″Once upon a time, I was an ocean.
But now I′m a mountain range.
Something unstoppable set into motion.
Nothing is different, but everything′s changed.″
— ″ONCE UPON A TIME THERE WAS AN OCEAN,″ PAUL SIMON, SURPRISE
(WARNER BROTHERS RECORDS, 2006)
For the most part, the signal-processing effects discussed in this book are not available to the music-buying public. These devices are for the professional studio, not the home stereo. Consumers are, however, allowed one of the most important signal-processing devices of all: the volume control. Sound engineers must not overlook the significance of this signal processor. This chapter reviews the important uses of some of the simplest signal-processing devices dedicated to volume: the fader, pan pot, and mute button.
The mother of all volume controls is, of course, the fader. Slide it up for more level and down for less volume. Pan pots coordinate two level settings to perceptually place the stereophonic image of a single audio track at or between the two loudspeakers. The mute button expresses level in binary: on or off, level or none.
Three types of faders are found in the recording studio: the variable resistor, voltage-controlled amplifier, and digital fader.
Variable Resistor
Electrical resistance is a property of all materials describing how much they restrict the flow of electricity. A high resistance does more to hinder the flow of electrical current than a low resistance. Materials with very high resistance are classified as insulators. Under normal operating conditions, they do not conduct much electricity at all. We appreciate this property when we handle things like power cords.
At the other extreme, devices with very low resistance fall into the category of conductor. Copper wire is a convenient example. The copper within that power cable conducts electricity from the wall outlet to the piece of audio equipment, enabling it to process audio signals, getting the light-emitting diodes (LEDs) to flicker, motivating the meters to twitch, and enabling us to make and record music.
In audio, special use is made of devices whose resistance is adjustable, from low to high. The volume knob on a home stereo, electric guitar, or analog synthesizer is (with a few model-specific exceptions) a variable resistor (Figure 8.1).Set to a high resistance, electricity has trouble flowing and the volume is attenuated. To turn up the volume, lower the resistance and let the audio waveform through more easily. Lowering the resistance raises the volume. These variable resistors define their own class volume control.
Electrical voltage represents the potential for electrical current to flow, depending on the impedance of the circuit that is made between the two points of contact on the voltage source. The potential is, to those who design and manufacture electrical devices, synonymous with voltage. Accordingly, variable resisters are also called potentiometers.
This same device gets yet another name. In order to specifically distinguish it from the voltage-controlled amplifier discussed next, variable resistors used in audio are sometimes simply called audio faders, or, more emphatically, true audio faders. This term indicates that the adjustable resistance of this fader is part of the very circuit that contains the audio. As the discussion of voltage-controlled amplifiers makes clear, this is not always the case.
Voltage-Controlled Amplifier
In the recording studio, it is necessary to look more closely at analog volume controls because there is a second type: the voltage-controlled amplifier (VCA). The idea behind them is simple and clever.
Be forewarned, it risks getting a little philosophical here. For audio circuits using a VCA, the fader that sits on the mixing console is separated from the audio itself by one layer. Instead of having that slider on the console physically adjust the resistance of a component of an audio circuit, it adjusts a control voltage instead. This control voltage in turn adjusts the amount of gain on an amplifier(Figure 8.2). That is, instead of using the fader to directly adjust the amplitude of the audio, the engineer uses the fader to adjust a thing that adjusts the amplitude of the audio. It may seem a tedious distinction at first.
Voltage-controlled amplifiers boost or attenuate an audio signal (that is the amplifier part) based not on the position of a fader control, but on the value of a control voltage (that is the voltage-controlled part). The result is that an engineer, machine, mix automation system, or compressor can control the level of the signal through a VCA. Without VCAs, the only way to have something other than the engineer adjust the level is to stick a motor on the fader. Motorized faders are certainly available, but they are a pricey, complicated option. But before motorized faders were feasible, machines controlled signal amplitude through VCAs. A machine can output a control voltage as easily as it can output an audio voltage. By using VCAs, faders can be manipulated by man or machine. This makes possible a range of effects, notably mix automation (see Chapter 15),compression (see Chapter 6), and expansion (see Chapter 7).
Figure 8.2 A voltage-controlled amplifier.
Sitting in front of any analog mixing console for the first time, an engineer is wise to determine (from the manual or the house engineer) whether the faders are VCAs or variable resistors. Variable resistors are comparatively easier to design and manufacture at a low price, with high audio quality. Voltage-controlled amplifiers, on the other hand, are a more sophisticated device, containing active electronics. Noise and distortion rear their ugly heads. A quiet, distortionless VCA is not a given. High-end manufacturers of analog consoles today can be counted on to offer sonically-transparent VCAs. With older devices (consoles from the 1980s and before), and cheaper devices, one needs to listen carefully.
Pay particular attention to low-level details in the signal, such as the imaging of a stereo pair of tracks or the envelopment of a lush reverb tail. Inferior VCAs can collapse stereo images ever so slightly, and they are particularly destructive to stereo reverb, where the reverberation narrows, shortens, and becomes less believable.
Digital Fader
Digital audio offers a third type of volume control. In digital audio, where the audio signal is represented by a long series of numbers, a fader is a multiplier. Multiply by a number less than one to reduce level. Multiply by a number greater than one to increase level. Provided the digital system is functioning properly, has headroom available to accommodate the mathematical result, and dithers the signal properly, it is fair to think of digital volume controls as a third, simpler, kind of level adjuster. The sound quality of a digital fader is driven by the type of math (fixed point or floating point) and bit resolution (storage formats are commonly 16, 24, and sometime 32 bit, but processing bit resolutions often run at much higher resolutions, using 48 bits or more for the math associated with faders, pan pots, signal processors, and mixers).
The action of the pan pot is so natural and intuitive that one might forget that it is a level control device. In fact, it is really two volume controls. A single user interface — a knob or a slider — adjusts the relative level of two outputs (Figure 8.3). The balance knob is similar, but is a stereo-in, stereo-out device, adjusting the relative volumes of the left and right side, but never mixing the two.
Figure 8.3 The pan pot is in a one-in, two-out device, while the balance control is in a two-in, two-out device.
The user, in a single pan pot gesture, turns up one output while turning down the other. Two variable resistors are adjusted by a single control. The variable resistors are wired with opposite polarity. A clockwise motion lowers the resistance of the right output while it increases the resistance of the left output by the corresponding amount. The result, of course, is the perception of sound coming from the direction of the louder speaker. Pan to make a track appear at one speaker, the other speaker, or some point in between.
Autopanners bring the panning capability into an effects device. A single input is fed to a voltage-controlled pan pot, which can be programmed to move in some desired patterns: left to right, right to left, left and right repeatedly, etc. The speed, depth, and shape of the panning are often controllable as well (much like the modulation section of a delay; see Chapter 9). In this way, a machine faithfully repeats a panning pattern, allowing the engineer to set it up once and then focus on other matters.
To mute a signal, a simple switch purposefully interrupts the flow! of the signal. Using relays or transistors, switches may be controlled or triggered by a machine as well as the engineer.
These humble level control devices are profoundly important production tools.
Consider the first step in building a mix. One carefully, systematically, and iteratively adjusts and readjusts the volume of each and every track until the combination starts to make musical sense. When the recording engineer has fine-tuned the level and panning of the core elements of the multitrack arrangement, thoughtfully muting those tracks that undermine the quality of the production, the mix is said to be ″balanced.″
In the course of overdubbing or mixing a tune, a single song may be played more than a hundred times. Artists will take advantage of multiple passes to refine and hopefully improve their recorded performance. Mixdown is such a complicated process that, in most pop productions, it takes from several hours to several days to mix a 3½-minute song.
In the course of just that first playback of the piece, however, the engineer must find the fader levels and pan pot positions that enable the song to stand on its own. The goal is to empower each and every track to make its contribution to the overall music without obliterating other parts of the multitrack production. Flawed tracks are muted. Tracks that are in conflict with each other are carefully evaluated and possibly muted. Balancing a mix is a fundamental skill all engineers must possess.
While there is no single right answer, it would be fair to say that in a typical pop mix, the vocal and the snare sit pretty loud in the mix, dead center. Sharing the center, but at lower levels, are the kick drum and the bass guitar.
These four elements — vocal, snare, kick, and bass — are essential ingredients for almost every pop mix. They should always be independently audible, never masking each other (see Chapter 3). The rest of the tracks sit below and beside, never masking these four tracks.
The vocal must reign supreme. Beware of other tracks with strong spectral presence in the middle frequencies. Spectral masking may make it difficult to hear the consonants of the vocals. The words then become difficult to understand. Beware of other instruments playing in the same range as the vocal. A piano, guitar, sax, or cello can render a vocal inaudible if the parts are too similar. Solo sections and instrumental tunes often have a featured solo instrument in lieu of a lead vocal. Give that solo instrument the same priority a lead vocal would get. Find fader positions for the tracks that keep midrange harmony instruments out of the way enough so that the listeners can understand and enjoy the phrasing of the singer or the soloist. Pan the various nonvocal tracks competing in the middle frequencies off to opposite sides, out of the way of the lead vocal and each other. Instruments with similar pitch and spectral content will likely blur together and cloud the enjoyment of each individual instrument. Pan them to different locations for an immediate improvement in clarity and multitrack independence.
While the vocal is typically the most important single track in the entire production, it is not unusual for the snare drum to rival it in loudness. In fact, the snare might even be a bit louder than the lead vocal in some songs. Because the snare is so short in duration, the temporary masking it causes is sometimes (but certainly not always) acceptable. There is a natural tendency to push the snare fader up. A snare that sounds great while soloed may sound disappointing when placed back into the full context of the mix. Keys and guitars will fight the snare in the midrange. Cymbals and distortion (see Chapter 4) can create unwanted masking in the higher frequencies. Bass can rob the snare of low-frequency punch and power. But do not turn the snare up too much; if too loud, the snare loses musical value, becoming a distraction rather than a source of rhythm and energy. The snare drum gets constant attention. Set the fader so it is always an enjoyable part of the multitrack production.
The kick drum and bass guitar may have a low-frequency tug-of-war. Set their levels very carefully. A song with a lot of bass can be thrilling on first listen, but if the bass guitar is too loud, the kick drum gets masked. Choose instruments, tunings, playing styles, musical parts, and recording techniques that aim for at least slightly different frequency ranges, kick versus bass. Use signal processing at mixdown to ensure the result. Typically, the deep tone of the kick drum is lower in frequency than any low emphasis of the electric bass, but there are exceptions. A small kick drum, tuned to an upper low-frequency heartbeat in a production using a five-string bass with the low string tuned down to the B slightly more than three octaves below middle C (and therefore having a fundamental frequency of about 31 Hz) can turn things around. In cases such as these, it might make sense to place the bass guitar below the kick drum, spectrally. In either case, these two instruments must be shown how to get along together. There is a natural tendency for them to fight. The fader positions for these two instruments are given deliberate attention. Raise their levels to points that are clearly too loud and then pull them back. Lower their level until they are too soft, and then push them back up. Somewhere between these two positions lives the most effective musical level. Engineers constantly message the levels to both teach themselves the dangers of losing control of these tracks and to make sure their levels make the most musical sense. The mix achieves new levels of refined clarity when the kick and bass cooperate.
The other parts of the multitrack arrangement (both tracks and effects) fill in around and underneath these most important tracks. If the guitar is louder than the vocals, the band is probably going to sell fewer disks and downloads. If the music fans can not hear the piano when the sax plays, the song loses musical impact (and the engineer probably loses the chance to hire that piano player on the next album). Work the faders hard to find a balance that is fun to listen to, supports the music, and reveals all the complexity and subtlety of the song.
Balancing a mix sounds so straightforward, in concept. Engineers who are early in their career will soon discover that it is not easy. Keeping a mix balanced requires experience, excellent listening skills, patience, focus, and a strong musical opinion. On every session of every project, the balance of the mix directly drives everyone′s opinion of the quality of the project. It also influences their ability to get their job done. In order to play, they have to be able to hear. The engineer is on the spot, expected to balance the tune and keep it balanced at all times.
Yet on any given session, there are other points of focus. The drummer needs attention during basics. The singer is everything during vocal overdubs. The compressors and reverbs are complex devices that can distract an engineer during mixdown. No matter what other session priorities are present, no matter what other difficulties and distractions arise, the engineer is always required to keep the production balanced. During a guitar overdub, the guitar might be allowed to be overly emphasized in the mix. As soon as the session advances from guitar to piano overdubs, however, that guitar must be turned down and tucked into the mix so that the vocal, snare, kick, bass, and all other supporting elements can be enjoyed.
This first step of a mix session is really a part of every session. For tracking and overdubbing, the players can not play, the engineer can not hear, and the producer can not produce until the signals from all the live microphones, all the tracks previously recorded, and all the effects are brought into balance. Relying almost entirely on volume controls, balancing a mix is one of the most important skills an engineer must master.
Every track of the multitrack production begins its life at that critical moment when it is recorded to tape or disk. If a microphone captures the music, a microphone preamp is needed. Of course, microphone preamplifiers are nothing more than volume devices. An engineer has got to set the microphone preamp level just right when recording to tape or hard disk.
Because of the noise floor of all equipment, it is desirable to try to record music at the highest level allowed so that the musical waveforms mask the noise floor. So it seems true that louder is indeed better. The question is how loud. The strategy for recording levels divides into two types, depending on whether the storage format is digital or analog.
Most engineers have heard that, for digital recording (whether to tape or hard disk), the goal is to ″print the signal as hot as possible, without going over.″ This deserves a closer look.
Digital recording systems convert the amplitude of the music signal into a string of numbers. Pressure in the air becomes voltage on a wire (thanks to the microphone), which then becomes numbers on tape or disk (thanks to the analog-to-digital converter). When the digital audio recorder has reached the largest number that it can handle, it maxes out. As the air pressure of the music increases in the air, the corresponding voltage on the microphone cable increases too. When the input into the analog-to-digital converter is too high, the numbers to be stored by the digital system can not get any bigger. When the signal gets too hot, a digital audio system runs out of the ability to keep up with that signal, sort of like a child not being able to keep counting once they run out of fingers. The child just stops at 10 fingers. The 16-bit system stops at: 1111 1111 1111 1111. The 24-bit system stops at: 1111 1111 1111 1111 1111 1111. No higher value can be stored. When the amplitude of the input forces the digital system to ″run out of fingers,″ the digital data no longer follows the musical waveform (Figure 8.4). This is a kind of distortion known as hard clipping, discussed in Chapter 4. Play back the numbers, and the digital recorder can not recreate the original music. The peaks are clipped off, gone forever. Harmonic distortion favoring odd harmonics is the unmistakable result.
To prevent this kind of distortion, it is important to make sure the analog levels going into the analog-to-digital converter never force the system past its maximum. Meters are essential here. Digital recorders generally have meters that measure the amplitude of the signal in decibels below full scale. At full scale, the digital system has reached its maximum digital value. Inputs below this level are successfully converted from analog to digital. Inputs above this level will not be stored with appropriately large numbers. Clipping commences. It is the engineer′s job to make sure the vocal, snare, or didgeridoo sound being recorded is never so loud that the digital system exceeds full scale and runs out of numbers.
Figure 8.4 Analog-to-digital converters will hard clip a waveform when overloaded.
Well … mostly. Engineers who are intrigued by the squared-off waveform shown in bold in Figure 8.4, and wonder what it sounds like, might want to overdrive the digital system on purpose. This is certainly allowed, if the engineer is careful. First, monitor at a low level. This kind of distortion is full of high-frequency energy and can melt tweeters. Second, listen carefully. This type of distortion is extremely harsh; it is not a particularly natural or musical effect. Use it sparingly, if at all.
A small amount of digital overload on a snare hit might give it a desirable, aggressive, unnatural sound that fits into some production styles. The instantaneous burst of harmonics associated with hard clipping adds complexity to the timbre of the snare sound during its brief percussive impact. Digital overload on a more continuous sound, such as a vocal or a rhythm guitar, should be reserved as a rare special effect, as it is almost always a rather unpleasant, fatiguing sound.
For analog magnetic recording systems, the typical approach is to record as hot as possible, and occasionally go over and into the red. Unlike digital audio, analog audio does not typically hit such a hard-and-fast limit. When the level of the music being recorded gets too high, analog magnetic tape distorts only gradually. This less-abrupt distortion at the peaks is called soft clipping, shown in Figure 8.5 and discussed in Chapter 4. At lower amplitudes, the analog magnetic storage medium tracks fairly accurately with the waveform. As the audio signal starts to reach higher amplitudes, the analog storage format can not keep up. It starts to record a signal that is a little lower in amplitude. As analog tape runs out of amplitude capability, it does so gradually and gracefully. Analog magnetic tape, when recorded at too hot a level, distorts, but in a less abrupt way than digital audio systems.
Figure 8.5 Analog magnetic tape recorders will soft clip a waveform.
A careful look reveals that overdriving analog tape might have a lot in common with compression (see Chapter 6).In fact, recording at high levels to analog tape is called tape compression. No compressor will exactly mimic tape compression, as the way it reshapes the amplitude detail of an audio waveform comes from a kind of low-threshold, variable-ratio compression with infinitely fast attack and release. Tape compression moves in and out on the peaks of the complex waves, affecting the amplitude of each individual peak, whether part of a low-frequency or high-frequency signal.
Augmenting those rack spaces and pulldown menus full of compressors, an engineer can overdrive analog magnetic recorders as an effect. It is not just a storage device; analog tape is part EQ, part distortion, part compression too. Learning to use analog formats to process sound is essential for experienced engineers.
On a split console, engineers dedicate certain input/output modules of the console or digital audio workstation to recording signals to the multitrack, and others to monitoring those signals (see Chapter 2). On an in-line console, every module has two faders: one for the record path (also known as the channel path) and the other for the monitor path (also called the mix path).
Two different forces motivate the engineer when setting the level of the channel faders and the monitor faders. On the record path, the fader determines the level of the signal to the multitrack recorder. Analog or digital, the engineer makes the best use of the dynamic range available. It is primarily a technical decision, balancing signal-to-noise ratio versus headroom.
On the monitor path, the level settings for these faders are almost entirely a creative decision. The engineer chooses the appropriate level of each and every track in the multitrack arrangement so that the desired musical statement is made. Balance the mix, and make it beautiful, sad, exciting, ___________ (insert production goal here).
In a digital audio workstation, almost every adjustable parameter can be automated. In the hands of a clever engineer, a digital audio workstation can be instructed to wake us to music first thing in the morning (noon), start the coffee maker, check email, and draw a warm bath. While this is all quite useful, automation is almost always used for just two very simple processes: fader rides and mutes. Automation is, in essence, a volume processor. Chapter 15 is dedicated to the subject.
For example, using the humble mute switch, the mix engineer controls the multitrack arrangement: cut the bass in the extra bar before the chorus, pull the flute out of the horn section until the last chorus, etc. These sorts of details can enhance the tune and help keep listeners interested. It is accomplished very simply, using some well-placed automated cuts. This sort of mix move happens throughout pop music. But check out an extreme example by listening to U2′s Achtung Baby. The album begins with some heavy cut activity as the drums and bass enter at the top of the first tune of the album, ″Zoo Station.″
Automating fader rides in support of the arrangement is a natural application of automation. Maybe it makes sense to push the guitar up in the choruses, pull the Chamberlain down during the guitar solo, and such. Ideally, the band (maybe with the advice of a producer) gets these dynamics right in their performance. Players know intuitively when to push their level up a little as the song develops. But in the studio, the full arrangement of the song may not come together for several months, as overdubs are gradually added to the tune. Fader rides may be just the ticket to help this assembly of tracks fall into a single piece of music with appropriate expressions of volume.
Volume changes are automated just to keep the song in balance (discussed above) as multitrack components of the song come and go. Typically, it is best to keep these moves quite subtle. These rides are aimed at the musical interpretation of the mix, trying to make the song feel right. With few exceptions, it should never sound like a fader was moved. Listeners want to hear the music, not the mixing console. Just a small ride here and another one there will help shape the energy level and mood of the mix.
Another automated volume effect is the automated send. Although the aux send knob is not usually automated, it can be (see Chapter 15). Some very sophisticated mix elements are created with automated sends. In this way, automation is employed to add rich and spacious reverb to the vocal in the bridge only, introduce rhythmic delay to the background vocals on key words, increase the chorus effect on the orchestral strings in the verses, add distortion to the guitar in the final chorus, etc. Automate the additional effect in and back out of the mix wherever it is wanted. The automated send — just another volume effect — offers the engineer a way to layer in areas of more or less effects, using nothing more than straight-forward faders and cuts automation.
When editing analog tape, one generally chooses a pronounced sonic moment, such as a snare hit, and cuts the tape at an angle. Placing the edit just before a snare hit helps keep the engineer oriented while editing. The snare hits provide a sonic clue identifying major increments in musical time in the song form. The snare hit also performs a bit of forward masking (see Chapter 3), making the cut in tape that occurs the instant before the snare hit difficult, if not impossible, to hear.
Cutting the tape at an angle makes the transition from one part of the tape to another gentler. Rather than having the pattern of magnetism on tape change suddenly from one performance to another, the angled cut magnetically cross fades between the two edit pieces.
Editing in a digital audio workstation offers more opportunities for rearranging a song. The ability to zoom in on the waveform enables great precision in choosing the edit point. The fact that digital edits may be nondestructive to the audio waveform, amounting only to a difference in the way the computer plays back the data, not an alteration to the digital audio data itself, makes cutting and rearranging an audio waveform as simple as cutting and pasting text. The engineer can try an edit, audition it, undo it, or modify it, ad nauseam. Digital audio workstations have made editing a much more frequent part of the music production process. Physical cuts into analog magnetic tape are more difficult to undo. Zooming in on analog tape is unproductive as the waveform is not visible.
Edits are quite volume dependent. Making edits between the notes, when the instrument is at least briefly silent (i.e., not playing) is a good first strategy. Further, performing all edits at zero crossings, so that the amplitude of the audio at the point of edit is zero, is a good habit. In fact, most digital editors can be persuaded to default to zero crossings for all edits. In this way, the amplitude of the signal at the instant of the edit is as low as practical.
Frequently the goal is to remove any audible artifact associated with the edit so that the listener feels the edited performance is the original performance. In this case, a cross fade is frequently required. The cross fade amounts to a fading out of the first event during the fading in of the second event. There are production situations in which cross fades need to be as short as a few milliseconds, or as long as several thousand milliseconds.
Steady-state signals, such as string pads, piano notes in long decay, long reverb tails, and hiss, can be unforgiving of edits. Cutting from one part of the steady sound to another similar steady sound often causes a noticeable click, thump, or other sonic discontinuity at the edit point. Any abrupt change from one steady-state sound to another, similar sound will reveal even slight differences between them. Editing during a distracting event can mask this. In addition, a cross fade between these similar, but not identical, steady-state sounds might make the edit aurally invisible.
Low-frequency signals, with their relatively slowly cycling waveforms, are intolerant of changes to the progress of those slow cycles. Low-frequency instruments are particularly difficult to edit. Often, it is not enough to work at zero crossings. A brief cross fade will smooth the edit.
Commercial radio stations can not endure a fade out. The DJ will start talking, the next song will begin, and/or the advertisement will be played before a song is ever allowed to fade to silence. Live concerts make the fade impossible. Yet, many music recordings fade out.
It is a practical solution to a sometimes difficult problem. Composers find it challenging to end a piece, and may not compose an actual ending. Bands find it difficult to stop cleanly; the end can be more difficult than the beginning of a tune. When the band flies through a take, thinking, ″This is the one! This is the keeper! We are going to be stars!,″ the last thing they need to worry about is a tight ending, which, if someone rattles on too long, renders the whole take unusable. It is sometimes deemed better to play several extra bars or an extra repeat of the chorus and fade it in the mix. Sometimes mistakes made by one or more performers may need to be hidden. The fade out covers any number of sins.
The fade is a musical statement, however. If used to simply cover over a problem, the result can be musically weak. Listeners may not be consciously aware of the cause, but will feel the music failed to finish, to take them somewhere, to achieve fruition. The fade out should not be the assumed ending, it should be a deliberate ending. Bands who arrange and rehearse actual endings to their songs are taking advantage of the ending of the tune to have a created, musical effect.
Artist: U2
Song: ″Zoo Station″
Album: Achtung, Baby
Label: Island Records
Year: 1991
Notes: The power of the cut button is displayed in the intro to the song and the album as the drums and bass enter.
Artist: The Smiths
Song: ″Hand in Glove″
Album: Hatful of Hollow
Label: Rough Trade Records
Year: 1984
Notes: The tune fades out, back in, and then out again. Listeners can only wonder what possibly went wrong and had to be hidden.