″The rubble over which you are stumbling
Just isn′t that hard.
What did you do with all that grace, now?
What did you do with all my wine?
What makes you think that just′ cause you dress bright,
it means that you shine?″
— ″HALF HARVEST,″ MICHAEL PENN, MARCH (BMG MUSIC, 1989)
Recording engineers are musicians. Sometimes, recording engineers are excellent piano players, drummers, or singers. Always, recording engineers are musicians. Their instrument is the recording studio. The loudspeakers will play back a multitrack production that is artistically successful only if the engineer knows a good deal about music and how to play their instrument, which is the equipment of the studio. The challenge of coaxing art out of technology pervades the entire production process, but nowhere is it more apparent than at the mixdown session. The technical mastery and musical ability of the engineer are revealed in the mix session as every bit of equipment is applied to any and all elements of the multitrack production. This chapter offers a starting point in the highly nonlinear, unpredictable creation of a successful multitrack mix.
How do you write a song? How do you create a web page? The answer varies from person to person and from project to project. Creating a mix is a similar experience. That is part of the pleasure of the recording arts. No rote tasks are repeated daily in the creation of multitrack music. Engineers keep learning, creating, and exploring. There is no single right way to do it. There are no hard-and-fast rules to follow. There is, however, something of a standard approach to mixing pop music that is worth reviewing. It is not the only way. In some cases, it is not even the correct way. But it is a framework for study, a starting point from which any engineer can takeoff in their own direction.
Consider a pop/rock tune with the following somewhat typical set of challenges: drums, bass, rhythm guitar (doubled), lead guitar, clavinet, lead vocal, and background vocals. Where does the session start?
First, lay out the console. Using either a mixing console or a digital audio workstation, it helps to preset the signal flow with as much as can be anticipated. Mixing requires the engineer to be creative in how they shape and combine the various tracks and effects. Yet the engineer must hook everything up correctly too. The latter tends to interfere with the former, so it helps to do a chunk of the tedious and technical thinking ahead of time so that it does not interfere with the flow of inspiration while mixing.
All of the various signal-processing effects needed in a mix cannot be fully known before the mix is begun. But some basic effects are so frequent that they are very likely to be needed and can be patched up ahead of time. Many mixes will make use of a long reverb (hall-type program with a reverb time over two seconds); a short reverb (plate or small to medium room with a reverb time around one second); a spreader (explained below); and some delays (dotted eighth note, quarter note, or quarter-note triplet in time). Before digging into the details of the mix, the engineer will often launch the appropriate plug-ins or patch in the appropriate hardware to get these anticipated effects going. By setting them up ahead of time, these effects are available with minimal additional effort. The engineer can instantly send a bit of vocal, snare, and lead guitar to the same effect. The way to have all these effects handy is to use aux sends (see ″Outboard Signal Flow″ in Chapter 2). For example, the mix engineer might lay out the mixer so that aux 1 goes to the long reverb, aux 2 feeds to the short one, aux 3 sends a signal to the spreader — whatever is most comfortable for the engineer. The effects are returned to empty, available monitor faders on the mixer or digital audio workstation.
If the mixer is hardware based, the session may eventually run out of aux sends. The trick then is to use the track assignment network to get to additional sends. Sending a signal to track bus 1 is the way to record to track 1, of course. When mixing, not multitracking, the engineer can use the track bus as a way of accessing additional effects units.
Consider the production challenge of adding a delay to the electric guitar, using an additional delay unit not previously connected to the mixer. Making matters worse, all aux sends have been used. The electric guitar signal is sent to the stereo mix, of course. In order to add a new delay effect to the guitar, send it to track bus 1 as well as the mix bus. Patch track bus 1 send into the delay input. Patch the delay output into an available return to the mix bus on the console. In this way, the guitar goes to both the mix bus and the signal processor, using repurposed track busses as extra aux sends on the console.
What is the spreader? It is often desirable to take a mono signal and make it a little more stereolike. A standard effect in pop music, an engineer will spread a single track out by sending it through two short delays. Each is set to a different value somewhere between about 15–50 milliseconds (ms); if too short, it starts to flange/comb filter (see ″Short Delay″ in Chapter 9), and if too long, it pokes out as an audible echo. One delay return is panned left and the other panned right. The idea is that these quick delays add a kick of supportive energy to the mono track being processed, sort of like the early sound reflections that would be heard if the instrument were played in a real room. It is certainly more satisfying for a musician to practice in the kitchen than in the linen closet. The kitchen has cookies, which is nice. But it also has some shiny hard surfaces (not the cookies) that help musicians hear themselves. The strong sound reflections can make the performance sound better, stronger, more exciting. A linen closet absorbs all this, making for a dull place to practice. Then there is the whole ″no cookies in the linen closet″ rule enforced in most homes. So the spreader takes a single mono sound and sends it to two slightly different, short delays to simulate reflections coming from the left and right.
Any two delay lines can be used as a basis for the spreader effect. The effect is further refined by pitch shifting the delayed signals ever so slightly. That is, detune each delay by a nearly imperceptible amount, maybe 5–15 cents. Again, a stereo effect is the goal, so the spreader requires slightly different processing on the left and right sides. Just as a slightly different delay time was specified for each output, the engineer dials in a slightly different pitch shift as well (maybe the left side goes up 9 cents while the right side goes down 9 cents).
Introducing these slight pitch shifts to the short delays abandons any basis in acoustic reality that the spreader might have had. The effect is not just simulating early reflections anymore. Now the mix engineer is taking advantage of signal-processing equipment to create a widened stereo sound that only exists in loudspeaker music; it is not possible in the physical world. This sort of thinking is a source of real creative power in pop music mixing: Consider a physical effect and then manipulate it into something that is better than reality (good luck, and listen carefully).
This effect might be added to the lead vocal track, among others. The lead vocal track is most likely going to be panned straight up the middle. In order for the spreading effect to keep the vocal centered, it helps to do the following. Consider the delay portion of the spreader only. If one listens to the two panned short delays (the reader is encouraged to step through this in the studio), the stereo image pulls toward the shorter delay. For example, with a 30-ms delay panned left and a 20-ms delay panned right, a listener sitting on the median plane will hear the sound coming from the right, even as the unprocessed track sits panned dead center. Now consider just the pitch-shifted part of the spreading equation. The higher pitch tends to dominate the image. With a 9-cent pitch up to the left and 9-cent pitch down to the right, the image shifts ever so slightly left, toward the higher, brighter sound. The full spreader calls for both delay and pitch shift. Arrange it so that the two components balance each other out (delay pulls right while pitch pulls left). This way the main track stays centered. Experiment with different amounts of delay and pitch change. Each offers a unique signature to your mix. When overused, the vocal will sound too mechanized, too processed. When conservatively applied, the voice becomes bigger and more compelling; singers like this, and so does the record-buying public.
With the console laid out, the engineer can start mixing. The vocal is almost always the most important element of every pop song, no matter how many tracks there are. So most engineers start with … the drums. Starting with the vocal makes good sense, because every track should support it. But easily 99% of all pop mixes start with the drums.
The drums are often the most difficult instrument to get under control in the recording studio. It is an instrument with at least eight, generally more, separate instruments playing all at once in close proximity to each other (kick, snare, hi hat, two or three rack toms, a floor tom, a crash cymbal, a ride cymbal, and all the other various add-ons the drummer has managed to collect). It is hard to hear the problems and tweak the sounds of the drums without listening to them in isolation. So engineers tend to start with the drums so that this instrument can be addressed in isolation. Once the vocals and the rest of the rhythm section are going, it is difficult to dial in just the right amount of compression on the rack toms.
What does an engineer do with the drums? The kick and snare are generally the source of punch, power, and tempo for the entire tune — they have simply inspire awe. So it is natural to start with these tracks. Step one: Keep them dead center in the mix. The kick, snare, bass, and vocal are all so important to the mix that they almost always live center stage. The kick needs both a clear, crisp attack and a solid low-frequency punch. Equalization (EQ) and compression are the best tools for making the most of what was recorded. The obvious: EQ boost at around 3 kHz for more attack and EQ boost at about 60 Hz for more punch. Not so obvious: EQ cut with a narrow bandwidth around 200 Hz to get rid of some muddiness and reveal the low frequencies beneath (see Chapter 5).
Compression does two things for the kick. The first goal of compression on kick is to manipulate the attack of the waveform so that it sounds punchy and cuts through the rest of the mix. Chapter 6 describes the sort of low-threshold, medium-attack, high-ratio compression that sharpens the amplitude envelope of the sound. Second, compression controls the relative loudness of the individual kicks, making the slightly weaker kicks sound almost as strong as the more powerful ones. Drumming is physical work. Drummers understandably get tired in the fifth chorus of take 17. Compression helps them out.
Placing the compressor after the equalizer lets the engineer tweak the sound in some clever ways. The notch around 200 Hz keeps the compressor from reacting to that unwanted murkiness. As the engineer pushes up that low-frequency boost on the EQ, the compressor reacts. With an aggressive low frequency boost, the compressor is forced to yank down the signal hard. In this way, heightened low-frequency punchiness makes the drums sound larger than life.
The snare is next. As a starting point, it likely gets a similar treatment: EQ and compression. The buzz of the snares is broadband, from 2 kHz on up. The engineer picks a desirable range: 8 kHz might sound to edgy or splashy, but 12 kHz starts to sound too delicate. Try 5 kHz. It is a subjective, musical decision. A low-frequency boost for punchiness is also common for snare. Typically the engineer looks higher in frequency than was done on the kick, maybe 80–100 Hz or so. The engineer might also look for some unpleasant frequency ranges to cut. Somewhere between 500 or 1,000 Hz lives a cluttered, boxy zing that does not help the snare sound and is only going to fight with the vocal and guitars anyway. One simply finds a narrowband to cut on the snare and the rest of the mix will go more smoothly.
The snare definitely benefits from the addition of a little ambience. In many mixes, it is not unusual to send it to the short reverb already set up and/or hope to find some natural ambience in the other drum tracks. The overhead microphones are a good source of supportive, natural snare sound. Any recorded ambience or room tracks should be listened to now. A gate across the room tracks keyed open by the close microphone on the snare can create a subtle touch of ambience on each snare hit (see ″Keyed Gating″ Chapter 7).
With the kick and snare punchy and nicely equalized, it is time for the engineer to raise the overheads and hear the kit fall into a single, powerful whole. The overheads have the best ″view″ of the kit and the snare often sounds phenomenal there. The mix engineer carefully blends the overhead tracks with the kick and snare tracks to make the overall drum performance congeal into a single, powerful event. It is tempting to add a gentle high-frequency boost across the overheads to keep the kit crisp. An engineer must listen carefully though, because if the tracks are already bright as recorded, additional high-frequency emphasis will lead to a harsh, un-pleasant drum sound. In fact a gentle and wide presence boost between 1 and 5 kHz can often be the magic dust that makes the drummer happy.
When the toms are on separate tracks, the engineer will again reach for those tried and true equalizers and compressors. Creativity is required, but nominally the engineer might EQ in a little bottom, and maybe some crisp attack around 6 kHz. One should consider EQing out some 200-Hz muddiness, as was done with the kick. Compress for attack and punch and the drum mix is complete, for now.
With drums tentatively set, the bass guitar needs the engineer′s attention. One often needs compression to balance the bass line. Some notes are louder than others, and some strings on the bass are quieter than others. Gentle compression (4 : 1 ratio or less) can even out these problems. A slow attack time adds punch to the bass in exactly the same way it did on the drums.
Release is tricky on bass guitar. Many compressors can release so fast that they follow the sound as it cycles through its low-frequency oscillations. That is, a low note at, say, 40 Hz, cycles so slowly (once every 25 ms) that the compressor can actually release during each individual cycle. This cycle-by-cycle reshaping amounts to a kind of harmonic distortion (see Chapter 4). Engineers typically slow the release down so that it does not distort the waveform in this way. The goal is for the compressor to ride the sound from note to note, not cycle to cycle.
The obvious EQ move is to add low end. But one must be careful as the track may already have a lot of low end. The bass player likely seeks it out. The tracking engineer likely emphasized it. The trick at mixdown is to get a good balance of low frequencies from 30 Hz through 300 Hz. The mix engineer must listen for a hump in the response — either too much or too little in a single low-frequency area. The engineer simply equalizes in the appropriate correction.
At this point, it makes sense for the engineer to glance back — with their ears — at the kick drum. If the kick sound is defined in the low end by a pleasing emphasis around 65 Hz, then one might need to make room for it in the bass guitar with a complementary, but gentle, cut. The trick is to find EQ settings on both the kick and the bass so that the punch and power of the kick does not disappear when the bass fader is brought up.
It is not unusual to add a touch of chorus to the bass (see ″Medium Delay″ in Chapter 9). This is most effective if the chorus effect does not touch the low frequencies. The bass provides important sonic and harmonic stability in the low frequencies. A chorus, with its associated motion and pitch bending, would undermine this. The solution: Place a filter on the send to the chorus and remove everything below about 250 Hz. The chorus effect works on the overtones of the bass sound, adding that desirable richness, without weakening the songs foundation at the low end.
For this basic mix, the rhythm guitars are doubled. It is a rock-and-roll cliché to track the same rhythm guitar twice. The two tracks might be identical in every way except that the performance is slightly, humanly different. Panned apart, the result is a rich, wide, ear-tingling wall of sound. The effect is better still as the subtle differences between the two tracks are stretched slightly. Perhaps the second track is recorded with a different guitar, different amp, different microphones, different microphone placement, or some other slightly different sonic approach.
In mixdown, one makes the most of this doubling by panning them to opposite extremes: one goes hard left, the other hard right. It is essential to balance their levels so that the net result stays centered between the two speakers. A touch of compression might be necessary to control the loudness of the performance, but often electric guitars are recorded with the amp cranked to its physical limits, giving it amplitude compression effects already. Complementary EQ contours (boost one where the other is cut and vice versa) can add to the effect of the doubled, spread sound.
The clavinet completes the rhythm section in this basic mix. It probably needs compression to enhance its attack using much the same philosophy that was employed on the kick, snare, and bass guitar. Giving it a unique sound through EQ and effects ensures it gets noticed. Good starting points would be to add some flange or distortion (using a guitar foot pedal or an amp simulation plug-in) to make it a buzzy source of musical energy. Panning it midway off to one side makes effective use of the stereo sound stage. Depending on how the performances interact, a good approach is to pan it opposite the high hat, or the most active toms or the solo guitar in an effort to keep the spatial counterpoint most exciting. Add a short delay panned to the opposite side of the clavinet for a liver feeling.
With drums, bass, guitar, and clav carefully placed in the mix, a first draft of the rhythm section is complete. It is time at last to add the fun and important parts: vocal and lead guitar.
The vocal gets a good deal of the engineer′s attention now. The voice must be present, intelligible, strong, and exciting. Presence and intelligibility live in the upper middle frequencies. Typically the engineer uses EQ to make sure the consonants of every word cut through that rich wall of rhythm guitars. A careful search from 1 kHz to maybe 5 kHz should reveal a region suitable for boosting that raises the vocal up and out of the clutter of the guitars and cymbals. The engineer might have to go back and modify the drum and guitar EQ settings to get this just right. Mixing requires this sort of iterative approach. The vocal highlights a problem in the guitars, so one goes back and fixes it. Trading off effects among the competing tracks, the engineer seeks to find a balance between crystal clear lyrics and perfectly crunchy guitars.
Strength in the vocal will come from panning it to the center, adding compression, and maybe boosting the upper lows (around 250 Hz). Compression controls the dynamics of the vocal performance so that it fits in the crowded, hyped-up mix that is screaming out of the loudspeakers. All of this compression and EQ track by track has so maximized the energy of the song that it will not forgive a weak vocal. Natural singing dynamics and expression are often too extreme to work, because either the quiet bits are too quiet or the loud screams are too loud, or both. Compressing the dynamic range of the vocal track makes it possible to turn the overall vocal level up. The soft words become more audible, but the loud words are pulled back by the compressor so that they do not over do it.
The vocal, a tiny point in the center, risks seeming a little small relative to the drums and guitars. The spreader effect is designed to combat this problem. Sending some vocal to the spreader helps it take on that much desired larger-than-life sound. As with a lot of mix moves, one may find it helpful to turn the effect up until it is clearly too loud and then back off until it is just audible. Too much spreader is a common mistake, weakening the vocal with a chorus-like sound. The goal is to make the vocal more convincing, adding a bit of width and support in a way that the untrained listener would not notice as an effect.
Additional polish and excitement comes from maybe a very high-frequency EQ boost (10 kHz or 12 kHz or higher) and some slick reverb. The high-frequency emphasis will highlight the breaths the singer takes, revealing more of the emotion in the performance. It is not unusual to add short reverb — try a plate — to the vocal to add midrange complexity and enhance the stereoness of the voice still further. The engineer likely adds a long reverb to give the vocal added depth and richness. Sending the vocal to an additional delay or two is another common mix move. The delay should be tuned to the song by setting it to a musically relevant delay time (maybe a quarter note). It is mixed in so as to be subtly supportive, but not exactly audible. The natural next step is to add some regeneration on the delay so that it gracefully repeats and fades. Taking it further still, the engineer might send the output of the delay to the long reverb too. Now the singer′s every word is followed by a wash of sweet reverberant energy, pulsing in time with the music.
EQ, compression, delays, pitch shifting, and two kinds of reverb represent, believe it or not, a normal amount of vocal processing. It is going to require some experimentation to get it all under control. By turning up the various pieces of processing too loud and then backing off, engineers learn the role each effect plays. It certainly requires going back and forth among every piece of the long processing chain. Change the compression, turn up the delay, turn down the reverb, and back to the compressor again. With patience and practice, one finds even elaborate combinations of effects easy to control.
And that is just a basic patch. Why not add a bit of distortion to the vocal? Or flange the reverb? Or distort the flanged reverb? Anything goes.
The background vocals might get a similar treatment, but the various parts are typically panned out away from center, and the various effects can be made more prevalent in the mix. One cliché is to hit the spreader and the long reverb a little harder with background vocals to help give them more of that magic, pop sound. Intelligibility might be less important for background vocals. Repeated words or call and response lyrics can often be understood through context, freeing the engineer to pull out some presence in the middle frequencies and emphasize other timbral details in less competitive spectral spaces, if they wish.
The lead guitar can be thought of as replacing the lead vocal during the solo. Because they do not occur simultaneously, the solo guitar does not have to compete with the lead vocal for attention. The mix challenge is to get the lead guitar to soar above the rhythm section very much as the lead vocal was required to do. An EQ contour like that of the lead vocal is a good strategy: presence and low-end strength. Unlike the vocal, however, many styles of guitar require little to no compression. Electric guitars are naturally compressed at the amp when tracked at maximum volume. Additional reverb is also optional for guitars. The overall tone of the guitar might be fully set by the guitarist when the amp was setup and the settings were dialed in — that includes the reverb built into the amp.
Solo guitar might get sent to the spreader, and it might feed a short slapback delay. The slap delay might be somewhere between about 100– 200 ms long. It adds excitement to the sound, adding a just perceptible echo reminiscent of live concerts and the sound of the music bouncing back off the rear wall. It can be effective to pan the solo about halfway off to one side and the slap a little to the other. If the singer is the guitarist, it might make more sense to keep the solo panned to center. Of course, one can always add a touch of phaser, flanger, or something from the long list of effects in the digital multi-effects unit. One can even add additional distortion. Such significant tone changes should probably be made with the guitarist present to get the blessing of an expert in the field.
The entire stereo mix might get a touch of EQ and compression. As this can be done in mastering, it is wise to resist this temptation at first. With experience, one should feel free to put a restrained amount of stereo effects across the entire mix. The engineer is trying to make the mix sound the best it possibly can, after all. For EQ, usually a little push at the lows around 80 Hz and the highs around or above 10 kHz is the right sort of polish. Soft compression with a ratio of 2 : 1 or less, slow attack, and slow release can help make the mix coalesce into a more professional sound. As the entire mix is going through this equipment, it is essential to use very good sounding, low noise, low distortion effects devices.
That sums up the components of one approach to one mix. It is meant to demonstrate a way of thinking about the mix, not the step-by-step rules for mixing. Every song demands different approaches. The sometimes frustrating fact is that every multitrack production triggers so many ideas that the engineer does not know where to start. Let this chapter be a guide when the long list of options becomes paralyzing. The best mix engineers can stay in control of the many forces that conspire to reduce the sound quality of the mix, while reacting freely to the countless creative urges that might help the music transcend mediocrity and become real audio art.