″Life′s the same
I′m moving in stereo.
Life′s the same
Except for my shoes.
Life′s the same,
You′re shaking like tremolo.
Life′s the same,
It′s all inside you.″
— ″MOVING IN STEREO,″ THE CARS, THE CARS (ELEKTRA/ASYLUM RECORDS, 1978)
No study of recording studio effects would be complete without the careful consideration of mix automation. The power of studio effects cannot be fully realized without it. This chapter is a tutorial on mix automation, appropriate for any automation system, analog or digital, on a console or in a digital audio workstation.
So what is so automatic about automation? An automation system plays back an engineer′s mix moves, however elaborate, automatically. That is all automation can do. It repeats the mix done by someone else. It is nothing without the engineer. In order to see what can be done with automation, it makes sense to take a look at what can be done without automation. Mixing without automation is called manual mixing. The following is pretty typical:
- Intro: all vocals cut, extra reverb on the strings, fade organ in. Bass enters at bar four.
- Verse 1: guitar, drums, and lead vocal in, keep background vocals out, less reverb on the strings, pan organ left, and make room for horns on the right at bar 12.
- Chorus 1: lead vocal double comes in, six background vocal tracks up and perfectly blended (three part harmony, doubled), gated room sound added to snare, strings out, acoustic guitar in.
That is a lot for one engineer to do all at once, so the assistant engineer, the studio manager, and the Chinese food delivery person help out. There is a lot to remember, so notes are scribbled on the track sheet, the console, and the Chinese food menu. It is hard to listen critically while doing so much, so in truth the engineer does not hear the mix objectively until the master is played back later.
Oops, the horns were too loud on verse 3? Try again. Do everything the same way as the last mix, but get the horns right on the third verse. This is not easy. It takes several more passes to get close to that last good version, the whole time the mix team is trying to remember the horns and not forget anything else.
When everyone has run out of the ability to remember another thing, it is time to start mixing in pieces: ″Well, we finally got the intro right. Now, let′s move on to the first verse …″ The song is mixed section by section and later edited together into what sounds like a single pass.
That is manual mixing, which can be summarized as:
- All hands on deck. Only an octopus with golden ears could make all these moves at once, so the production pulls in the help of others in the control room.
- Extensive documentation. Make creative but informative notes about everything that changes in the course of the mix, putting tick marks next to the faders for their key levels at each part of the song, sticking red tape on the pan pot that gets twisted left in every verse, sticking a post-it on the reverb that gets cut in the bridge, etc.
- Trial and error. Print several passes to the master deck. Then have a listening session with the band and choose the best one.
- Cut and paste. Be prepared to edit the good pieces from several good mixes into a single best mix.
Manual mixes often become an intense process of choreographing the contortions of the various helpers grabbing knobs and faders on cue, while speed-reading the notes, scratches, and scribbles all over the studio. This can be an adrenaline-filled experience, pulling the engineer into the musical performance on the multitrack; the engineer starts to feel like part of the band. It is important that the excitement not cloud the engineer′s judgment. Sometimes these engineering thrill rides are not fun for anyone except the engineer. Sometimes the music suffers. The thrill of mastering the complicated logistics in a manual mix can mask any opinion the engineer has about the music. One must remember that those who listen to the mix later will not, for the most part, have any idea what occurred in the studio. Listeners will react to the sound of the art, not the complexity of the craft.
Automation to the rescue. Because it can control any number of faders, automation can be several sets of hands doing several mix moves at once. Because it ″remembers″ different settings and fader positions by storing them, automation can make all the crazy documentation unnecessary. Automation makes trial and error and cut and paste obsolete, turning each mix pass into a controlled, repeatable fine-tuning of the mix.
There are degrees of automation capability. In the world of digital workplaces (digital audio workstations, digital hard disk recorders with built-in mixing capability, or digital consoles), it is often possible to automate nearly every knob, switch, slider, or parameter in the mixer. This is more difficult to do with an analog work surface where it is quite likely that only the faders and cut buttons are automated. Somewhere in between is musical instrument digital interface (MIDI) automation, where simple note on/note off and other performance gestures are used to drive a mixer instead of a synthesizer or sequencer.
15.2.1 FADERS AND CUTS AUTOMATION
Some consoles only automate the faders and the cut buttons. It is possible to spend more than $200,000 and only get faders and cuts automation. That might seem disappointing at first. There is a lot more to a mix than faders moving and mute buttons cutting tracks in and out. What about pan positions, equalization (EQ), reverb time? Of course these other settings are important to a mix. But a key to successful automated mixing is being honest with oneself in assessing whether or not these settings really need complex changes throughout a mix. Keeping the mix moves simple helps keep the mind free enough and calm enough to think creatively and listen carefully. It is possible to have too much of a good thing, even in rock and roll. By the way, it is safe to say that at least 80% of all the hit pop and rock records made in the 1980s and 1990s were mixed with fader and cut automation only. Wonderfully elaborate and complicated mixes can be built with this relatively limited amount of automation capability.
The horn arranger has the creative mandate to build colors and feelings through the controlled use of different kinds of horns. Which horn plays which part of the chord? How will the chords be connected to each other? The arranger answers these sorts of questions creatively to produce a horn chart. Multitrack mixes are also arranged; they are arranged by the mix engineer. The engineer decides the sound and texture of each track, using signal processing to tweak or mangle the sound as desired. The mix engineer decides which part plays when, using the cut buttons.
The mixer/arranger often wishes to push the sonic development of the song further by changing the signal-processing structure of a given track for a special part of the song. For example, it might be desirable for the vocal to take on a less aggressive persona during the bridge. This can be accomplished through the use of a different EQ contour, less compression, more reverb, and a touch of chorus on the reverb tail — signal-processing details that were not a part of the vocal sound during the rest of the tune.
Not surprisingly, automation is the mixer′s/arranger′s tool. What perhaps is surprising, however, is the discovery that such elaborate changes to different elements of a mix can be achieved through simple faders and cuts automation. Often it is not necessary to automate the equalizer as it transitions from its primary sound to the less radical tone desired in the bridge. Nor is it necessary to automate the rest of the signal-processing components as the compression decreases, the reverb increases, and the chorused reverb appears. All that is needed is a parallel vocal channel, getting the same original vocal track, but sending it through the signal-processing chain required for the bridge. During the bridge, automated cuts simply mute the aggressively treated vocal sound that is open for the rest of the tune, and turn on the parallel, sweet, and gentle one. The action of two mute switches, turning one vocal patch off while the other is turned on, affects this significant change to the mix.
The sonic result feels like an elaborate mix move and, hopefully, a compelling musical statement. But on the console, it is created through the use of one additional channel using different effects, and a couple of mute and unmute commands of the automation system.
Automated Send
A variation on the theme above is the automated send. It may not be necessary to create an entirely different effects structure to accomplish a creative twist in the mix. For example, it might be desirable to have extra reverb on the acoustic guitar during the intro, but back off once the band kicks in. In this case, the acoustic guitar track is routed to an additional fader that sends the guitar to the reverb only, not the mix bus. Most consoles can do this. The input into the additional channel is a mult (short for ″multiple,″ the mult is a copy of the signal created simply by splitting the signal at the patch bay) of the guitar sound. The aux send patched to the reverb is turned up, but this channel′s output is specifically not assigned to the mix bus. The cut button on this extra guitar channel amounts then to a reverb on/off button; the fader is a reverb send level. Pull the fader down for less reverb; push it up for more. This automated send offers the engineer a way to layer in areas of more or less effects, again using only straightforward faders and cuts automation.
The opportunities for automated sends are limited only by the engineer′s imagination — and good taste. Add a triplet echo to key words through some automated cuts on a send to a delay. Consider the delay accent on the promising line, ″My baby′s got a new pair of ear plugs … ear plugs … ear plugs … ear plugs.″ The automated send comes from a module with a mult of the vocal, feeding a delay. It is muted the entire tune until it unmutes for the words, ″ear plugs,″ sending it to the delay, only to mute again for perhaps the rest of the tune. Alternatively, the engineer can ride a fader-sending signal only to reverb and gently add width and depth to some background vocals during the chorus. In all of these cases, the mix engineer is creating very sophisticated layers to their mix using faders and cuts automation only.
MIDI Equivalents
Depending on the extent of the desired sonic change, it may be possible to use MIDI messages sent to the effects units to achieve some mix goals. The engineer creates and saves two different reverbs on the same effects device and uses program change commands to switch to the ″verse reverb″ and back to the ″main reverb.″ These reverbs might be the same exact patch with a single parameter, like reverb time, changed.
Some of the more clever effects units allow not just program changes, but also the changing of many parameters while the effect is running. That is, using a MIDI controller, the reverb time might be shortened slightly without any audio artifacts; the reverb smoothly transitions into the smaller sound. Once again, detail and complexity are added to the character of the mix through the use of very simple commands: program changes or mod wheel motions.
Automated Comps
Ever done a session in which the lead vocal ends up scattered across several tracks? Actually, the real question is, has anyone ever done a session when it was not? Take one: sounds really good, especially the last verse and chorus — the singer really dug in there. Save that track and do another take. Take two: Everything is sounding good and the missing parts of track one are all now covered, except a couple of words drift flat. Now record those problematic words, but use a third track as punching in risks erasing a portion of the keeper part. Naturally these tracks are to be ″comped″ into a single track: Bounce the best part from each of these three source tracks onto a fourth track, feeling no pressure when punching in and out at each transition. If a punch is missed, the engineer just tries again. The original magic moments are safe on other tracks. As shown in Figure 15.1, the process has created a best, composite take from the many different options tracked.
Automate the cuts associated with creating that composite track for a more relaxed experience. Better yet, do not bounce the tracks to a new one. Using automated cuts, create a virtual comp that plays back the best take at all times, even those single words over on track 17 that were resung to correct pitch. Used wisely (i.e., not constantly), this can enable the project to enjoy a single, convincing, consistent, and powerful performance. Faders and cuts automation is all that is needed.
Printed Mix Moves
Sometimes, and this might be once every dozen mixes or so, the engineer gets really inspired and throws together a mix move too complicated for the automation system, and too magic to let go. The solution, tracks permitting, is to record the effect. In the interest of creating a distinct sound for the vocal on the last chorus, one might feel inspired to run it through an old guitar amp with tremolo, using a wah-wah pedal, sticking the amp in the shower (the water is off!). Wait, there′s more. It is too tempting to use that strange looking ribbon microphone that the tech says, ″Doesn′t work,″ but that the assistant engineer says, ″Sounds freakin′ wicked.″ Without really knowing what ″freakin′ wicked″ means, one tries it and it sounds good. While the singer plays the wah-wah pedal and the engineer pans it in time with the music, this crazy effect is printed to multitrack. Not only will it be impossible to recreate the amp tone, microphone placement, etc. of this sound tomorrow, there is the possibility that the microphone will quit working altogether. This mix move is not saved in automation, it is saved to multitrack. Simple automated mute commands will bring it into the mix anytime.
Figure 15.1 Compositing a vocal from three takes.
Of course, when the audio or the mixer live in the digital world, automation as a technology becomes a lot like a word processor or video game. If software controls the music, there is almost no limit to what the automation system can do. Digital- or computer-based mixers offer this sort of opportunity. The alternative signal path explored in faders and cuts automation is still useful, but it lives in software not hardware. To have an alternative vocal sound, just program the changes in: automate the equalizer, compressor, reverb, and chorus. There is no need for an entirely different signal path for the vocal to have this altered state, just a different kind of signal processing. In addition, the automated sends discussed above be-come trivial to set up if the echo send level control itself can be automated. That is, one need not dedicate an entire channel to automating an effect. For more ambience on the intro, ride the automated echo send level up on the intro. Nice and easy. Everything automation is a real ally to the engineer′s creativity. The mix engineer can do more sonically because the automation system can do more. Automate the EQ or the compressor for de-essing. Automate a mid-frequency sweep for wah-wah. The imagination is the limit.
The discussion of automation approaches for faders and cuts automation above reveals that it sometimes takes an elaborate signal flow structure just to create a small effect. The engineer often has trouble feeling inspired if every sonic idea (″Wouldn′t it sound great if …″) is followed by an analysis of the automation manual (″How the heck is that automated?″). With everything automation, sound engineers do not have to restrain their creative impulses to enhance or manipulate a sound. The mix moves can be programmed almost as soon as the engineer thinks of the type of sound effect wanted.
The possibilities are endless, but the risk is that it′s too much of a good thing. Too many options can paralyze the mixing engineer: ″Let′s see … I′ll push the lead vocal fader up on the word, ′baby,′ roll-off a bit of high end on the hi-hat, cut the snare track between each hit, pan the piano and guitar back and forth in triplet time, add a nasally EQ to the bass, and shorten the delay on the digeridoo for the first two bars of the first verse. Then, I′ll …″
Clearly the music could suffer. There is a real temptation to explore mix moves because the gear can do it, not because the music needs it. With everything automation, the recording studio has lost a sort of ″reality check″ that the more limited automation system imposed. It takes restraint and maturity. Most engineers are at least occasionally seduced by the equipment in this way, no matter how committed they are to the music. The equipment will take over or just plain interfere with the process if left unchecked. The trick is to know when to explore the automation possibilities and when to take a step back and let the music be.
Everything automation is made less intimidating through what is known as snapshot automation. While it may be musically desirable to have the bridge of the tune live in another sonic world (e.g., different fader, pan, EQ, and effects settings), often it is not necessary to continuously modify all the controls with subtle shifts from one setting to the other. Snapshot automation can store the configuration of every parameter of the console at one instant (snapshot A), and then store all the settings wanted for the bridge of the tune (snapshot B). The snapshot automation system enables the mixer toggle from one setting to the other, smoothly, all at once. The mix engineer does not have to program the changes in fader by fader, pan pot by pan pot, effects parameter by effects parameter, and so on. Just store the best sounding chorus setup and the best sounding bridge setup, and let the snapshot automation system connect the dots.
The clever snapshot system gives the user some control over how the change from one snapshot to another is made. Adjustable cross-fade or morph times, perhaps with adjustable slopes, can make the snapping process itself more musical. A very complicated set of mix moves accomplished with very little tedious automation programming results. This often more than adequately serves the musical needs of the mix.
While there are as many different automation systems as there are brands of consoles and workstations, they have enough in common that the reader can be prepared to use any automation system through the following orientation.
Either the engineer is performing the mix moves, or the automation system is. Terms like ″write″ or ″read″ are common ways of making this rather important distinction. Think of it as automation record and automation play. When a mix move has been designed and rehearsed, the engineer enters automation write mode and records the move in. Automation read mode makes the automation system reproduce the mix move.
Write and read can usually, and sometimes dangerously, be done globally across the entire console or workstation. Alternatively, write and read may be applied in a more focused way, a few channels at a time, fader by fader, cut button by cut button.
Like any piece of software, there are some quirks that seem strange at first but become more natural with experience. For example, some automation systems need you to write at least a rough version of the whole mix for all faders and mutes on the entire console before one is allowed to go back and do those smaller tweaks on a single fader within a small part of the tune. Sometimes something as trivial as the start time of the mix is an unmovable anchor for the automation system. Getting expressive ideas into a computer is never easy. The artist′s approach to creating computer graphics is at least a little different from how they draw on paper. Similarly, entering a mix into the computer requires some navigation through menus and mastering of peculiar syntax that slightly alters the routine from just manual mixing. Once these quirks are mastered, one writes and reads mix automation moves as naturally as one records a track, rewinds and does another pass, plays it back, moves onto another track, and so on. Expect at least a short learning curve.
15.3.2 WRITE: ABSOLUTE OR RELATIVE
When recording mix moves, there are usually two broad approaches, often described by words like absolute mode and relative mode. Absolute mode tells the system that the engineer wants the automation to store the exact mix parameters currently being adjusted. Any previously written automation moves at this point of the song are completely forgotten, erased and replaced by the new mix moves.
Relative (a.k.a. update or trim) mode lets the engineer revise an existing mix through nudging and tweaking. If the mix engineer likes where the vocal is sitting in the mix, but wishes to push a couple of key words up a little louder, a relative mode might be a good approach. Entering write mode in this way does not make the automation system disregard the mix information already recorded; rather, it updates it based on the new relative moves. Push the fader up and it adds that increase to the move already there. These relative mode trims can be accumulated through additional automation passes to create a very complicated set of fader rides in a simple and intuitive way.
Using Figure 15.2, consider this pretty typical scenario. The first pass sets the general level of the vocal in the mix. Absolute mode writes this into the computer as a starting point. The second pass might address some overall arrangement issues: Turn the vocal up from this basic level at every chorus, and have it be a little lower at the bridge. Relative mode would be a good way to do this. With all other faders safely in read mode, enter write mode on the vocal with the fader in any position, push it up in the choruses, pull it back down to the original starting point in the verses, and lower it in the bridge. Because the mix is being trimmed in relative mode, the automation system looks only at the changes in the fader′s position, not the actual, absolute position of the fader itself. On the third pass the engineer might then modify those already changing fader settings on a word-by-word basis, reducing the words that poke out of the mix and raising the words that are getting lost. Relative mode would be the appropriate choice again. As the mix plays, there is no need to find and match fader levels already written, just write automation moves in relative mode and make these quick fader rides to revise and refine the evolving fader moves already stored in automation.
Figure 15.2 Fader automation in four easy steps.
15.4 Automation Strategies: Organizing the infinite Options
It is essential to bring some amount of order to the mix approach. The ability to do everything makes it difficult to do anything. Here is a common way to break it down into bite-size increments that grow into an elaborate, musical mix.
Consider the various stages of building a mix. First, one lays out the rough balances that start to make musical sense. Listening to the entire song, one finds fader positions and pan pot settings that enable the song to stand on its own. As discussed in Chapter 8, the pop music vocal and the snare sit pretty loud in the mix, dead center. Kick and bass are also in the center, likely at a lower level. The other pieces of the arrangement fill in underneath and around these critical tracks. Guitars and keys must not mask the vocals or harm the intelligibility of the lyrics. The musical role of each and every track must be understood, and the mix balanced accordingly, so that elements playing together blend or achieve counterpoint as needed. Work hard to find a balance that is fun to listen to, supporting the music while revealing the complexity and subtlety of the song. Importantly, the engineer does not yet automate the mix.
15.4.2 PHASE II: SIGNAL PROCESSING
With the fine tuning that comes from the addition of various forms of signal processing, the engineer builds up a mix that tastefully highlights every element of the pop arrangement that needs it, while achieving subtle blending and layering for the behind-the-scenes, supporting tracks. Various forms of distortion, EQ, compression, limiting, expansion, gating, tremolo, flanging, chorus, echo, pitch shift, reverb, and any number of other effects are introduced, tested, rejected, adjusted, and refined. The sonic traits that will ultimately define this mix are developed now. Every effect introduction requires the engineer to check the balance again. Effects on the snare likely affect its apparent loudness. Effects on the snare are likely to influence its relative balance versus the guitars and the vocal. The engineer constantly and iteratively works to keep the multitrack arrangement balanced as effects are added and deleted.
Please keep in mind that, so far, the whole mix is static. That is, the mix engineer has got a pretty decent sounding tune coming out of the loudspeakers without doing any clever fader moves and effects twiddles. These two steps (phases I and II) actually do the most to determine the overall sound of the mix. The engineer works hardest here, and still has not entered automation yet. As they are so important, these first two phases should take the most time and consume the most creative energy. In fact, because automation is so darn fun, this is rarely the case.
With the mix very near where it needs to be sonically (i.e., the producer, engineer, vocalist, and the rest of the band pretty much approve of the sound), the engineer will at last begin to automate the mix. The first step is to apply the appropriate cuts to channels that either are not being used or that were decided against for performance or arrangement reasons. This amounts to making the mix arrangement official.
If the producer does not want horns in the first chorus, the engineer automates the mutes on all of the horn tracks accordingly. If, in the engineer′s judgment, the bridge sounds better using doubled harmony vocals, the appropriate mutes and unmutes are automated. If the singer likes the second chorus from the third take of the lead vocal, the engineer ″comps″ it in using cuts automation.
Now with these cuts happening automatically on cue, the producer and engineer can listen carefully to how the song feels to make sure those are the right decisions. It comes as a surprise at first, but just the process of diving for the cut buttons to mute and unmute tracks at the appropriate times is enough to interfere with one′s hearing, psychologically. That is, it is hard for the engineer to form a certain opinion of the mix idea while also remembering all these manual moves. It is much more effective to plug the cuts into automation, sit back, and listen to the song unburdened by any other activities or distractions. Hands folded, eyes closed, the engineer can join the producer and the artist as they decide how much they like the production so far. Does the arrangement make musical sense? Does it grow musically? Does it sag and feel empty in the bridge? Is the detail of the snare lost when the doubled electric guitars enter each chorus? Is the piano in tune with the bass? All these important questions can best be answered when one is just listening and not pressing buttons.
To complete the multitrack arrangement of the song, the cuts described above are followed by some general fader rides. This is where the engineer does things like push the vocal up in the choruses, pull the piano down during the guitar solo, and such. Generally, these are pretty subtle rides. These fader moves are aimed at the musical interpretation of the mix, trying to make the song feel right, whatever that means. A little ride here and another one there helps shape the energy level and mood of the mix.
15.4.5 PHASE V: TWEAKS AND SPECIAL EFFECTS
Only after the musically compelling and well-organized automated mix built carefully through the four phases described above is complete, should the mix engineer attempt the more elaborate, the silly, and the downright crazy moves that are now so tempting. At this point the engineer can add quarter-note delays to keywords of the lead vocal. Now is the time for the mixer to bump up the lost notes of the solo. At last, the engineer can experiment freely with some more elaborate effects to dress up the bridge. With the fundamental elements of the mix being faithfully replayed by the automation, the mix engineer′s mind is free to explore the complicated stuff: ″In the last chorus, let′s run the returns of the long piano reverb through a distortion pedal. Then run that through a noise gate being keyed open by a swing eighth-note delayed snare hit. Then, if we have any more patch cables, it would sound cool if we …″ Anything goes at this phase of the automated mix.
It is not enough, after all this, just to know the theory of how to operate the automation system. The mix engineer has to develop some performance ability on the automated mixing console.
Truly musical mixes come only after the engineer is comfortable with the automated mixer, mentally and physically. A goal of automation is to make the console more like a musical instrument — a device on which all good engineers can perform. It follows then that engineers need to practice. Like practicing scales, the mix engineer needs to have down a set of typical moves and be able to do them quickly, under pressure, without thinking. The ″scales″ of mixing to be practiced are things like the quick cuts associated with comping, appropriate level rides on a vocal track or horn solo through a performance, musical fades for the end of a tune, quick fader moves to attenuate an unwanted squeak or highlight a subtle phrase, setting up an automated send, etc.
Building on this set of often used basic moves, the engineer also develops the techniques needed to attempt more unusual moves. Automation, just like piano, requires practice, practice, practice. It may sound a little silly, but it is also recommended that engineers find some calm time to simply noodle around with the automation system. Think of practicing a musical instrument. Scales, arpeggios, etudes, and prepared pieces are part of that discipline. But all musicians depart, during practice sessions, and just jam. Engineers should do the equivalent on the mixer. Experiment with tracks when the client is not there. Design and explore elaborate automation moves just to see if it can be pulled off. Practicing with the mixing console in this way does more than make the engineer a better mixer. The engineer also develops the ability to work quickly and to improvise on the mixing console. Musicians and producers notice this ability in an engineer.
The techniques developed for mixing are intimately related to the type of mixer used, be it an analog work surface, a digital console, a digital audio workstation, or some combination thereof. Some interfaces welcome a pretty natural approach, while others require special techniques. To get an understanding of how to automate all controls (faders, pan pots, aux sends, equalizers, compressors, and so on) it is helpful first to look closely at faders alone. They serve as an excellent example that provides insight into automating the rest of the console.
The faders on an automated console take instructions from two possible sources: the engineer or the automation system. When the automation system is in control, one of two rather unusual things happens. One possibility is that the faders start moving on their own, without being touched. This can be rather creepy, especially when one is all alone with the console, late at night, lacking sleep, overloading on caffeine. Perhaps more bizarre is when the mix is played back by the automation system, with all of those carefully engineered fader rides happening, yet the faders are not moving. It is difficult to say which ghost is preferred within the machine.
Moving Faders
The performance techniques — the physical, dexterity part — are slightly different based on this sole criterion: Do the faders move? Consider moving fader automation first, as it is a little more intuitive. When the console has the ability to move the faders, there is less chance for confusion. Moving faders offer the engineer that much-desired feature of WYSIWYG — what you see is what you get. That is, one can see the mix moves during playback. Guitars too loud in the bridge? One glance at the guitar faders reminds the engineer — oops, I forgot to pull them down.
Not only is there excellent visual feedback about the mix, there is also a nice physical feature. Most moving fader systems will automatically leave read mode and enter write mode whenever a fader is touched. Moreover, they can return to read mode as soon as the engineer lets go of the fader. This gives the engineer a real opportunity to perform. Could not hear that word? Wind the mix back to a point a couple of bars before the word. Play the mix and simply nudge the fader up on that word. When the engineer touches the fader and pushes it up, the automation system instantly starts recording the mix move. When the engineer lets go, the fader returns naturally to the old level and the automated mix playback resumes. It could not be more intuitive. Automation systems without moving faders can achieve the same mix move, but without motorized faders it is less intuitive how the change gets programmed in.
The ease of use as well as the immediate visual, aural, and physical gratification that moving fader systems offer makes them the most desirable way to work. Most world-class studios have moving faders. All digital audio workstations offer a click and drag equivalent. And, thankfully, many less expensive automation systems have moving faders too. But moving fader systems are not without drawbacks. Perhaps most obvious is cost. Consoles are already expensive devices full of seemingly countless components. Putting a reliable, accurate, consistent, small, and quiet motor on every fader is not cheap. If they are cheap, then one has to worry about their reliability, accuracy, consistency, size, and noise. Those world-class studios build this into their big-ticket studio fees. And they probably have spare motors in the tech room. The solution is to have high revenue that justifies this, or to shop very carefully. Such a financial commitment to automation is not always the best choice when it might be more productive to buy another compressor rather than maintain a set of servomotors. There is an additional potential drawback to moving faders, and this one is a little scary. If the automation system can move the faders when the engineer tells it to, then it might also be possible for the automation to move the faders when the engineer did not (mean to) tell it to.
Trade shows and studio tours often show-off demos of the moving fader systems on large-format consoles. In the demo, the faders move together, not for a musical mix, but for some sort of visual effect — sort of a Radio City Rockettes sort of entertainment. The faders arrange themselves into a sine wave that moves left to right and then right to left across the console. First slow, then fast. Those are nifty demonstrations. But imagine spending a not unreasonable four or more hours finding the balance for a tune (phase I) and, just before entering automation, someone accidentally hits the ″demo″ button. Buzz, click, hum. The faders start doing the wave like they are at the Super Bowl. A careful balance is destroyed by the pre-programmed dance of the faders. It sounds silly, but it really happens. Worse, it is not just the demo that wipes away hours of hard work.
Imagine the following situation: The engineer has found a decent static balance for the tune when, on accident, the multitrack winds back too far, into the previous song. If the engineer is not careful, locating to the previous song can cause the faders to race to the appropriate levels from that mix. The automation has seen that time code address before; it knows what to do. The balance for this tune is gone, baby, gone. The solution is to know how to turn the motors off. If fader motors are disabled until after the careful balance is safely stored in the automation memory, then those important fader levels will not be lost. Even after doing this, your author has to admit, he still feels a little rush of terror every time he turns the motors on.
VCA Faders
If the faders cannot be moved by the automation system, how can they be automated? Good question. The solution relies on a voltage-controlled amplifier (VCA) (see Chapter 8). VCAs are faders whose amplifiers boost or attenuate a signal based not on the position of a fader control, but on the value of a control voltage. The result is that the engineer and the automation system share control of the fader level. The engineer uses the slider on the console to adjust the control voltage, while the automation computer uses software. Either control voltage then determines the fader level, the engineer′s or the automation system′s. But it is tricky mixing when the faders are not moving. What one sees definitely is not what one gets.
No problem. With some adjustment to technique and some practice, most engineers find VCA automation perfectly easy to use. As with true audio faders, the first automated pass is done in an absolute mode. All subsequent passes, however, are generally done in a relative mode. A typical series of mix automation passes is shown in Figure 15.2. Once the static level is written in absolute mode, additional adjustments are made with the VCA in relative mode.
Broadly, two approaches to VCA automation have evolved. The first approach is to return all faders to their zero position after the first pass. This way, the engineer can cut or boost the level, and the scale next to the fader quantifies just how far it has been moved. If the fader is at -12 dB, the engineer has reduced the level 12 decibels from the last stored fader level. There is little confusion as to how extreme the fader move on any automation pass has become, and one can always return to the level of the last pass by returning to the 0 dB marking. This is a good way to keep track of all changes within a pass on a nonmoving fader system.
On the down side, however, it does remove entirely the visual representation of all the fader levels relative to each other. That is, if all faders are set to zero after the first automation pass, then the careful, overall balance built before automation — where the relative levels of each and every track in the multitrack production were thoughtfully and iteratively coaxed into position — lives in the computer′s memory only, and the engineer cannot see it. Some engineers find this discomforting. Another option is to do even the later relative mode passes with the faders left where they originally were in the rough balance. That rough balance is usually the level for most of the faders most of the time. Leave the faders at that position for the start of any relative trims and then the actual, physical position of the fader will stay at or very near the fader level that is programmed into the automation. It is a little more confusing to enter write mode at -8 dB, nudge it up briefly to -5 dB, and then remember to return to -8 dB before finishing the revision. With practice, it becomes more natural.
When the faders live on a computer screen, they can move without motors. So much of the benefit of moving fader automation is preserved in digital audio workstations without the expense and maintenance disadvantages. Perhaps the biggest frustration is that, without an external hardware controller with motorized faders, one has to click the fader, not touch the fader. For old-school engineers who were weaned on large-format analog consoles, clicking on faders is an acquired taste. For anyone who is reasonably computer savvy, this is not a problem. Power-users, who are quick on the word processor, efficient surfing the Internet, and agile playing computer games, probably have enough mouse dexterity to find mixing by clicking on a screen perfectly comfortable.
In the software domain, all the other controls on the computer screen can usually be automated as well. The approach is the same, so an engineer′s knowledge of write/read and absolute/relative modes is easily applied to pan pots, echo sends, reverb parameters, etc.
Yet another automation feature appears when one can interact with it on screen: graphic editing of mix parameters. There are times when one might not need to perform an additional mix pass just to fix a mix problem. Wish the trombone solo were just a little louder? Click on the graphical re-presentation of the level and drag it up a bit. Listen to it to confirm it is right. Accidentally uncut the track containing the organ solo a hair late and clipped off a bit of the first note? No need to repeat passes and hope through trial and error to eventually time the uncut move at the right instant. Just type in a slightly earlier unmute time in the computer to trim it until it sounds right. These screens, menus, lists, and pictures are often a powerful way to program a mix.
Mix automation gives engineers the power to achieve more. Sloppily handled, that power turns on the engineer, leading almost certainly to lower-quality mixes.
15.6.1 MASTER THE GEAR BEFORE IT MASTERS YOU
Most engineers do this sort of work because they love music. The focus during mixdown is to make the most of the music already recorded, using as much gear as it takes and employing every trick ever seen or heard. It is not easy work. The proverbial magic dust that gets sprinkled on a single vocal track might realistically include two stages of compression; a ″secret recipe″ of EQ; the subtle addition of rhythmic delays to create a pulsing, highly customized reverb; and fader rides that do not just change from verse to chorus, but word to word, and sometimes syllable by syllable. It is easy to spend several hours of mix time on the vocal alone. The quantity and quality of the signal processors used — on just the vocal — might take a year′s salary to acquire. It is not easy to dive that deep into the detail of the vocal sound and still keep track of the musicality of the mix. Inexperienced engineers find it frustrating when the bass player walks in while they are working on the vocal. Likely, the bass player is focused on the bass. The engineer has been focusing on the vocal. The unfortunate engineer ″glances″ at the bass sound to find a twangy bit of sonic meatloaf, overpowered by every other track in the mix, including the shaker.
Mixing requires an engineer to see the forest and the trees. The gear pulls all engineers into a microscopic level of resolution. The music asks them to zoom back out and listen to the whole. Experienced mixers can handle this sort of conflict because they have exercised themselves in this discipline and because they know the gear backwards and forwards. New and intermediate engineers should endeavor to do the same. Recognize that the automation system itself is a very seductive amount of signal-processing capability, and it is often so elaborate as to be at least a little intimidating. Overcome this through overpreparation. Great engineers become so comfortable with the syntax, techniques, tricks, and limitations of the automation system that they can freely move between big-picture musical issues and highly focused automation moves. They can listen to the bass, and the snare, and the kazoo, even as they refine the vocal.
Avoid the common trap of entering automation too soon. This is tricky, and it takes most engineers a number of painful experiences to get over this temptation. Automation employed carefully (and sometimes that means sparingly) is a lifesaver. When misdirected, automation will fall somewhere on a spectrum from distracting to crippling. That is, at best one is faced with the need to constantly undo or redo moves as the mix unfolds. At worst, the automation system so takes over and interferes with the engineer′s ability to make changes that, musically, the mix unravels. The solution is to activate the automation system as late as possible in the course of the mixdown session. It is easier to explore mix ideas when the automation does not have to be rewritten. For example, the producer might like to hear the background vocals panned left, without the doubled tracks, but with the addition of some heavily flanged reverb panned right. Trying this out is no problem preautomation. The engineer just turns the knobs and pushes the faders until it sounds right. If the mix is already automated, the engineer will have to revise the automation data governing the fader levels, pan positions, and cut buttons on every track involved. That will slow the session down. It may take so long to get the idea sounding good that everyone gives up on it. A good mix idea is suppressed by the interference of the automation system.
Only the most complicated mix ideas need automation. Explore those late in the game after the basic mix arrangement is automated. The creative process associated with most mix ideas will be much more successful if they are explored without automation.
The performance gestures that make a mix soar live in software. They must be treated with the same level of paranoia that other software documents inspire: save often, make safety backups, and be prepared for frustration and irretrievable loss when working during a thunderstorm with its associated power spikes and flickers.
Through organization, study, and practice, automation can become an easy-to-use asset in the studio. It does not make manual mixing easier or faster. On the contrary, automation enables mixes to become much more elaborate. Mix sessions can sometimes be much, much slower.
Automation does not make everyone a better mix engineer. If the manual mixes were pretty dull, the automated mixes will be too. It is only in the hands of a talented engineer that automation makes mixes more interesting and sophisticated. Automation makes the rather uninteresting device known as a mixing console into more of a musical instrument. It stores the engineer′s performance. It frees them to improvise. Many engineers do not think of automation as a helpful feature, they think of it as a required tool that empowers them to orchestrate all of the effects associated with making recorded music.