Additional material by PAUL WHITE.Conventional wisdom has it that to make a sound appear louder in a mix, you just turn up the level -- but by applying a little psychoacoustic theory, you can change the perceived loudness of a sound while barely affecting its level at all. By applying the correct techniques, you can make a sound sit at the front of a mix, or have it settle demurely into the background, and the results can appear far more natural than working with level alone.
Audio perception is inextricably linked with survival, and the human hearing system is most interested in the first few hundred milliseconds of a sound, because that gives clues as to whether the source of the sound is a threat -- whether or not it's a sabre-toothed tiger, for example. Consequently, what happens during that first few hundred milliseconds greatly affects the perception of how 'loud' that signal is, as well as its relationship to other sounds happening at the same time. In other words, the sounds we register as most important and potentially threatening are the ones that grab our attention. If we develop this idea, we can devise different ways of changing the prominence of a sound in a music mix, without using traditional level changes.
Given two sounds of equal loudness that play at almost the same time, the one that started first will appear the more prominent. This is the basis of the Haas or precedence effect. For example, suppose you have kick drum and bass guitar hitting at the same time; if you want the bass guitar to be a little more prominent than the kick drum, try moving it ahead of the kick slightly. Conversely, to push the bass behind the kick, move it so that it sounds slightly later than the kick.
The way to move sounds depends on your recording medium. With MIDI sequencers or modular digital multitracks, a track shift function will do the job. With hard disk recorders or 'Audio plus MIDI' sequencers, you can simply grab a part on-screen, and shift it or type in a track delay. With analogue recording, a simple delay unit can be used to delay any track. You'll need to set the feedback to zero, the effect level to max and the direct level to off. The delay should be patched in via the appropriate mixer insert point, and you need a range that will allow you to set up between 1 and 50ms of delay. The effect must be set by ear, because if you go too far, the musical feel will be affected.
In nature, the closer a sound, the brighter it is, which means that both high-frequency EQ and enhancement are useful tools in creating a sense of proximity. If you want to bring just a couple instruments out from a mix, patch an exciter/enhancer or 'tube distortion' device (depending on whether you're looking for a cleaner or a grittier sound, respectively), and use it to process the sounds you want to bring out.
Most exciters are designed to work in console insert points, but with most models, you'll find that you can also get away with using them in the aux send loop, providing you turn the effect balance full up. If you have a unit that works OK this way, you have the benefit that you can turn up the aux send for individual channels, to make them stand out from the mix to a greater or lesser degree. Most people use enhancement devices to process an entire two-track tape, but they also work well when used in the way described above, because they allow you to create contrast between the different sounds in a mix. If your enhancer isn't happy working in the aux send system, you could patch it into a group insert and then route all the sounds to be treated to that group. Figure 1 shows an enhancer patched into the aux send return system of a console.
Just about any anti-digital diatribe focuses on how tubes, and analogue gear in general, create a mellower, more rounded sound compared to digital media such as CDs or DAT. No matter which side you take, one thing is clear: the sound is at the very least different -- and you can use this to your advantage. As just one example of how to change the mix with tubes, try recording background vocals through a tube preamp, and the lead vocal through a solid-state preamp (or vice-versa). Assuming quality circuitry, the 'tubed' vocals should be less 'in your face' than the solid-state one, although some tube preamps behave as exciters and sound even more upfront, so let your ears decide.
Percussion seems to work well through tubes too, especially when you want it to feel less prominent than the trap drums. If you don't have access to tube and digital sound processors, try out your mic collection to see how different they sound. The warmer mics can be used for sounds that you need to keep at the back of the mix, while the brighter mics can be used to push sounds to the front. Often, you'll find that different mics are capable of tonal nuances that can't be replicated with EQ.
Human hearing has a tendency to pay less attention to constant sounds, and you can use this to your advantage by slightly detuning your synth patches to make them more noticeable. This involves doing a little programming of your synth, but the effect is worth it. Take a choir patch that has two layered chorus sounds (the dual layering is essential). If you want this sound to grab the ear more, use a pitch envelope to add a slight downward pitch bend from slightly sharp to concert pitch on one layer, and a slight upward pitch bend to concert pitch on the other layer. The pitch difference doesn't have to be very much to create a more animated sound. By the same token, adding a little delay vibrato to a sound breaks up the monotony, making it more noticeable.
While movement is supposed to make a sound stand out, the regular movement imposed by a chorus unit has the effect of diffusing your attention, so if you want to weaken a signal, a chorus/flanger can help a lot. The phase interaction can also cause a real (as opposed to subjective) level drop, if you set the effect up properly.
Set the effect for a short delay (under 10ms or so), and use an out-of-phase output mix, where the output control which blends straight and delayed sounds reads -50 instead of +50. Alter the mix by starting with the straight sound, then slowly add in the delayed sound. As the delayed sound's level approaches the straight sound's level, a comb-filtering effect comes into play which essentially peppers the signal's frequency spectrum with holes. If you're trying to make a piano or guitar take up less space in a track, this technique works well. Similarly, adding a lot of reverb can create the effect of distance. So, now you know why your heavily effected lead vocal never sounds quite as in-your-face as you might like!
As hinted at earlier, EQ is a very under-utilised resource for mixing. Turning the treble down instead of the volume can bring a track more into the background, without it becoming 'smaller'; just less 'present.' A lot of engineers go for really bright sounds for instruments like acoustic guitars, then turn down the volume when the vocals come in (or some other solo happens). Try turning the brightness down a tad instead. If you have an effects unit that can create EQ effects under external MIDI control, you can use a sequencer to automate these EQ changes for you -- so you don't even have to expend the effort of moving little faders and switches.
Though most mixes are initially balanced using the level faders, by applying the techniques mentioned here, you can make a mix appear more dynamic and less cluttered, without having to keep juggling the levels. If you're painting a picture, you don't place items in the background just by making them smaller, you also choose more subdued colours -- and so it is with audio mixing. You can also use the ideas discussed to make background sounds more noticeable, which will allow you to mix them at a lower level without them getting lost. This can be important in a busy mix, where you find yourself running out of space all the time.
Even in a very basic home studio, you have a lot of options when it comes to mixing -- and more coming along every day. Try some of these ideas, and I'm sure you'll add at least some of them to your repertoire of mixing techniques.
With a hard disk recorder or 'Audio with MIDI' sequencer, you can do some cool fade-ins to make an attack less prominent. However, if you do a fade starting from the beginning of a sound, you'll lose the attack altogether, so instead, extend the start of the fade to before the sound begins, as shown here. The audio then doesn't come up from zero, yet you definitely reduce the attack, which makes the sound less attention-grabbing.
One common technique used to strengthen voices is doubling, where a singer sings a part, then tries to duplicate it as closely as possible. The slight timing variations add a fuller effect than doubling the sound electronically. However, panning or centring these two tracks makes a big difference during mixing. When centred, the vocal lays back more in the track, and can tend to sound less full.
When panned out to left and right (this needn't be an extreme amount), the sound seems bigger and more prominent. Some of this is also due to the fact that when panned together, one voice might cover up the other a bit, and this doesn't seem to happen as much when panned. You also remove the risk of phase cancellation by panning the two takes left and right. Some singers are so accurate that when you pan their doubled vocals to mono (centre), you can hear the levels beating as the signals drift in and out of phase.