Panning for Gold

Mar 1, 2002 12:00 PM, By Randy Neiman

Once you've mastered the arcane science of signal routing, learned the ancient secrets of gain structure, and been initiated into the mysterious ways of equalization, compression, and myriad other stages in the black art of signal processing, you'll face the ultimate challenge: the final mixdown. That is where all those perfect takes — often separated by months, miles, and styles — must combine seamlessly beneath your skilled fingers to create something sonically balanced and musically cohesive.

Moving the faders comes easy: up is louder. Equalization skills come with time: find the pain and reduce the gain. But what's the rule of thumb for creating an effective soundstage? Where the heck do you place each instrument and effect in the stereo field? Is there some secret pan-pot code for creating a good stereo image? Randy Hoffner, a former audio engineer at NBC, once said, “Stereo does not equal mono times two,” and though that may seem obvious, stereo imaging is more than first meets the eye — or rather, the ear.

There is no rule of thumb for creating a soundstage; as with all things musical, the only regulation is that it sounds good (or at least the way you want it to sound). Still, most engineers agree that the general goal is a clear, uncluttered, three-dimensional soundstage in which the elements can all be clearly heard and their positions can be readily identified. For the novice, that is best achieved by learning first how to create a realistic soundstage — that is, one that sounds believable, as though the band were set up and playing right in front of you. Just as Picasso mastered realism before venturing off into cubism, the mix engineer is well served by an initial apprenticeship to nature — that is, to making things sound the way they actually do. Naturally, once you've mastered the rules, you will have a much better idea of when and how to break them.

This article covers what you need to know to create realistic soundstages. I'll explain how stereo imaging works, discuss critical tools of the trade, lay out general strategies, and offer practical tips, tricks, and caveats for constructing an effective, true-to-life soundstage. In addition, I have provided two detailed, real-world examples of soundstaging (see the sidebars, “Figuring It Out” and “Spinal Tap Dancer”) based on actual mixes you can download from and analyze at your leisure.

SOLID FOUNDATION

Although the word stereo commonly refers to any 2-channel system (whether audio or some other type), the original Greek word stereo meant “solid” in the sense of three-dimensional (having breadth, depth, and height). Combine stereo with the word phonic, which means “sound,” and you get stereophonic, or solid sound. Just as stereoscopic, or solid-vision, imaging employs two slightly offset photographs to create the illusion of depth (based on the fact that having two eyes a given distance apart allows for depth perception), stereophonic recordings create the illusion of three-dimensional space from two speakers. Simply put, stereophonic imaging works because having two ears allows for localization, the ability to perceive the direction a sound is coming from.

The human ear is an extremely complex and precise instrument. The structure and interrelationship of the outer, middle, and inner ears allow for maximum energy transfer at specific frequencies. Impulses generated by the nerve fibers in the cochlea (inner ear) are sent to the brain, which can accurately determine, among other things, pitch (to within 0.3 percent accuracy), distance, direction, and, if the source is moving, speed. Other cues related to timing and relative intensity provide subtle clues as to the nature and size of a room or environment — whether, for example, a sound is coming from a small room, a large auditorium, or the depths of a forest. All of that happens incredibly fast, with no need for the listener to think about the complex physics involved.

PANORAMIC VIEW

A primary factor in localizing sound is intensity, also known as volume. Volume gives an indication of distance, because a sound's intensity decreases as it moves away from the source. In a free field, sound emanates from the source in a spherical pattern and with a determined amount of energy created by the source; as the sphere gets bigger, the energy is reduced. The inverse square law states that sound pressure decreases proportionally to the square of the distance from the source. That works out to approximately 6 dB attenuation each time the distance doubles. In other words, louder sounds closer.

Timing is another primary audio cue. The Law of the First Wavefront states that when two coherent sound waves are separated in time by short intervals (less than 28 ms), the first signal to arrive at the ears will provide the dominant directional cues. Direct sound arrives at a listener's ears earlier than reflected sound (assuming the sound source is in an enclosed space); thus you can clearly detect location, even in very reverberant spaces. For a mono sound played through two speakers at equal volume (with the signal panned dead center), a delay of even 1 ms on either side can shift the image significantly to the right or left. Signals arriving more than 35 ms later than the original are interpreted as a distinct, separate echo from the source.

Although knowing the mathematics is not essential to creating an effective soundstage, it is important to understand how these two principles, intensity and timing, underlie the phenomenon of stereo imaging. By altering the relative intensity and timing of signals between two identical speakers in an enclosed space, you can create a believable sense of space in a mix. The mind does the rest of the work for you by localizing the source.

During the making of the 1940 movie Fantasia, audio engineers at Walt Disney Studios were asked to create the illusion of sound moving back and forth across the screen. Based in part on the earlier work of Dr. Harvey Fletcher (of Fletcher-Munson-curve fame) and his team at Bell Labs, the engineers determined that a sound source that is faded between two speakers seems to move between them, provided the total sound-pressure level (SPL) in the room remains constant. A special two-gang potentiometer (essentially a two-output, variable-voltage divider) was developed for which the sum of the log attenuations equaled a constant. The Disney engineers dubbed it the Panoramic Potentiometer, or pan pot for short.

As the pan pot is turned, a mono signal is sent to two channels simultaneously; the total intensity remains constant, but the difference in intensity between the two speakers provides the cues for localization. That is called intensity stereo. Thanks in part to the pioneering work of Les Paul, using pan pots and intensity stereo to create a lateral soundstage has been standard practice in popular music mixing since the mid-1950s.

TOOL TALK

Of all the tools necessary for creating a good mix, none is more important than two you already possess: your ears and brain. Those instruments collect the sound from the room, process and interpret it, and let you build a mix. Needless to say, keeping them in good shape is important. Listening to extremely loud music can dull and eventually damage your hearing. Recreational chemical use, too, can be deleterious. Not to be preachy, but some effects that make booze and other drugs fun — time distortion and heightened sensory perception, for example — can lead to bad or even disastrous results in the studio. Your mix may sound great to you now, but in a few days, once your head is clear, it very well may not.

If the ears and brain are the most important tools for mixing, the next is surely the monitors and, by extension, the room. I include the room because the premise of intensity stereo is intrinsically linked to keeping a constant volume within an enclosed space. The listening environment is thus a functioning part of the speaker system — a resonator, if you will. Together, the monitors and room provide the information to the ears and brain.

SPEAKER OF THE HOUSE

Many people (myself included) do much of their mixing in spaces that are far from ideal acoustically. But by applying acoustical treatments to the listening environment, you can produce better results immediately. A number of good articles about room treatment are available, in print and on the Internet, so I won't delve deeply into the subject. However, you can do some simple things now to guarantee optimal performance from your monitoring setup.

First, if possible, make sure your monitor speakers are not positioned in corners or directly against a wall, as that can cause low-frequency buildup and standing waves, making low frequencies seem louder than they really are. A simple Rule of Thirds can be employed to determine where to position the monitors in relation to the nearest wall: place the monitors at about one-third the depth of the room. For example, if your room is ten feet deep, place the monitors a bit more than three feet from the wall.

Even more critical is speaker placement in relation to the listener. In a proper mixing setup, the listener sits at one vertex of an equilateral triangle formed by him or her and the two monitor speakers; that is, the speakers are the same distance from each other as they are from the listener (see Fig. 1). That setup allows the sound from each transducer to arrive at the listener's ears at approximately, if not precisely, the same time. An offset of just a couple of inches on either side or a delay of just a few milliseconds can shift the resulting image significantly to the left or right, seriously degrading the stereo image.

It's also critical that speaker levels be matched. A discrepancy of less than 1 dB between the two speakers can shift the stereo image several feet in terms of the perceived location of the source; a difference of 20 dB moves the image completely to one side. Simply setting the levels of your amp or matching the trim on your active monitors until they look the same won't do — the potentiometers used in those devices can have variances of 5 percent or more. Furthermore, the cumulative effects of the component tolerance variances in the amplifiers can cause 1 to 5 dB of difference in loudness between the channels, even with the level knobs set identically.

An inexpensive solution is to buy an SPL meter from an electronics store. Even an inexpensive SPL meter, though not laboratory accurate, can measure the relative SPLs coming from your speakers well enough. Use a constant tone to begin with — 1 kHz is suitable for a quick measurement. First, pan the signal hard left and take a measurement. Next, pan the signal hard right and take another measurement. Then, alternate between several frequencies or use pink noise to measure the relative SPLs, adjusting amplifier or trim levels until they match.

DAILY MONITOR

The quality of your studio monitors is another important part of the equation. Buy the best system you can afford. Some people will readily spend a couple grand on a great microphone or a mic preamp and then grumble about paying that much for a pair of speakers. But this is no place to compromise; you use your monitors constantly to hear (and evaluate) every bit of audio in your studio, so ultimately, they are the most important tools you possess.

People interpret what they hear somewhat differently, so there is no one best choice for everyone. If there were, only one company would be making studio monitors. What's important is how they translate — that is, how your mixes sound once they leave your studio and get played on hundreds of other, quite different systems. Speakers that “sound good” or are “flattering” do not necessarily make the best choice. If everything played through a pair of monitors sounds great, some pretty awful mixes will leave your studio, because you'll have stopped working too soon.

That's why some engineers intentionally choose lower-quality, quite nonlinear speakers: to be able to judge how the mix will sound on similarly inexpensive consumer-playback systems. When used by someone with trained ears, such monitors (a certain white-coned model comes to mind) let mix engineers familiar with them make good decisions; I do not, however, suggest using inexpensive monitors as your main tools. Neither do I suggest using headphones as primary mix tools, because headphones are actually biphonic rather than stereophonic. Unless you intend to have your music heard only through headphones, use them sparingly, as a reference.

Select monitors that are as accurate, uncolored, and revealing as possible. A good test is to listen to material you're familiar with. If you don't hear parts that you know are there or if the image drifts or seems out of balance, those speakers aren't a good choice for you. If, on the other hand, the monitors allow you to hear subtleties you hadn't noticed before or cause you to begin to pick apart mixes you have long admired, buy them immediately; they're telling you what you need to know to make informed decisions about your mix.

If possible, try to audition monitors in your mixing space, because what sounds great in the store may cause resonance problems or give you listening fatigue at home. Mix a few songs with them and then listen to how well the mixes translate on different playback systems.

SIMPLE PLAN

Before you turn that first pan pot, take a quick inventory of all the elements to be placed in the mix. That initial assessment will help you form a mental picture of the soundstage, so you can better determine where each sound might best go and how much area it should occupy. If mixing an acoustic duet, for example, you may decide that each instrument should occupy a fair amount of the soundstage. On the other hand, in a dense mix with, say, bass, vocals, background vocals, percussion, two keyboards, three guitars, and ten tracks of drums, there obviously isn't room for each instrument to take up much of the soundstage. In that case, you will be seeking to create ample space for each instrument.

In general, each instrument should occupy a distinct area of the soundstage. The process of creating an area for an instrument is often referred to as carving out a space. That may mean not only finding the optimal panoramic placement for the instrument but also equalizing the sound so it doesn't mask or interfere with other instruments in the same or a similar frequency range.

When done well, that approach lets each instrument be heard distinctly, results in more dynamic range, and requires less volume from each element. Ideally, when the mix is complete, the soundstage will be clear and coherent. You should be able to close your eyes, clearly “see” the room in which the musicians are playing, and point to the position or area of each instrument.

However, don't hold to that (or any) plan rigidly; rather, use it as a starting point and let your ears be the judge from there. Imagine how you want the music to sound and begin to visualize each instrument in its place on the imaginary stage. As you start to place each instrument, some things may need to be moved or swapped around with others. That's fine as long as you leave sufficient sonic space for each element.

The practice of carving out space for instruments also helps prevent the common pitfall of fader creep. Fader creep results from, say, bringing up the bass to be able to hear it above the drums, then bumping up the guitar so it doesn't get lost behind the bass, then raising the vocals so as to hear them distinctly, and so on. Eventually, the console runs out of headroom, and the mix becomes unintelligible, incoherent mush. At that point, it's usually best to begin the mix anew.

It is good practice, especially if you're new at mixing, to closely examine and dissect mixes that you admire or that sound the way you want yours to sound. That will give you a foundation from which to work. But even if you don't try to cop a particular mix style, comparing your work closely with that of others can help you see where you may need work, enlighten you to new mixing techniques, and give you greater appreciation for masters of the craft. Take note of those mix engineers whose work you admire and listen to more of it; many veterans can pick out specific techniques, or even a particular engineer, based solely on the style of the mix, regardless of who the artist is.

THE EARS HAVE IT

A fairly common mistake, especially among beginners, is to mix with the eyes — that is, according to what looks right rather than what sounds right. Some engineers, for example, work through what seems a logical progression: placing the vocals and bass in the center, panning stereo signals (such as keyboards and drum machines) hard left and right, and putting mono instruments into standard 9:00 and 3:00 or 10:00 and 2:00 positions. At the other extreme, I have seen “perfectionists” take excruciating pains to measure and duplicate exact pan-pot positions from left to right, assuming that doing so will produce a more exact stereo image.

Neither method, however, is likely to produce natural or even pleasing results, because each is based on logic associated with visual rather than aural cues. The problem is pan pots used in many budget consoles show variances in the 6 to 10 percent range. Therefore, pan-pot positions on the mixer won't necessarily correspond precisely to instrument positioning in the stereo field. The best approach is to ignore the position of the pan pot and simply listen to the results of turning it.

Likewise, don't have too much faith in the meters on your mixer's stereo bus. Not only are they, too, probably less than perfectly calibrated, but on a typical budget console (with, say, 12-step LED ladders and a dynamic range somewhere between 84 and 90 dB), each segment represents at least 7 dB of gain — a resolution far too low for exactitude. I bring that up because I've seen engineers attempt to balance their stereo mixes by offsetting the master-level faders (on boards that offer separate left and right master faders) in order to have the LEDs line up perfectly.

Don't rely on mechanical pots and meters to tell you what your ears, which are much more sensitive instruments, can better discern. In short, mix with your ears, not your eyes.

HARDLY A HARD-PAN

Because most sound modules these days present you with stereo outputs, it would seem logical to simply hard-pan each left/right output to its corresponding left/right position. Indeed, that may sound great when you solo the stereo source. But what sounds great soloed may not blend well with the other instruments. Typically, when you pan, say, a stereo piano part (from a sound module) hard left and right, the keyboard will appear more or less centered on the soundstage. That may work fine in a mix of a duet — piano and flute, for example; but if you have five stereo keyboard parts and a stereo drum machine, all of them hard-panned left and right, the images will all appear across the center of the mix, largely on top of one another. Obviously, you can't get separation and a broad stereo image if everything is in the same place.

That doesn't mean you have to make all stereo parts mono. Rather, you can retain a sense of space by slightly offsetting the pan pots, thus creating a smaller space for the instrument to sit in. For a stereo piano, for example, you could position the pan pots at 9:30 and 10:30. That would still give it some space — a sense of dimension on the soundstage — but the piano would appear smaller and to the left of center stage.

The same thing applies for guitar preamplifiers with stereo outputs; although the soloed guitar may sound enormous with the preamp outputs hard-panned left and right, the sound will likely lose definition and luster in a busy mix. Again, carve out a smaller, more appropriate space.

Note that drum machines typically have preset pan positions for each drum, cymbal, or percussion hit within a particular kit. You should therefore learn how to change the panning of individual instruments from inside the drum machine; otherwise, you'll be stuck always having to work around the preset pan positions.

String sections and pads are perhaps the best candidates for wide stereo spreads; after all, a full string section would naturally take up a large area on a true soundstage. Even there, though, I would avoid hard-panning left and right.

In fact, I rarely pan anything but effects — or those elements that I want to stand out without getting lost in the center — to the hard left and right positions. After all, it would be rare in a concert performance for any instruments to be positioned directly to the left or right of the audience. So rather than filling all of the available space with a wall of sound, it's usually best — at least, in the interest of realism — to reserve the extreme pan positions for what is sometimes called headphone candy: reverb tails and other effects that provide some breathing room and a sense of space around the musicians. For example, in a band mix for which I pictured the band as playing on a 20- to 30-foot-wide stage, I would pan all of the instruments somewhere between 10:00 and 2:00 or at least no farther out than 9:00 and 3:00 (see Fig. 2); that would leave the areas between 7:00 and 9:00 on the left and 3:00 and 5:00 on the right for reverb and other effects to mimic the natural dispersion and reflection of sound from the band.

CONFLICT RESOLUTION

Earlier I mentioned the related roles of panning and equalization in the task of carving out a space for each instrument. There are further considerations when mixing instruments with similar frequencies and timbres — two electric rhythm guitars, for example. In short, try to keep them apart spatially. That is not to say they should be as far apart as possible or even directly opposite from each other (which often works, by the way); just make sure they are clearly discernible as separate instruments. In this case, the pan positions should not overlap, and generally, I recommend at least an “hour” or two of panoramic separation between them.

Fig. 3a shows some typical frequency distributions for a standard rock piece and how the frequency ranges overlap. Fig. 3b shows those instruments panned in such a way as to allow space for each as well as to account for natural sound-dispersion patterns as heard from the listener's perspective. Note that instruments of the lowest frequencies are typically panned toward the center. Low frequencies are less directional and thus more difficult to localize; they also require more power to be pushed from the transducers. Therefore, distributing low frequencies more evenly between the two monitors lets the speakers work more efficiently.

Note that frequency and timbre are not the only two realms in which elements may conflict. Elements may also conflict rhythmically or in terms of importance or centrality to the mix. In such cases, look for interesting ways to distinguish the elements so they no longer conflict but rather complement or offset each other.

3-D, BUT NO CRAZY GLASSES

Because pan pots are two-dimensional controllers and you're panning between two speakers, it's easy to fall into the habit of just laying everything out in a linear fashion — drums here, guitar there, and keys over there — as though the music were happening on a straight line. But when you listen to live music, you also hear the dimensionality or depth — the clear sense, for example, that the drummer and the percussionist are located behind the singer and the guitarist.

As mentioned previously, the sound localization is based on apparent loudness and timing. However, though louder elements tend to sound closer than quieter ones, what really creates a sense of depth within a mix is a judicious use of timing cues, usually through delays. In general, delays of less than 25 ms help create a sense of space; anything over 35 ms is perceived as a separate image or echo.

Under normal conditions and at sea level, sound travels at about 1.13 feet per millisecond. Therefore, 5 ms of delay will seem to move an image a bit more than five feet back into the soundstage (assuming equal volumes of the source and delay). Another way of saying the same thing, but in real terms, is that the sound from a snare drum positioned five feet behind a guitar amp will take about 5 ms longer to reach your ears than the sound from the guitar amp (assuming you are listening from the audience's perspective).

Understanding that principle can help you considerably in constructing a natural soundstage. For example, if you want the drums in a mix to sound as though they are close to the back of the soundstage, near the wall (as in a standard band setup), you could put a shorter predelay on them than you would on the instruments located in front of the drums, because the distance from the drums to the first reflective surface (the rear wall) is shorter. Depending on the depth of the stage, they'd also take about 6 to 8 ms or longer to arrive than anything closer to the front of the stage. Likewise, if you want to position a piano stage right near an imaginary side wall and you use a stereo reverb on the piano, you might want to set the predelay several milliseconds shorter on the right channel than on the left, because in reality the right side of the piano would be closer to a reflective surface.

When using effects such as room reverbs to help create a sense of space, be careful not to combine clashing or contrary-sounding rooms — for example, a small tile room for the drums and a concert hall for the vocals. That doesn't mean you have to use the same type of effect on everything; it just means that the effects should go well together to create a coherent sound. Again, though a particular effect sounds really cool on an instrument in Solo mode, it may clash or contradict when mixed in with everything else, destroying the illusion of a natural soundstage.

COMPATIBILITY ISSUES

A debate is still going on these days about mono compatibility, the main question being, “Does it matter anymore?” My answer is yes: it most certainly does matter.

Mono compatibility refers to how a mix holds up when played through a mono system (that is, when the two channels are summed to one). Phase problems, though perhaps not apparent in stereo playback, can result in dropouts, comb filtering, and other weirdnesses when the mix is played mono. Such phase problems may exist not only between the two channels of a stereo-recorded source but also between mono sources recorded simultaneously. Stereo effects, too, can be a culprit — what sounded spacious in stereo may evaporate or turn to mud when reproduced in mono.

Many consoles provide a Mono button that, when engaged, sums all channels of the mix to mono. That function provides a quick and easy way to check for mono compatibility. Use it. Mono playback remains ubiquitous in people's lives. Many television stations and cable networks still broadcast in mono, as do most AM and some FM stations. In addition, countless televisions, clock radios, computers, car radios, and other sound sources have only one speaker. Moreover, the stereo sound systems in many vehicles automatically sum to mono at lower volume levels so that half the music isn't lost to the driver.

TAILS OUT

You won't always want to create a natural-sounding stereo soundstage, so if the project calls for something decidedly unnatural sounding, by all means, go for it. But in many cases, a true-to-life soundstage is best for the music, and it's practically always a good starting point, no matter how much you end up deviating from natural by the time the mix is complete.

Regardless of the soundstaging strategies that you employ, avoid falling into a habitual approach — that is, automatically panning particular instruments to the same spot every time you mix. Not only will you bore yourself (and eventually your listeners) but, undoubtedly, you'll also fall short of turning out your best work. Every song and performance is necessarily unique, so it goes without saying that a formulaic approach to soundstaging will result in mixes that don't sound as good as they could.

Finally, don't rely on spatial enhancers or similar processors to improve your soundstage. If the soundstage doesn't sound right before such processing, it certainly isn't going to sound right afterward. Although I wouldn't claim there is never a time and place for spatializers and the like, you are cheating yourself if you rely on them to cover up mistakes. You're cheating your listeners, too.

---

Randy Neiman is an independent audio and marketing consultant living and playing golf in sunny Los Angeles. Share studio stories and golf tips with him at audioguru@mail.com.

FIGURING IT OUT

I was called on to remix a medium-energy pop single called “Figure It Out.” Instrumentation included an acoustic guitar, an electric bass, a drum machine, live percussion, a twangy Stratocaster, and female lead and backing vocals in the vein of Natalie Merchant and Sarah McLachlan. Upon arriving at the studio, I found a somewhat typical panning setup on the console: drum-machine tracks hard-panned left and right, the two channels of stereo acoustic guitar (miked with a large-diaphragm condenser near the sound hole and a small diaphragm near the neck) also hard-panned, vocals and bass sitting dead center, and the Strat and backing vocals panned to 11:00 and 2:00, respectively. The stereo reverb returns for the vocals were hard-panned left and right, as well.

The drum machine had its own effects, which were also hard-panned left and right. I panned them to the 8:00 and 4:00 positions to tighten the space, and then I placed a tabla part at about 2:30. Two percussion instruments of similar sonic spectra, shaker and tambourine, needed to be separated to avoid coherence (masking of one element by the other), so I put the shaker at around 3:45 and the tambourine at 8:30. The wide spacing between the two left plenty of room for the tonal instruments.

I found a nice spot for the stereo acoustic guitar by positioning the track recorded with the large-diaphragm mic just left of 11:00 and the other track slightly right of 12:00. That put the center image of the guitar a bit left of dead center and allowed for some interesting fret noise to the right. I panned the twangy Strat track far over, just shy of hard right, which let me bring its level down in the mix. I panned the Strat's reverb to the left side, which brought out a nice contrast to the main acoustic-guitar part.

People commonly pan kick drum and bass guitar right on top of each other, usually dead center, but a hint of separation between the two can provide clarity and a more natural sound — some breathing room, as it were — between two low-end instruments that often compete for a piece of the same sonic territory. I put the bass just right of center, at about 12:15, and used the drum machine's Pan menu to position the kick at 11:50. Snare drum fell in just right of the bass guitar and hi-hat slightly left of the kick, at 11:30. The result was a nice, tight image of the band playing together on a 12-by-15-foot stage in a room 30 feet wide and 50 feet deep.

Lead vocal is another element traditionally panned dead center, and there is an advantage to doing it that way: equal distribution between the speakers results in greater apparent volume, making the vocal sound more up front. But I often break with tradition and shift the voice slightly off center, usually to the right. Why? Assuming the song is released commercially, it will most often be heard in people's cars: that's the only controlled environment you can truly count on for playback. There the primary (or only) listener is usually the driver, which means the listening position is left of center, sometimes quite far so. A centered vocal will thus arrive at the listener's ears from the left speakers first, which shifts its arrival time and intensity. By mixing the lead vocal slightly right of center, I essentially compensate for the driver's offset sweet spot; the result is an apparent centered vocal for the driver. (No worries — most people will never catch the change, even when listening in their homes.) This trick also thwarts center-channel elimination devices, the boxes that attempt to derive a karaoke mix by removing any midrange elements that are panned dead center (which typically means vocals).

On “Figure It Out,” the female singer had a nice but not especially strong voice, so getting her vocal to stand out was a challenge, short of unnatural processing, excessive compressing, or pushing the volume too high. Fortunately, I had several takes to choose from, and I found one that, though similar to the primary take (and almost identical timingwise), had some interesting changes in inflection. I panned the primary take just right of center, to 12:45. I then put a hint of the alternate take even farther to the right, at about 4:30, and sent the reverb return almost exclusively to the left. That added some strength to the voice without causing the overt chorusing that can happen when you simply combine two takes. The left-panned reverb, set fairly hot and coming in from opposite the main vocal, helped make a thicker and wider vocal sound, and the slight changes in inflection gave the track more depth.

With the lead vox now happening, I began placing the four background vocals in spaces left open by the other instruments. This was no hairsplitting, knob-assessing affair; rather, I added the voices one at a time, without regard to pan-pot positions, simply listening until I could hear each in its own little space and then adjusting the levels so the four parts sat just right behind the lead vocal without detracting from it. To increase the sense of a real space, I delayed the reverbs slightly differently on each background vocal, based on the track's location in the stereo field. With all the instruments positioned on the soundstage, I increased the overall sense of space by panning those backing-vocal reverbs to hard left and right positions and opposite to their source tracks.

Keep in mind that none of the pan-pot positions I described (so precisely) was mapped out or clocked in advance. Rather, the mix was done by ear, without visual regard to panning. What mattered was naturalness and a sense of space, not commitment to an exact rotation of knobs. Only after the mix was complete did I bother to note pan positions, in case they were needed later (which they were — for this sidebar).

SPINAL TAP DANCER

More challenging than “Figure It Out” was “Tap Dancer,” from David Bryce's new CD, UltraMaroon (www.mp3.com/DavidBryce). This power ballad, centered around a piano part, has an acoustic guitar; two distinct, alternating bass lines (both synths); string pads; lead and backing vocals; four electric guitars; and a Hammond B-3 organ. All that is augmented in the B sections by a “horn section” — a blend of real horns and arranged samples. Percussion was done using an Alesis DM Pro module, which supplied tabla, tambourine, and shaker samples. The project's recordings were done in Mark of the Unicorn's Digital Performer 2.7, and many parts were already submixed as stereo pairs.

Listen to Tap Dancer on

My first step was to place the primary instruments (stereo acoustic piano, acoustic guitar, basses, and drums) into a space within which everything else would work. The song breaks down to those few instruments — piano and acoustic guitar floating over a soft string pad — at several points, so they needed to stand out to be heard distinctly at all times, without lots of volume changes.

I started the mix by establishing the intensity of the lead vocals, which are the loudest and most important element and therefore the one that largely determines the dynamic range for the entire mix. For many people, it seems most logical to begin a mix with the drums. That method is called additive mixing: building a foundation from the rhythm section (drums, bass, guitar) and then adding auxiliary and lead instruments in over the foundation, culminating with the vocal. Additive mixing, however, often hits a snag: as you begin to add in the other parts, the total volume keeps rising, and you eventually max out the console's dynamic range to get the vocals hot enough. To avoid that situation, I usually do the reverse, subtractive mixing, which means beginning with the loudest part and placing everything else in beneath it.

Having established the song's overall dynamic range, I positioned the stereo-recorded piano tracks, putting the left channel at 1:00 and the right at about 2:30. That gave the piano some space — a sixfoot grand, after all, won't sound natural coming from an area seemingly two feet wide — without taking up the entire soundstage. Next, I panned the acoustic guitar to about 10:45; that placed it about six to eight feet left of the soundstage's center.

I panned all the drums and their effects (except for the snare) from within the DM Pro module. The shaker ended up at just inside 3:00 and the tambourine at 10:30. I panned the drum overheads to just outside the 9:00 and 3:00 positions. Generally, I try to keep the drums themselves tucked around the center, as they would be on a real stage; however, I often make the reflections from the drums sound more live and intense as compared with those from other instruments. I brought kick and snare in on their own channels for better control; almost no EQ was applied to either.

As I added more instruments to the mix, it became increasingly difficult to find logical and distinct places to put them. For example, I'd ordinarily keep the Hammond organ part away from other keyboard instruments, but to get the guitars and horns to sit right, it ended up just to the right of the piano, around 3:15. This example illustrates the importance of using your ears and basing your pan and other mix decisions on the piece itself. No matter what the style of music, if you listen closely, it will tell you what needs to be done.

In “Tap Dancer,” the electric guitars take precedence over the horns, both in terms of frequency range and number of bars played; therefore, I brought in the electric guitars next. Three parts were submixed to a stereo pair from within Digital Performer, so I did the panning in Digital Performer and left the console channels hard-panned left and right. The power chords fit nicely at around 9:00. I slipped in the delayed “seagull” guitar parts at 3:00, between the piano and Hammond. The delay was set at about 3 ms, just enough to place the sound back a smidgen. The solo guitar occupied its own channel and was quite easily placed just inside 11:00, which put it about four to six feet to the left on the soundstage. I delayed it about 8 ms and set its reverb predelay about 25 ms longer than that on the other reverbs.

The horn section comprised four live horns submixed to a pair of channels and a five-part arrangement of horn samples submixed to another pair of channels. Building the section took some time, and I don't have sufficient space to describe the process in detail; suffice it to say that the net result was an interleaving of the nine parts onto the four channels. I positioned the horn section to the left of the solo guitar, between 8:30 and 9:30, so it took up about five feet of the soundstage.

As in “Figure It Out,” I panned the lead vocals just to the right of center, at 12:45. I then inserted several channels of individual backing vocals and vocal sections into the few remaining openings and panned each corresponding reverb return a bit farther out than opposite its source. String pads were the hardest panned parts, with one channel at 8:15 and the other at 3:45; that translates to nearly the entire span of the 30-foot-wide soundstage I imagined for the song.