of course, one of your band members is really a barking dog). The only exception here is when you are creating music for picture and want to position the effects to match the action.

Working with synthetic sounds essentially gives you carte blanche to create exciting mixes with sounds coming from all over the stereo image and frequency spectrum. And using creative dynamics control and multi-effects processing, you can mold those sounds into practically anything you want.

Spatial Placement

You have much more creative liberty with a mix of electronic instruments than you do with a mix of acoustic ones. I like to create a natural-sounding blend of all the elements. Artists such as Beck, Nine Inch Nails, Jane’s Addiction, and Alanis Morrisette often employ contrasting timbres that don’t blend smoothly—but for a lot of music, a smooth blend is preferable.

In most cases, instruments and sounds should not compete with one another either spatially or spectrally. You should be able to hear every part of a mix and immediately identify which instrument is which. To do this, I try to conceptualize the mix as a three-dimensional stage (see Fig. 1). Panning instruments moves them across the width of the stage; altering their level and adding reverb or other delay effects determines how far back they are. The vertical axis represents frequency response (for example, cymbals would be toward the top of the stage, with the kick drum sitting near the bottom). This way you can graph the stage from either the front or above and see the two most important relationships of a mix: frequency/pan position and volume/pan position.

The goal is to make sure that no two components are centered at the exact same place in either graph. I don’t mean that things can’t overlap—the lower keys of a piano will inevitably be situated in the same area of the frequency spectrum as the bass—but elements shouldn’t blatantly sit on top of each other. This is what causes a mix to become cluttered and muddy sounding. A clear mix is achieved through careful planning and adjustment of level, pan, and EQ.