Audio Intermediate

Introduction

There is a moment in every audio career when the system stops being the problem. The microphones arrive on the right channels, nothing hums, nothing feeds back, and the show goes out to the room without incident. And it is fine. It is completely fine. It is also, if you are honest with yourself, a little flat, a little cluttered, and not quite the thing you hear when somebody good is mixing.

That gap is what this book is about. It is not a gap in equipment and it is rarely a gap in effort. It is a gap in craft: in knowing which of forty available adjustments is the one worth making, in hearing a problem precisely enough to name the frequency, in understanding what a compressor is actually doing to a transient rather than what the preset is called.

Audio Basics got signal from a microphone to a loudspeaker. Live Sound Production got you through a show day. This book is the craft in between those two things: what you do to a signal once it is arriving cleanly, and how you make a loudspeaker system and a room agree with each other. It assumes you can already patch a stage and run a line check without help.

You will also find the gear named here. Consoles, loudspeaker systems, amplifiers, measurement software and microphones appear by manufacturer and model throughout, because that is how the job is actually spoken. An engineer who walks up to a desk and knows what it is starts the day ahead of one who does not, and the difference is nothing but vocabulary.

How to Use This Book

Read it in order the first time. The chapters are built so that gain structure comes before EQ, EQ before dynamics, and both before any discussion of mixing, because that is the dependency order of the decisions themselves. A compressor set on a badly gain-staged channel will behave in ways no amount of reading about ratios will explain.

After that, treat it as a reference. Each chapter ends the same way: a reference table of the things worth memorising, a procedure you can follow on a real job, a failure table matching symptoms to causes, a practical exercise, and a checklist that tells you when that part of the work is genuinely finished.

• The reference tables are the parts worth returning to on a show day.
• The failure tables are written as symptoms first, because that is the order you meet them in.
• The exercises need a real system or a real recording; none of them can be done by reading.
• The checklists are the standard, not a suggestion.

Keep a notebook

Write down the settings that worked and the room they worked in: the threshold and ratio that held a vocal, the crossover point and the delay time for that system, the frequency you cut and by how much. At this level the same problem keeps arriving in different rooms, and a notebook is what turns a lucky fix into a method you can repeat.

A note on the words

Several words here name a control on one machine and something different on another. Gain, trim and input level may be the same knob or three separate ones depending on the desk, and a preset called Vocal tells you nothing about what it is doing to a transient. This book names what the processor is actually doing first and the controls that do it second, because that is the order in which the explanation stays true when you move to a different desk.

A note on safety

Two hazards sit behind everything in this book. The first is your own hearing: noise-induced hearing loss is permanent, cumulative and completely preventable, and it is the most common reason audio careers end early, so wear filtered plugs and count the long afternoon of soundcheck as exposure too. The second arrives as systems get bigger than a pair of tops on sticks -- mains power, and the weight of what you are lifting. Nothing here qualifies you to do electrical work or to rig anything overhead.

Where this book sits

This book assumes you have read Audio Basics and Live Sound Production, or that you already know what is in it.

If none of that is true yet, read Audio Basics first.

What pairs with this

• Acoustics Fundamentals — why rooms do what this book keeps working around
• Wireless & RF Systems — the sources that arrive without a cable

Where it leads

These are not steps in an order. Take whichever one fits.

• Monitor Engineering — the specialism where this craft is pointed at the stage
• Stage Audio & IEM Systems — the stage end, in depth
• Front-of-House Mixing — the room end, in depth

Beyond the Basics

Everything in this book rests on one idea: at the intermediate level you stop reacting to problems and start making decisions. The beginner asks "is it working?" The intermediate engineer asks "is this the sound I meant?" Those are different questions, and only the second one has craft in it.

Basics make audio work; intermediate craft makes it sound good.

From working to great

Audio Basics and Live Sound got you a working mix — clean gain, balanced levels, no feedback. Audio Intermediate is about making audio sound genuinely good: shaping tone with EQ, controlling dynamics with compression, adding space and polish with effects, and combining these into a professional mix. The difference between a functional mix and a great one is largely this processing craft, applied with skill and taste. This book develops that craft, deepening the channel-strip tools you met in Audio Basics into deliberate, musical use.

The mindset shift is from making it work to making it sound right. You already know the signal flows cleanly; now you learn to sculpt it — to make a vocal sit perfectly in the mix, to tighten a loose sound, to add a sense of space, to balance everything into a coherent whole. This is where audio becomes an art on top of the technical foundation, and where an operator becomes a mixer.

The processing chain on a channel

Every channel on every console is a chain of processes in a fixed order, and each one hands its result to the next. What comes out depends not only on the settings but on where in that line each adjustment sits. This sounds obvious until you meet an engineer wondering why their compressor keeps ducking on a rumble they have already filtered out visually but not actually.

The conventional order, from input to output, is: input gain, then high-pass filter, then gate or expander, then EQ, then compressor, then the fader, then the sends, then the output routing. Some consoles let you swap the EQ and the compressor, and a few let you move the gate. Most do not, and the default order is the default for good reasons.

• Input gain sets how hard the signal hits everything downstream. Nothing after it can fix a gain that is wrong.
• The high-pass filter removes energy nothing downstream needs to waste effort on -- especially the compressor.
• The gate decides whether the channel is heard at all, and it works best on a signal that has already lost its rumble.
• EQ before the compressor means the compressor responds to the tone you chose; EQ after means it responds to the tone you had.
• The fader is the last thing before the sends on most desks, which is why post-fade effect sends follow your mix moves.

Put an EQ before a compressor and you are telling the compressor what to listen to. Cut 200 Hz before the compressor and the compressor stops reacting so hard to the low mids. Cut it after, and the compressor is still ducking on energy you are about to throw away. Neither is wrong. They are different instructions, and the intermediate engineer gives the one they meant.

What changes when the system stops being the problem

A beginner's attention is almost entirely downward and outward: is there signal, is it on the right channel, is anything buzzing, is anything about to feed back. That attention is not wasted. It is the foundation, and an engineer who has not automated it cannot afford to think about anything else. But once it is automatic, something uncomfortable happens. The attention is freed, and there is nothing obvious to spend it on.

What you spend it on is the difference between a functional mix and a good one. It goes into listening to the tone of a single source and deciding whether it is the tone you want. It goes into noticing that the vocal is clear but sitting on top of the band rather than inside it. It goes into hearing that the room has a ring around 250 Hz that you have been mixing around all evening without naming.

None of that is more equipment. Every tool in this book was on the console the day you learned to plug it in. What changes is that you now know what each one is for, which one this particular problem calls for, and -- most importantly -- when the answer is to leave all of them alone.

There is a trap here worth naming immediately. The newly intermediate engineer, having discovered processing, processes everything. Every channel gets an EQ curve, a compressor and a reverb send, because using the tools feels like doing the job. It is the audio equivalent of a young cook discovering salt. The mixes that result are busy, tired and oddly lifeless, and the way out is not more skill with the tools but the judgement to use fewer of them.

Deciding before adjusting

The habit that most reliably separates levels is a small one: name the problem before you touch anything. Not "this needs work", but "the snare has a ring about an octave above the fundamental that is making it sound like a biscuit tin". The second sentence contains its own solution. The first one leads to twenty minutes of turning knobs.

This is harder than it sounds, because the ear is much better at noticing that something is wrong than at saying what. Building that vocabulary is a large part of what the rest of this book is for, and it is the thing the ear-training work in the final chapter is designed to grow.

Until it is fluent, use a discipline: make one change, listen, and decide whether it helped. Then keep it or undo it. An engineer who stacks six changes without evaluating any of them has not made six improvements; they have made one unknown state.

The channel chain, in order

Each stage hands its result to the next. Knowing the order tells you where a problem can and cannot be solved.

How to do it

1. Set the input gain with the source at its loudest realistic level, leaving headroom.
2. Engage the high-pass filter and sweep it up until the source thins, then back off.
3. Listen to the raw channel alone and say out loud what is wrong with it.
4. Make one EQ move that addresses what you just said, and listen again.
5. Only then decide whether it needs dynamics at all.
6. Put it back in the mix and check that your decision still makes sense there.

What goes wrong

Practice it

On a console with a channel you can experiment on, using recorded material.

Gain Structure and Headroom

Almost every problem that feels like a processing problem is a gain problem wearing a disguise. A compressor that will not behave, a channel that is noisy, a mix that distorts only when it gets exciting, a digital desk that sounds brittle -- all of these are usually one badly set level somewhere in a chain of twenty.

What gain structure actually means

Gain structure is the practice of setting the level at every stage of a signal path so that the signal sits comfortably above the noise floor and comfortably below the point where anything distorts. That is the whole idea. What makes it difficult is that a modern signal path has far more stages than most people count, and getting it wrong at any one of them costs you at all the others.

Consider a single vocal. It passes through the microphone, a preamplifier, an analogue-to-digital converter, a channel's internal processing, a group, a matrix, a digital-to-analogue converter, a system processor, an amplifier, and finally a loudspeaker. That is ten or more places where level is set, and most of them are invisible from where you are standing.

The signal has a floor and a ceiling. Below the floor is the inherent hiss of the electronics; too little gain early and you amplify that hiss along with the music for the rest of its journey. Above the ceiling is distortion; too much gain anywhere and the signal is clipped, permanently, in a way that no downstream adjustment can undo. The working space between them is headroom, and headroom is the thing you are protecting.

Unity, and what it does and does not mean

Unity gain means a stage passes the signal through without making it louder or quieter. On a fader it is usually marked 0, and it is usually about three-quarters of the way up rather than at the top -- a detail that surprises people who assume the top of the throw is neutral.

Faders are designed to be most precise around unity. The marks are spread out there and bunched together at the bottom. An engineer mixing with every fader at the bottom of its travel has almost no resolution: a millimetre of movement is several decibels, and fine balance becomes impossible. An engineer mixing with everything pinned at the top has nowhere to go when something needs more.

So the practical rule is to set input gains such that, with the show running at the level you intend, the faders sit somewhere near unity. If they are all at the bottom, your gains are too high. If they are all at the top, your gains are too low. This is not aesthetics; it is the difference between having control and not having it.

Digital headroom, and where the ceiling really is

Analogue systems distort gradually. Push an analogue console's input and it begins to sound thicker, then harder, then unpleasant, over a range of several decibels. Generations of engineers learned to use that region deliberately.

Digital systems do not work that way. Below the maximum level a digital system is clean; at the maximum level it is clean; above the maximum level there is nothing at all, because there is no number available to describe a larger sample. The result is not warmth but a hard, splintery distortion that appears without warning and sounds nothing like overdriven analogue.

This is why digital practice leaves more headroom than analogue practice did. Aim to have your loudest peaks land somewhere around minus eighteen to minus twelve on a digital meter rather than flirting with the top. It feels conservative. It is the reason your mix survives the moment the drummer hits harder than they did at soundcheck.

• The ceiling is absolute: there is no forgiving region above it.
• Peaks are much higher than averages, and meters that show averages will lie to you about how close you are.
• Clipping early in the chain travels all the way to the loudspeaker and cannot be repaired downstream.
• Leaving headroom costs you nothing; the noise floor of modern digital equipment is far below anything you will hear in a venue.

Gain structure across a whole system

The same logic applies beyond the console. A system processor with its input attenuated and its output boosted is amplifying its own noise. An amplifier run with its sensitivity turned down while the console runs hot is wasting headroom at one end and risking clipping at the other.

The convention that solves this is simple: run the console at a sensible working level, run the processing at unity, and set the amplifier sensitivity so the system reaches the level the room needs with the console near unity. Then the limiter protecting the loudspeakers is the only thing that should ever intervene, and when it does, it means something.

On a touring system this is set once by the system engineer and left alone. On a house system you may inherit whatever the last person did. It is worth an hour to find out, because everything you do afterwards depends on it.

Gain on a source that will not sit still

Not every source has a predictable level. A singer who whispers in the verse and screams in the chorus, a presenter who leans in and then away, a guest handed a microphone for the first time in their life -- these defeat a single gain setting by design.

The intermediate answer is not to ride the gain. It is to set the gain for the loudest realistic case, so nothing ever clips, and then use a compressor deliberately to lift the quiet parts. That is what the compressor is for, and it is why the next chapters come in the order they do.

There is one exception worth knowing: some digital consoles offer a trim after the preamplifier, separate from the analogue gain. Where a preamplifier is shared between front of house and monitors, the analogue gain belongs to whoever agreed to own it, and everyone else uses their own trim. Changing a shared gain without saying so is one of the few genuinely rude things an engineer can do.

Where level is set, and what it costs to get it wrong

A signal passes through more gain stages than most engineers count.

How to do it

1. Play known material at a realistic show level.
2. Look at the console output meters and note where the peaks land.
3. Look at the system processor input and output meters and compare.
4. Check whether the amplifier or processor limiters are active at show level.
5. Note where the faders are sitting with a normal balance.
6. Adjust at one end only -- usually amplifier sensitivity -- and re-check the whole chain.

What goes wrong

Practice it

On any system you can meter end to end.

EQ & Dynamics

These are the two tools you will reach for most, and they are the two most often used without understanding. This chapter is about the controls themselves -- what each knob physically does to a signal. The two chapters after it are about when to use them, which is a different and harder question.

The two workhorses of audio processing: shaping tone, and controlling level.

EQ in depth

EQ (equalization) shapes the tone of a sound by boosting or cutting specific frequency bands — the frequency theory from Audio Basics made into a precise tool. Intermediate EQ use is deliberate: cutting a frequency that makes a sound muddy or harsh, boosting one that adds clarity or warmth, carving space so instruments do not clash in the same frequency range, and removing problem frequencies (like rumble or feedback-prone areas). Good EQ is often subtractive — removing what does not belong more than adding — and it is guided by listening: you hear a problem or a need and use EQ to address it. EQ is the most-used processing tool, and using it skillfully transforms a mix.

EQ serves both correction and enhancement: correcting problems (a boomy room, a harsh vocal, competing instruments) and enhancing sounds (adding presence, warmth, air). The intermediate skill is hearing what a sound needs and knowing roughly where in the frequency spectrum to act, then making focused adjustments. Over-EQing (boosting everything) muddies a mix; targeted, often subtractive EQ clarifies it. Developing your ear for frequency and a deliberate, restrained EQ technique is central to intermediate audio.

Dynamics and compression

Dynamics processing controls the level of a signal over time, and the key tool is the compressor, which reduces the difference between the loudest and quietest parts of a sound — taming peaks and bringing up quieter moments so the sound is more consistent and controlled. Compression keeps a vocal steady so it stays present without spiking too loud, tightens a bass or drum, and generally gives a sound a controlled, professional feel. Used well, compression makes sources sit consistently in the mix; used heavily or poorly, it can squash the life out of sound, so intermediate skill is using it appropriately.

Understanding compression conceptually — it reduces dynamic range, controlling loud peaks and lifting quiet parts — lets you use it to solve real problems: a singer whose level varies wildly, a source that keeps poking out or disappearing. The intermediate mixer applies compression where it helps (control and consistency) without overdoing it. Together, EQ (tone) and dynamics (level control) are the two workhorse processing tools, and using both deliberately is the heart of shaping individual sounds well.

Filters, shelves and bells

An equaliser makes some frequencies louder or quieter than others. Everything else is detail, but the detail is what you are paid for. There are three fundamental shapes, and every EQ you will meet is made of them in some combination.

A filter removes everything beyond a point. A high-pass filter lets high frequencies pass and progressively removes the lows below its corner frequency; a low-pass filter does the opposite. Filters have a slope, usually described in decibels per octave -- 12 dB per octave is gentle, 24 dB per octave is firm. The corner frequency is not where the sound stops; it is where the removal has reached a defined amount, usually three decibels, and the effect extends above it.

A shelf lifts or lowers everything beyond a point by a fixed amount, then stops changing. A high shelf boosted by three decibels at 8 kHz raises everything above roughly that region by about three decibels and leaves it there. Shelves are broad, gentle and good for tonal decisions -- "this needs more air" -- rather than surgical ones.

A bell -- also called peaking or parametric -- lifts or lowers a region centred on one frequency, tapering away on both sides. It has three controls: the frequency at its centre, the gain applied there, and the width, called Q. A high Q is narrow and surgical. A low Q is broad and musical. Most corrective work uses a narrow bell; most tonal work uses a wide one.

• High-pass filter: removes lows below a corner, at a chosen slope.
• Low-pass filter: removes highs above a corner.
• Shelf: a fixed lift or cut of everything past a point.
• Bell: a lift or cut centred on a frequency, width set by Q.
• Notch: an extremely narrow, deep cut -- a bell taken to its limit.

Q, and why width matters more than people think

Q describes how wide a bell's effect is. It is defined as the centre frequency divided by the bandwidth, which means a high number is a narrow curve -- the opposite of most people's intuition on first meeting it.

The practical consequence is this: a narrow cut removes a problem and leaves the surrounding sound largely untouched, which is exactly what you want when fighting a specific resonance. A wide cut changes the character of the whole source, which is exactly what you want when a source is simply too dark or too honky across a region.

Beginners tend to use narrow curves for everything, because narrow feels precise and precise feels skilful. The result is a channel full of small deep cuts that fix nothing audible while making the source sound hollow. The intermediate habit is the reverse: use a wide curve unless you have identified a specific narrow problem, in which case use a narrow one and go deep.

The compressor's six controls

A compressor reduces the level of a signal once it exceeds a chosen point. That is all it does. Every use of a compressor -- control, polish, loudness, character, the glue on a drum bus -- is that one behaviour applied with different settings.

Threshold is the level above which it starts working. Below the threshold nothing happens at all. Ratio is how firmly it reduces what exceeds the threshold: at 2:1, two decibels of input above the threshold become one decibel of output; at 10:1 it is effectively a limiter. Attack is how quickly it begins reducing once the threshold is crossed. Release is how quickly it stops once the signal falls back. Knee describes whether the transition at the threshold is abrupt or gradual. Make-up gain restores the level the compression removed.

Two of those controls cause almost all of the confusion. Threshold and ratio together decide how much compression happens, and they interact: a low threshold with a gentle ratio and a high threshold with a firm ratio can produce similar amounts of gain reduction and sound completely different. Attack and release decide what the compression sounds like, and they are the subject of most of the next-but-one chapter.

Read the gain reduction meter, not the settings. The meter tells you what is actually happening; the settings only tell you what you asked for. Three to six decibels of gain reduction on peaks is a normal, useful amount for control work. Fifteen decibels is a deliberate effect, and if you did not intend it, something is wrong.

Gates and expanders

A gate is the mirror image of a compressor: instead of reducing what is above a threshold, it reduces -- or removes -- what is below one. On a live stage where every microphone hears every instrument, that is how you buy back separation.

A gate has a threshold, an attack, a hold and a release, and often a range. Range is the most useful control and the least used: rather than closing the channel completely, it reduces it by a set amount. A gate with a range of twelve decibels cleans up spill without the unnatural sound of a channel snapping shut.

An expander is a gate with a gentler hand: below the threshold it reduces level progressively rather than shutting off. On sources where a hard gate sounds obviously mechanical -- a floor tom, a quiet vocal -- an expander does the same job without announcing itself.

The EQ shapes and what each is for

Choose the shape before the frequency; the shape is the decision.

The compressor's controls

Threshold and ratio set how much. Attack and release set what it sounds like.

How to do it

1. Take a bell, set a narrow Q and boost it substantially -- ten decibels or more.
2. Sweep the frequency slowly across the range while the source plays.
3. Listen for the point where the problem becomes obviously, unpleasantly worse.
4. Stop there; that is the frequency.
5. Now reverse the gain into a cut, and reduce the Q until it sounds natural.
6. Use the smallest cut that solves the problem.

What goes wrong

Practice it

With a recorded multitrack and one channel you can experiment on.

EQ in Practice

Knowing what a bell curve does is not the same as knowing which one this guitar needs. This chapter is about the decisions: what to cut, what to leave alone, where the common problems live, and why the answer is so often less rather than more.

Why cutting usually beats boosting

Ask an experienced engineer to fix a muddy mix and watch what they do. Almost without exception they will cut something. Ask a beginner and they will boost something else. Both are trying to change the same relationship, and only one of them is also managing headroom, phase and the loudness of the result.

There are three reasons cutting wins. The first is headroom: every boost adds level that has to go somewhere, and a channel with four boosts on it is several decibels hotter than it was, eating the headroom you carefully protected. The second is phase: every conventional EQ shifts phase around the frequencies it changes, and a boost shifts it in a way that is generally more audible than a cut of the same size.

The third reason is psychological, and it is the largest. Boosting makes things louder, and louder sounds better -- for about ten seconds. An engineer boosting their way through a mix is receiving constant positive feedback for decisions that are collectively making things worse. Cutting removes that flattery and leaves you judging the actual balance.

This is a strong default, not a religious rule. If a source genuinely lacks air, a gentle high shelf is the right answer and no amount of cutting elsewhere will produce it. But reach for the cut first, and boost only when you have established that nothing can be removed to achieve the same thing.

The high-pass filter as a discipline

Of every habit in this book, routine high-pass filtering buys the most clarity for the least effort. A vocal microphone picks up stage rumble, foot noise, air conditioning and handling noise, all of it below the range where the voice lives and all of it consuming headroom and driving compressors. Removing it costs nothing audible.

The technique is to sweep rather than guess. Engage the filter, set it low, and raise the corner frequency slowly while the source plays until you hear the source begin to thin. Then come back down a little. On most vocals that lands somewhere between 80 and 120 Hz; on a spoken presenter it may be higher; on a bass guitar or kick drum it may be as low as 30 Hz and it is still worth doing, because there is energy below that which nothing in the room can reproduce.

The one source that usually should not be filtered aggressively is anything whose fundamental genuinely lives down there. Filter a five-string bass at 80 Hz and you have removed its lowest notes. The discipline is to filter everything deliberately, not to filter everything identically.

• Vocals: typically 80-120 Hz, swept to taste.
• Guitars and keys: often 100-150 Hz; they rarely need anything below.
• Overheads and cymbals: can go surprisingly high, 200-400 Hz, to remove spill.
• Kick and bass: still filtered, but low -- 25-40 Hz -- to remove subsonic energy.
• Speech microphones: often 100-150 Hz, and a low-pass at the top helps too.

Where the common problems live

Frequency ranges are not magic numbers, and any table of them is a starting point rather than a prescription. But the same complaints recur, and knowing roughly where to look saves a great deal of sweeping.

Below about 100 Hz is weight and rumble. Between roughly 100 and 250 Hz is warmth, and also the region people mean when they say "muddy" -- too much energy here from several sources at once is the single most common problem in an amateur mix. From about 250 to 500 Hz is the "boxy" region, where a source can sound like it is in a carton. Around 800 Hz to 1.5 kHz is honk and nasality, and where snare rings usually live.

From about 2 to 5 kHz is presence and intelligibility -- this is where consonants live, and where a vocal fights to be heard over a band. It is also where the ear is most sensitive, so changes here are more audible per decibel than anywhere else. From about 5 to 8 kHz is edge and sibilance. Above 8 kHz is air, and also hiss, cymbal wash and the sound of a microphone's self-noise.

The intermediate skill is not memorising this. It is being able to hear a problem, guess the region, sweep to confirm, and be right often enough that the guess is worth making.

Making room instead of making louder

When two sources compete, the instinct is to turn one up. The better move is usually to remove, from one of them, the region the other one needs.

The classic case is a vocal and an electric guitar. Both have substantial energy between 2 and 4 kHz. Turn the vocal up and the mix gets louder and no clearer, because the guitar is still occupying the same space. Cut a couple of decibels from the guitar in that region instead and the vocal appears, with no change in overall level at all.

The same logic applies to kick and bass guitar, to snare and rhythm guitar, to piano and almost everything. Decide which source owns which region, and let the others give way. This is what experienced engineers mean by "carving": it is not a trick, it is the recognition that a mix is a shared and finite space.

A useful discipline: when you want to boost something, first try cutting the same region on whatever is competing with it. If the result is what you wanted, you have solved the problem without adding a decibel to the mix.

Dynamic EQ and multiband, and when they are the right answer

Some problems are intermittent. A vocal that is fine most of the time but harsh on the loud notes. A bass that is even on most notes and booms on one. A presenter whose sibilance is savage only when they get close to the microphone. A static EQ fixes these by removing the region all the time, which dulls everything else. A broadband compressor fixes them by pulling the whole signal down whenever the problem appears, which is audible and unmusical.

A dynamic EQ is the correct tool: it is an EQ band that only acts when the energy in that band exceeds a threshold. The harsh note gets cut; the rest of the performance is untouched. A multiband compressor is the same idea arranged differently -- the signal is split into bands and each is compressed separately.

A de-esser is simply a dynamic EQ or multiband compressor aimed at the sibilance region, usually between 5 and 8 kHz. It is worth knowing that this is all it is, because then you can build one on any console that has a dynamic EQ, whether or not it offers something labelled "de-esser".

These are precision tools and they are easy to overuse. The question to ask is always: is this problem constant or intermittent? Constant problems belong to static EQ. Intermittent ones belong here.

Where to look first

Starting points, not prescriptions. Sweep to confirm before cutting.

Static or dynamic?

The question is whether the problem is always there.

How to do it

1. Solo the two sources together and identify which one must win.
2. Find the region where they overlap most by sweeping a boost on the one that must give way.
3. Cut two to three decibels there, with a moderate Q.
4. Return both to the full mix and listen without touching any faders.
5. If the winner is now audible, stop; if not, cut a little more rather than boosting.
6. Check that the source you cut still sounds right on its own.

What goes wrong

Practice it

With a multitrack of a full band.

Dynamics in Practice

Compression is the process most often applied by preset and least often understood. This chapter is about what the controls do to the actual shape of a sound, and about the difference between using a compressor to keep something under control and using one to make it sound a particular way.

What attack and release do to a transient

Every sound has a shape in time. A snare drum has a sharp initial spike lasting a few thousandths of a second, then a body, then a decay. A vocal has a consonant, then a sustained vowel. The attack and release controls on a compressor decide which parts of that shape get reduced and which survive intact.

A fast attack -- say one millisecond -- catches the initial spike and pulls it down. The result is a snare that sounds rounder, softer, further away, and more even. A slow attack -- say thirty milliseconds -- lets the spike through before reduction begins, so the crack survives while the body afterwards is compressed. The result is a snare that sounds punchier and more aggressive than it did uncompressed, which surprises people who expect compression always to soften.

That is the single most useful thing to understand about compression: a slow attack makes things sound more punchy, not less, because it compresses the body and leaves the transient standing proud of it.

Release decides how quickly the compressor lets go. Too fast and you hear the level surging back up between words or hits, an effect called pumping or breathing. Too slow and the compressor never recovers between events, so it sits holding everything down and the source sounds squashed and lifeless. The useful setting usually follows the tempo of the material: the compressor should recover roughly in time with the music.

Compression for control, and compression for character

These are different jobs and they want different settings. Control means keeping a source within a usable range -- a singer whose quiet lines would otherwise disappear and whose loud lines would otherwise dominate. Character means deliberately changing how a source sounds, using compression as a tonal effect.

Control work is gentle and mostly invisible: a ratio around 3:1, a threshold set for three to six decibels of gain reduction on the loudest moments, a moderate attack, a release that follows the music. Done well, bypassing it makes the source harder to place in a mix but does not obviously change its sound.

Character work is deliberate and audible: heavier ratios, more gain reduction, attack and release chosen for their effect on the envelope rather than to be transparent. Parallel compression on a drum bus, a heavily squashed room microphone, a vocal pushed until it sits forward and breathy -- these are effects, chosen on purpose.

The mistake is doing the second while intending the first. An engineer who sets a 10:1 ratio and pulls the threshold down until the vocal "sits better" has not controlled it; they have flattened it, and the performance's dynamics -- the thing the singer was doing on purpose -- are gone.

Gating a live stage

On a live stage every open microphone hears everything. A tom microphone hears the whole kit; a vocal microphone hears the guitar amplifier behind it. That spill is not merely untidy: it accumulates, it muddies, and it introduces comb filtering as the same sound arrives at multiple microphones at slightly different times.

Gating buys separation back. The classic application is drums, where gating the toms removes the wash between fills and dramatically tightens the kit. It is also useful on a lectern microphone that would otherwise carry room noise all evening, and on a guitar amplifier that hisses between songs.

The controls matter more than on a compressor because a gate's mistakes are so audible. The threshold must be above the spill and below the quietest real hit -- a window that can be narrow. The attack should be fast enough not to swallow the start of a drum but not so fast it clicks. Hold keeps the gate open for a set time after the signal drops, preventing chatter. Release decides how naturally it closes.

The single most useful refinement is range. Set a gate to reduce by twelve decibels rather than closing completely and you get most of the separation with none of the unnatural sound of a channel vanishing. On a live show, where a gate that closes at the wrong moment is a missed hit in front of an audience, that margin is worth having.

Bus compression, parallel compression and sidechains

Compressing a group of channels together rather than individually produces something individual compression cannot: the sources move together. Compress a drum bus and when the kick hits, everything on that bus dips fractionally with it, which the ear reads as the kit being one instrument rather than seven microphones. This is what engineers mean by glue.

Parallel compression takes the idea further. Instead of compressing the source, you send a copy of it to a heavily compressed bus and blend that back underneath the original. The uncompressed signal keeps all its transients and dynamics; the compressed copy adds density and body underneath. You get the weight of heavy compression without losing the life of the original, and the blend control gives you exactly how much you want.

A sidechain makes one signal control the compression of another. The standard use is ducking: a compressor on the music bus keyed from the presenter's microphone, so the music drops automatically whenever they speak. In a band context, a sidechain from the kick into a bass compressor makes room for the kick on every beat without the engineer touching anything.

A sidechain filter is a related and very useful refinement: it filters what the compressor listens to without filtering what it passes. Filter the low end out of a vocal compressor's sidechain and it stops reacting to plosives and stage rumble while still compressing the whole signal.

Limiting, and the last line of defence

A limiter is a compressor with a ratio high enough that the output effectively cannot exceed the threshold. Its job is not musical; it is protective. Somewhere between your mix and the loudspeakers there should be a limiter whose only function is to make sure no decision of yours, and no accident of anyone else's, can damage the system or the audience.

On a modern system that limiter lives in the amplifier or system processor, set by the manufacturer's preset for the specific loudspeaker. It is not yours to adjust. What matters is that you know it is there, you know how to see whether it is working, and you understand what it means when it is.

A limiter that engages occasionally on the biggest moments is doing its job. A limiter that is working continuously is telling you the system is being asked for more than it has, and the answer is a bigger system or a quieter show -- not a heavier hand on the master fader.

Attack and release, by intention

The transient decides the character. Choose the attack for what you want to survive.

The dynamics tools and their jobs

All of them reduce level. What differs is when, and by how much.

How to do it

1. Bypass any EQ for the moment and listen to the raw channel.
2. Set ratio to 3:1 and attack and release to the middle of their range.
3. Lower the threshold until the loudest phrases show three to six decibels of gain reduction.
4. Adjust attack until the consonants sound the way you want.
5. Adjust release until the compressor recovers in time with the song.
6. Set make-up gain so bypassing the compressor does not change the loudness.
7. Now judge it in the mix, and only then re-enable the EQ.

What goes wrong

Practice it

With a recorded drum multitrack.

Effects & Building a Mix

Effects are the part of mixing that most obviously sounds like production, and the part most obviously overdone by people who have just learned to use them. Used well they create space and depth; used badly they create distance and wash. The difference is almost entirely in restraint and in routing.

Add space and polish, then balance everything into a coherent whole.

Time-based effects

Effects add space, depth, and polish to a mix, and the main time-based effects are reverb (simulating the sound of a space, adding a sense of room or ambience) and delay (repeating a sound, adding depth or rhythmic interest). A touch of reverb can make a dry vocal sound natural and present; delay can add space and dimension. Effects are usually applied via the aux sends (from Audio Basics) to shared effects, so multiple channels can share a reverb. Used tastefully, effects make a mix sound polished and three-dimensional; overused, they make it washy and indistinct, so intermediate skill is restraint and appropriateness.

The craft is using effects to serve the mix, not to show off: enough reverb to give a sense of space without drowning clarity, delay where it adds and not where it muddies. Different material wants different amounts — a clear spoken-word event wants minimal effects for intelligibility, while music may want more for richness. Knowing when and how much effect to use, and applying it via sends to shared effects, is the intermediate effects skill that adds the final polish to sounds.

Building a polished mix

Building a great mix combines everything: clean gain, balanced levels, deliberate EQ, appropriate dynamics, and tasteful effects, brought together into a coherent whole. A great mix has balance (every element at the right level, nothing too loud or buried), clarity (each important element clearly heard, achieved partly through EQ carving space), and tone and space (EQ and effects giving each sound and the whole mix the right character). The mixer works both on individual channels (shaping each source) and on the overall balance (how everything sits together), constantly listening and adjusting toward a mix that sounds clear, balanced, and polished.

This is where the intermediate skills come together as a craft. You are not just setting levels; you are sculpting a complete sonic picture — making the vocal sit perfectly over a balanced band, each instrument clear and in its place, the whole sounding professional and pleasing. It takes listening, judgment, and the deliberate use of all the processing tools. A polished mix is the goal everything in this book builds toward, and achieving it consistently is what marks an intermediate (and beyond) audio professional.

Reverb -- what it is actually simulating

A reverb simulates what happens to a sound in a space: the direct sound reaches the listener first, then a set of distinct early reflections from nearby surfaces, then a dense decaying tail as the energy bounces around. Every control on a reverb is adjusting one of those three things.

Decay time, sometimes called reverb time, is how long the tail takes to fade away. Short decays -- under a second -- read as a small room and keep a mix tight. Long decays -- three seconds and beyond -- read as a large hall and will smear a fast song into mush. Match it to the tempo of the material: a rough guide is that the tail should have largely gone before the next beat arrives.

Pre-delay is the gap between the direct sound and the start of the reverb, and it is the most underused control on the device. Without pre-delay, reverb starts at the same instant as the source and obscures its attack, which is why heavily reverbed vocals lose their intelligibility. Add twenty to forty milliseconds of pre-delay and the consonants arrive cleanly before the space opens up behind them. The vocal stays clear and sounds further forward even though there is just as much reverb.

The common reverb types are worth knowing by name because you will be asked for them. A hall is large, slow and lush. A room is smaller and faster. A plate is a dense bright tail with no obvious early reflections, historically made with a sheet of steel, and it is the classic vocal and snare reverb. A chamber sits between room and plate. A spring is the boingy sound built into guitar amplifiers.

Delay, and putting it in time with the music

A delay repeats the signal after a set time, optionally feeding some of the output back to its input so the repeats continue and fade. Its three controls are time, feedback and level, and the first of them is the one that decides whether it sounds intentional or accidental.

A delay whose time bears no relationship to the tempo is noise. A delay set to a note value -- a quarter note, an eighth, a dotted eighth -- locks to the music and becomes part of the arrangement. Most digital consoles let you tap the tempo, and it is worth re-tapping between songs rather than leaving a ballad's delay running under an up-tempo number.

The dotted eighth is worth singling out because it is the sound of a great deal of popular music: the repeat lands between the beats rather than on them, filling the gaps and creating movement without competing with the downbeat.

Feedback decides how many repeats you hear. One or two repeats read as depth; six or seven read as an effect; infinite feedback is a special effect that will fill the room and should be used on purpose and briefly. Live, be cautious: high feedback plus an open microphone is a path to a sound you did not intend.

Sends, returns, and why effects belong on buses

An effect can be inserted directly on a channel, replacing the dry signal with the processed one, or it can be fed from a send, with the processed signal returned on its own channel and blended with the original. For reverb and delay, the send-and-return approach is almost always correct, and understanding why is a genuine intermediate milestone.

The first reason is control: with a send, the dry signal and the wet signal have independent levels, so you can have exactly as much space as you want without changing the source. The second is efficiency: one reverb fed from twelve channels puts all twelve in the same space, which is what makes a mix sound like one performance in one room rather than twelve separate recordings.

The third reason is that you can process the return. EQ the reverb return -- high-pass it at 300 Hz and low-pass it at 8 kHz -- and the reverb stops muddying the low end and stops adding hiss at the top, while still doing its job. This one move improves more live mixes than almost anything else in this chapter.

Pre-fade and post-fade matters here. A post-fade send follows the channel fader, so when you pull a source down its reverb goes with it and the balance holds. A pre-fade send does not, which is what you want for monitor mixes and almost never what you want for effects.

Modulation, pitch and the rest of the rack

Beyond reverb and delay there is a family of effects based on modulating the signal, and you should know them by name because performers and producers will ask for them by name.

A chorus delays the signal by a few milliseconds and modulates that delay time, producing a thickened, shimmering version of the source. A flanger does the same with a shorter delay and feedback, producing the familiar sweeping jet sound. A phaser shifts phase rather than time, producing a subtler sweep.

A pitch shifter transposes the signal. Shifted a few cents and blended back in, it produces a doubling effect that thickens a vocal without sounding like an effect; shifted by a musical interval it produces a harmony. A harmoniser is a pitch shifter that follows a key so its harmonies stay musical.

For live work the honest guidance is that most of these are used sparingly, on specific songs, at the artist's request. The two you will use constantly are reverb and delay, and doing those two things well matters far more than owning the rest.

Reverb settings by intention

Match the decay to the tempo; use pre-delay to protect intelligibility.

Delay note values

Tap the tempo, then choose the subdivision that fits the arrangement.

How to do it

1. Put the reverb on a bus fed by sends, not inserted on a channel.
2. Set the sends post-fade so effects follow your mix moves.
3. Bring the return up until you hear the space, then back it off slightly.
4. High-pass the return around 300 Hz to keep the low end clear.
5. Low-pass the return around 8 kHz to remove hiss and harshness.
6. Set pre-delay so the source's attack arrives before the reverb.
7. Check it in the full mix, and check it again with the room full.

What goes wrong

Practice it

With a multitrack and one reverb and one delay available.

The Craft of the Mix

Everything up to here has been about individual tools on individual sources. A mix is the arrangement of all of them into something that sounds like one intentional thing. That is a separate skill, and it has an order, a set of dimensions, and a relationship to the room you are standing in.

The order a mix gets built in

A mix built in the wrong order fights itself all night. The reason is simple: every decision you make constrains the ones after it, so the decisions that constrain the most should come first.

Start with the thing the show is about. For most music that is the vocal; for a corporate event it is the presenter; for a conference panel it is whoever is speaking. Get that one source sounding right and sitting at the level the room needs. Everything else is then built around it, and you always know what you are protecting.

Then build the foundation: kick and bass, which together occupy the low end and must be made to agree rather than compete. Then the rest of the rhythm section. Then the harmonic instruments. Then whatever is left. At every stage, check that the thing the show is about is still clear.

The opposite approach -- starting with the drums, building a beautiful kit sound, and then discovering there is no room left for the voice -- is extremely common and extremely hard to recover from, because by then every drum decision is load-bearing.

• The most important source first, at the level the room needs.
• Kick and bass next, made to agree in the low end.
• The rest of the rhythm section around them.
• Harmonic and melodic instruments into the space that remains.
• Effects last, and less than you think.

Balance, panning and depth

A mix has three dimensions and an intermediate engineer arranges all three deliberately. Balance is level: how loud each source is relative to the others. Panning is left-to-right in space. Depth is front-to-back: how close or far each source seems.

Balance is the one everyone works on, and the one that matters most. If the balance is right almost everything else can be mediocre and the mix will work. If the balance is wrong, no amount of processing rescues it.

Panning in a live room is more constrained than in a studio, because a large part of the audience is not sitting between the loudspeakers. Anything panned hard will be missing entirely for some of the room. The convention is to keep the essential things -- vocals, kick, snare, bass -- in the centre, and to pan only what can afford to be uneven: overheads, keyboard spreads, guitars in a two-guitar band, effects returns.

Depth is created by three things: level, high-frequency content, and reverb. Quieter, duller and wetter all read as further away. This means you can push a source back without turning it down, by rolling off its top end and adding a little space -- useful when something needs to be present but not prominent.

Mixing the room you are actually in

The mix that worked in rehearsal is not the mix this room needs, and the mix that worked at soundcheck is not the mix the full room needs. An audience is an enormous quantity of soft absorbent material. A room with people in it has a shorter reverberation time, less low-mid build-up and a duller top end than the same room empty.

The practical consequence is that a mix which sounded slightly bright and thin at soundcheck is often exactly right at showtime, and a mix that sounded perfect in the empty room will sound dull and muddy once the doors open. Experienced engineers deliberately leave a little brightness in at soundcheck, knowing it will be absorbed.

Where you stand matters too. A mix position against a back wall hears reinforced low end that most of the audience does not, so an engineer mixing there tends to under-supply bass to the room. A position under a balcony hears a dull, contained version of the system. If you cannot move the mix position, walk the room during support acts and playback and build a mental correction for what your position is lying to you about.

Walk the room. It is the single most valuable habit in this chapter, and the one most often skipped because it means leaving the desk.

Masking, and why a loud mix can still be unclear

Masking is the phenomenon where a loud sound makes a quieter sound at a nearby frequency inaudible. It is not a matter of opinion or taste; it is how hearing works. Two sources occupying the same region do not add up to both being heard, they add up to one being heard and one being wasted.

This is why turning things up does not create clarity. If the guitar is masking the vocal, a louder vocal is masked by a proportionally louder guitar the moment you turn that up to compensate, and the mix arrives at the same relationship at a higher and more fatiguing level.

The answers are all about space rather than level: carve the competing region out of the less important source, use panning to separate them where the room allows, use depth so one sits behind the other, or -- most effective and least used -- ask whether both sources need to be there at all in that moment of the song.

Arrangement is a mixing tool. A source muted in a verse is not a failure; it is the reason the chorus sounds bigger.

Knowing when it is finished

A mix is finished when further changes stop making it better. That sounds glib, but the practical test is real: if you have made three consecutive changes and undone all of them, you are done and have been for a while.

The risk at the intermediate level is the opposite of the beginner risk. A beginner stops too early because they cannot hear what is wrong. An intermediate engineer keeps going too long, because they can now hear everything that is wrong and feel obliged to fix all of it. Some of it is not worth fixing, and some of it will be inaudible the moment there are eight hundred people in the room.

Take the ears out of it periodically. Listen at a genuinely low level for thirty seconds -- balance problems are ruthlessly obvious quietly and can hide behind loudness. Listen from a different part of the room. Listen to a reference track you know well. Then come back and decide whether the thing that was bothering you actually matters.

The three dimensions of a mix

Balance decides whether it works. The other two decide whether it is good.

What an audience does to a room

Expect these changes between soundcheck and showtime.

How to do it

1. Mute everything, then bring up the most important source alone.
2. Set it at the level the room actually needs, not a comfortable level.
3. Add kick and bass, and make them agree before adding anything else.
4. Add the rest of the rhythm section, checking the main source stays clear.
5. Add harmonic instruments into whatever space is left.
6. Add effects last, on sends, less than you think.
7. Walk the room and correct for what the mix position was hiding.

What goes wrong

Practice it

With a full multitrack, from silence.

System Tuning & Bigger Systems

Up to this point the book has been about what you do to signals. This chapter is about what happens after the signal leaves the console: how a loudspeaker system is made to agree with a room, what measurement actually tells you, and which problems EQ can solve and which it absolutely cannot.

Make the PA sound right in the room, and handle more complex systems.

Tuning a system to the room

A great mix still depends on the PA sounding right in the room, and system tuning is the process of adjusting the sound system so it delivers even, accurate sound across the audience — compensating for how the room and the speakers themselves colour the sound (the acoustics ideas from Acoustics Fundamentals, addressed at the system level). Tuning uses EQ and other tools applied to the whole system (not individual channels) to even out the response, so the mix you create translates well to every seat. Understanding that a system is tuned to the room — and that this is different from mixing individual channels — is key intermediate knowledge.

You do not need to master system tuning at this level (it is a deeper specialty often using measurement tools), but you should understand its purpose and place: before you mix, the system is set up and tuned so it reproduces sound accurately in that room; then your mix sits on top of a good-sounding system. A poorly-tuned system undermines even a great mix, while a well-tuned one lets your mixing shine. Recognizing the role of system tuning connects your mixing craft to the room and the PA it plays through.

Routing and digital workflow

Intermediate audio also means handling more complex systems and deeper console work. Routing — directing signals through more complex paths (multiple mixes, sub-groups, matrix outputs, effects routing) — becomes more involved on bigger shows, and understanding how to route signals flexibly is an intermediate skill built on the signal-flow foundation. Digital consoles, which dominate, offer deep capabilities: storing and recalling complete settings (scenes), so a show can be set up and recalled instantly; extensive built-in processing; and flexible routing in software. Working fluently with a digital console's routing, scenes, and processing is core intermediate craft.

These deeper capabilities let you handle bigger, more complex audio systems and shows efficiently. Sub-groups let you control related channels together; scene recall lets you manage complex shows or quick changeovers; flexible routing handles the many mixes a large show needs. The intermediate audio professional moves beyond a basic mix to managing a fuller system — more channels, more routing, more processing — using the console's depth. This systems-level fluency, combined with the mixing craft, is what intermediate audio develops.

What a tuning day is, and who does what

On any system larger than a pair of speakers on sticks, somebody tunes it before anybody mixes on it. On a touring production that is the system engineer, sometimes called the system technician or crew chief. In a house venue it may be the house engineer, and on a small show it may be you.

The work has a shape. The system goes up and is aimed. Each element is confirmed to be working and loaded with the correct manufacturer preset. The subs and mains are aligned in time. Any delay loudspeakers are aligned to the mains. Then the system is measured at several positions, an overall equalisation is applied, and the result is checked by listening to material the engineer knows intimately.

As a mix engineer your job is not to do this work but to understand it, because you inherit its results. A system that has been tuned well gives you a mix position that tells the truth and a room that sounds broadly the same everywhere. A system that has not will have you fighting problems on channel EQs all night that no channel EQ can solve.

The most valuable question you can ask a system engineer is simply "what did you have to do?" The answer tells you what the room is like, where the system is weak, and what to expect when the room fills.

Measurement -- magnitude, phase and coherence

System measurement is done with a dual-channel transfer function analyser -- in practice, almost always Rational Acoustics Smaart, with Meyer Sound's SIM and a handful of others in use. The principle is straightforward even if the mathematics is not: the analyser compares the signal sent to the system with the signal a measurement microphone hears in the room, and reports the difference.

That difference is reported three ways. Magnitude is the frequency response: which frequencies arrived louder or quieter than they were sent. This is the curve everybody looks at. Phase shows the timing relationship across frequency, and it is what you use to align elements. Coherence is the one beginners ignore and experts read first: it is a measure of how much of what the microphone heard is genuinely related to what was sent.

Coherence matters because where it is low, the other two traces are meaningless. Low coherence means the microphone is hearing reflections, noise, or a signal too quiet to measure -- so the magnitude curve at that frequency is describing the room's interference pattern, not the system. An engineer who equalises a dip that sits in a low-coherence region is equalising noise.

The rule that follows is simple and worth memorising: do not fix what the coherence does not support.

Where the microphone goes, and why one position is a lie

A measurement at one position describes that position. Move the microphone two metres and the trace changes, sometimes dramatically, because the interference pattern between direct sound and reflections is different there.

This is why measurement practice uses spatial averaging: take measurements at several positions across the area you are tuning and average them. What survives the averaging is what the system is genuinely doing; what cancels out was a property of one position. A problem that appears at every position is worth addressing. A deep null at one seat is a room artefact, and equalising it would ruin every other seat while failing to fix that one.

Placement also matters in the ordinary sense. The microphone belongs in the audience area, at roughly head height, away from walls and aisles, on a stand rather than in a hand, and pointed at the system element being measured. Measuring from the mix position alone is a common shortcut and a poor one, because the mix position is frequently one of the least representative places in the room.

Delay alignment -- the adjustment that changes the most

When two loudspeakers reproduce the same sound and their output arrives at a listener at different times, they interfere. At some frequencies the arrivals add; at others they cancel. The result is a comb-filtered response that no equaliser can repair, because the problem is in time, not in level.

Delay alignment fixes it by delaying the element that arrives first until both arrive together. This is arithmetic, not taste. The two places it matters most are between subwoofers and mains, and between delay loudspeakers and the main system.

Subwoofers and mains are often physically separated -- subs on the ground, mains flown several metres up -- so their output arrives at different times in the crossover region. Aligning them typically transforms the low end from vague and disconnected to tight and punchy, and it is the adjustment that most often makes people say a system "sounds bigger" when nothing got louder.

Delay loudspeakers covering the back of a long room must be delayed to match the sound arriving from the main system, or the audience under them hears everything twice. Sound travels roughly one foot per millisecond, or about three metres every ten milliseconds, so a delay position thirty metres behind the mains needs roughly ninety milliseconds. Most engineers then add a few extra milliseconds deliberately, so the main system arrives very slightly first and the ear localises the sound to the stage rather than to the speaker overhead.

What EQ can fix in a room, and what it cannot

This is the most important discipline in system work, and it is constantly violated. Equalisation changes level at a frequency, everywhere the system plays, for everyone. It is therefore the right tool only for problems that are the same everywhere.

EQ can fix: a system whose overall tonal balance is not what you want; a loudspeaker's inherent response; a broad tendency of a room to emphasise a region, when that tendency shows up at every measurement position. These are real and EQ is the correct answer.

EQ cannot fix: a reflection, a null caused by interference between elements, coverage that does not reach part of the audience, an arrival-time problem, or anything that differs between positions. A null from interference is a cancellation -- boosting it puts more energy into a cancellation and heats the amplifier while changing nothing audible. A reflection arrives later than the direct sound and EQ has no access to time.

The practical version: if the problem moves when you move, do not use EQ on it. Use aiming, delay, coverage, or absorption. Those are the tools that work in time and space, and EQ is not one of them.

Reading a transfer function measurement

Read coherence first. Where it is low, the rest of the trace is not evidence.

What is fixable, and with what

If the problem changes as you walk, it is not an EQ problem.

How to do it

1. Place the measurement microphone in the audience, on the centre line, at head height.
2. Measure the mains alone and note the arrival time.
3. Measure the subwoofers alone and note their arrival time.
4. Delay whichever arrives first by the difference between them.
5. Measure both together through the crossover region and check the result adds rather than cancels.
6. Invert the polarity of the subs and confirm the combined response gets worse; if it gets better, your alignment was wrong.
7. Confirm by listening to material you know well.

What goes wrong

Practice it

On any system with a measurement rig available, or alongside a system engineer.

Coverage, Subwoofers and the Shape of a System

The boxes above your head are arranged the way they are for reasons, and an engineer who understands those reasons makes better decisions about what can and cannot be fixed from the console. This chapter is about the shape of a system: why arrays curve, what subwoofers are doing on the floor, and what the manufacturer's processing has already decided for you.

Coverage, and why it decides the box count

The job of a loudspeaker system is to deliver roughly the same sound to every seat. Not the same loudness in the physical sense -- that is impossible across a long room -- but the same tonal balance and the same clarity, at a level within a few decibels across the audience.

That requirement, not loudness, decides how many boxes are required. A small club needs enough coverage to reach the back wall and the side walls without shouting at the front row. An arena needs a great many boxes not because it must be louder but because the distance between the nearest and furthest listener is enormous, and closing that gap takes carefully aimed elements at different angles and levels.

The enemy is the inverse square law: sound from a point source drops about six decibels every time the distance doubles. A listener at forty metres receives far less than one at ten. A line array partly defeats this -- over its useful range a long array behaves more like a line than a point and loses closer to three decibels per doubling -- which is precisely why arrays exist.

Why arrays curve

Look at a line array and you will see the boxes at the top angled only slightly apart, with the angles increasing toward the bottom. That curve is not decorative and it is not arbitrary.

The boxes at the top of the array are aimed at the furthest seats. Those seats are a long way away, so they need the combined energy of several boxes working together, which means those boxes must be nearly parallel. The boxes at the bottom are aimed at the nearest seats, which are close and need much less energy spread over a much wider vertical angle -- so those boxes splay apart.

The result is an array that delivers an even level from the front row to the back wall. When somebody says the system is "shaded" or "the splay is wrong", this is what they are talking about, and the symptom in the room is a level that changes noticeably as you walk from front to back.

Nothing on a console can fix an array whose splay is wrong, because the problem is different at every distance. This is worth knowing so that you recognise it and ask for it to be addressed rather than spending an afternoon on channel EQ.

Subwoofers, and the arrangements they are put in

Low frequencies behave differently from everything else: they are long waves, they are largely omnidirectional from a single box, and they go around obstacles rather than being stopped by them. That creates two problems. The first is that a row of subwoofers across the front of a stage interferes with itself, producing strong and weak regions across the audience. The second is that subs radiate backwards onto the stage, where the low end muddies monitors and makes life hard for musicians.

The arrangements you will meet are answers to those problems. A left-right arrangement, with subs under each main hang, is simple and common, and produces a pronounced interference pattern down the centre of the room. A centre cluster puts all the subs together in the middle, eliminating that pattern but concentrating low end centrally. A distributed or gradient arrangement spaces subs across the stage front with deliberate spacing and delay to even out the coverage.

Cardioid and end-fire arrangements address the backwards radiation. A cardioid arrangement reverses and delays one sub in a stack so that its output cancels behind and adds in front. An end-fire arrangement places subs in a line front to back, each delayed so their outputs arrive together forwards and scatter backwards. Both trade some forward output for a quieter stage, and on a loud show with in-ear monitors that trade is almost always worth making.

• Left-right: simple, common, interference down the centre.
• Centre cluster: even across the width, concentrated centrally.
• Distributed / gradient: even coverage, more boxes and more setup.
• Cardioid: quieter stage, some forward output traded away.
• End-fire: quieter stage, needs depth on the deck.

Presets and manufacturer processing

A modern loudspeaker is not a passive object with a response you equalise into shape. It is a system: a cabinet, drivers, an amplifier and a substantial amount of processing designed together by the manufacturer. The preset loaded into the amplifier or processor contains the crossover filters, the corrective equalisation, the phase alignment between drivers, the protective limiters and the delay for the specific configuration.

This is why loading the wrong preset is such a serious error. A box running its neighbour's preset may sound broadly plausible while its limiters are set for a different driver -- which is how equipment gets damaged.

It is also why "flat" is not the goal of system EQ. The manufacturer has already made the box do what they intended. Your system EQ addresses the room and the tonal preference of the show, on top of work that has already been done. An engineer who equalises a modern system to a ruler-flat measured response has usually made it sound worse, because a flat measured response in a real room is not what a good system sounds like.

The control platforms are worth knowing by name, because you will be asked whether something is patched into them: d&b audiotechnik use their own amplifier control; L-Acoustics use LA Network Manager with LA12X and LA4X amplified controllers; Meyer Sound boxes are self-powered and managed through Galileo processors; Lab.gruppen and Powersoft amplifiers with Lake or Armonia processing appear under a wide range of systems.

Subwoofer arrangements

Each one trades something. Know which trade this show made.

What the preset already decided

A modern system arrives partly tuned. Know what is already done.

How to do it

1. Ask which system it is, and which preset is loaded.
2. Count the boxes and identify mains, subs, fills and delays.
3. Confirm every element is actually producing sound.
4. Identify the subwoofer arrangement and what it implies for the centre of the room.
5. Play familiar material and walk the room front to back and side to side.
6. Note where level or tone changes most, and raise it with the system engineer.

What goes wrong

Practice it

At the next show with a flown system you can inspect.

The Digital Console in Depth

An intermediate engineer is expected to walk up to a desk they have never opened and be useful on it within the hour. That is possible because digital consoles are far more alike than they appear: they share an architecture, and once you understand the architecture the differences are mostly where the buttons are.

Layers, banks and why the desk is smaller than the show

A digital console has more channels than it has faders. A desk with twenty-four physical faders may be running ninety-six inputs, and the way it manages that is layers: the faders are a window onto the channels, and you slide the window.

Layers are usually fixed banks -- channels 1 to 24, then 25 to 48 -- while custom layers let you assemble any channels you like onto one set of faders. Building a custom layer with the eight things you actually touch during the show is one of the most useful habits in live mixing, because it removes the most common cause of live mistakes: changing the wrong channel because you were on the wrong layer.

The related danger is real and worth naming. Every experienced engineer has muted something on the wrong layer in front of an audience. The defences are to build layers that make sense, to look before you touch, and to use the desk's colour and naming tools properly rather than working from memory.

DCAs, groups and the difference between them

These two are constantly confused and they do different things. A group -- sometimes called a subgroup or audio group -- is a bus: the audio of the assigned channels is actually summed into it, and it passes through the group's own processing on the way to the output. A DCA, or VCA on older analogue desks, does not carry audio at all. It is a remote control for the faders of its assigned channels.

The practical difference matters. Put a compressor on a drum group and it compresses the summed drums -- that is bus compression, and it is the glue described earlier. Put the drums on a DCA and there is nothing to compress; pulling the DCA down simply pulls all the drum faders down together, preserving every relationship you built between them.

So: use a group when you want to process several sources as one. Use a DCA when you want to control the level of several sources while keeping their internal balance and their individual sends intact. Most live engineers use DCAs for control constantly and groups only where processing is genuinely wanted.

• Group: audio is summed into it; it can be processed as one signal.
• DCA: no audio passes; it moves the assigned faders together.
• Effects sends follow a DCA move only if they are post-fade.
• A channel can be in a group and on a DCA at the same time.

Matrices, and where a signal really goes

A matrix takes a set of sources -- usually groups and master buses -- and combines them into a new output with independent levels. This is how a modern desk feeds the many destinations a show actually has: the main system, the delay speakers, the front fills, a broadcast feed, a recording feed, a lobby feed, the dressing rooms.

The value is that each destination gets its own blend. The broadcast feed can have more vocal and less audience microphone than the room does. The lobby feed can be the whole mix at a quieter level with the sub content removed. None of that requires a second mix, only a matrix with sensible sources.

An intermediate engineer should be able to draw the desk's routing on paper: what feeds what, what is post-fade, what is post-processing, and which single fader change would take the broadcast feed off the air. Consoles hide this behind screens, and the only cure is to look.

Scenes, snapshots and recall safes

A scene, or snapshot, is a stored state of the console that can be recalled instantly. In theatre and corporate work they drive the entire show, changing routing, levels and processing cue by cue. In music they are more often used per song or per act.

The critical concept is scope: what a scene actually recalls. A scene that recalls everything will helpfully overwrite the gain change you just made to fix a problem, restoring the fault you were fixing. That is a bad evening.

Recall safes and scope filters solve this by excluding specified things from recall. The usual practice is to safe input gains, anything you expect to adjust live, and any channel whose state must persist regardless of cue. Setting these correctly before a show is not optional on a scene-driven event.

Every desk names these differently -- recall safe, scope, isolate, focus -- and the terminology is one of the real differences between manufacturers. The concept is identical everywhere.

Latency, and where it accumulates

Every digital process takes time. Conversion in and out, the console's own processing, any networked transport, any outboard the signal visits and returns from, and the system processor all add a small delay. Individually these are fractions of a millisecond to a couple of milliseconds. Together they add up.

Two situations make it audible. The first is a musician on in-ear monitors, who hears their own voice through the system at the same time as through their own head; more than a few milliseconds of delay there is uncomfortable and singers notice it immediately. The second is a signal that takes two different paths of different lengths and is then summed -- one path through an outboard device, one direct -- which produces comb filtering rather than delay.

The rule is to know your path and to avoid summing signals that took different routes. Most consoles compensate automatically for their internal paths; almost none compensate for a trip out to a piece of outboard and back.

Networked audio in a live rig

The stage box at the end of a copper multicore has largely been replaced by a network. Audio travels as data over standard Ethernet cabling, and a single cable carries what used to need a drum of multicore. The dominant system by a wide margin is Audinate's Dante, with AES67 as the interoperability standard that lets different systems talk, and MADI still in service as a robust point-to-point multichannel link.

What an intermediate engineer needs from this is not network engineering but literacy. Know that one device is the clock master and everything else follows it. Know that a network carrying audio should not be sharing a switch with a production office's file transfers. Know that sample rate must match across the system. Know that a patch in the network is separate from a patch on the console, so a channel can be correctly patched on the desk and still be silent.

That last point causes more confusion than anything else on a modern rig. There are two patches now: the network patch, which decides what audio arrives at the console, and the console patch, which decides which channel it lands on. Checking one and assuming the other is how an input goes missing at soundcheck.

Group or DCA?

The question is whether you want to process, or only to control.

Where latency comes from

Individually small, collectively audible -- especially on in-ears.

How to do it

1. Find the input patch and confirm what is arriving, on the network and on the desk.
2. Find the layer structure and build a custom layer with what you will actually touch.
3. Identify which DCAs and groups exist and what is assigned to them.
4. Trace the matrix outputs and note what feeds each destination.
5. Check scene scope and recall safes before touching anything else.
6. Save your own show file under your own name before you change a thing.

What goes wrong

Practice it

On any digital console you can access for an hour.

The Gear You Will Meet, By Name

Everything in this book applies to any equipment. But the job is spoken in product names, and an engineer who cannot name the desk in front of them starts every conversation at a disadvantage. This chapter is vocabulary. None of it is endorsement, and none of it is a claim that one maker is better than another -- it is simply what is in the trucks.

Why naming the gear matters

A production manager asks whether you are comfortable on a Quantum. A local crew chief says the house desk is a CL5 and the PA is a K2 rig. A colleague mentions the support act is bringing their own S6L. None of that is technical difficulty -- it is language, and an engineer who does not speak it appears far less capable than they are.

There is a practical dimension too. Knowing that a desk is from a particular family tells you roughly what to expect: how its processing is arranged, whether it is scene-driven, what its network options are, how its layers behave. Walking up to a desk already knowing its shape is most of the hour you would otherwise spend finding things.

Specialisation makes this sharper still. If you intend to work in theatre, in arena touring, in corporate events or in houses of worship, each of those has a set of equipment that dominates, and being fluent in that set is a substantial part of being hireable in it.

The consoles

At the top of touring you will meet DiGiCo, whose SD range and newer Quantum range are ubiquitous on large tours, and Avid, whose VENUE S6L is the other common choice and is especially prevalent where recording and live work overlap. Both are scene-capable, both are heavily networked, and both are desks people list on their CVs by name.

Yamaha's CL and QL series are probably the most widely installed digital consoles in the world, found in venues, houses of worship and corporate work everywhere; the Rivage PM series sits above them for large-scale work, and the smaller DM7 and DM3 cover compact jobs. If you learn one console family thoroughly, this is the one you will meet most often.

Allen & Heath's dLive is a serious large-format contender -- the dLive S7000 is the surface you will meet most -- with the smaller SQ series beneath it, and Midas -- long the name in analogue touring desks -- continues with the PRO series and HD96. Soundcraft's Vi series remains in many venues. At the accessible end, Behringer's X32 and the newer WING have put capable digital mixing within reach of almost every small venue and band, and you will meet them constantly on smaller shows.

• DiGiCo SD and Quantum -- large touring, theatre, broadcast.
• Avid VENUE S6L -- touring, and where live meets recording.
• Yamaha CL, QL, Rivage PM, DM7, DM3 -- the most widely installed family.
• Allen & Heath dLive and SQ -- large format and compact.
• Midas PRO and HD96 -- touring heritage, still widely specified.
• Soundcraft Vi -- common in installed venues.
• Behringer X32 and WING -- small venues, bands, budget productions.

The loudspeaker systems

Four names dominate large-scale live sound, and engineers refer to rigs by them constantly. L-Acoustics is arguably the most widely specified touring system in the world, with the L-Acoustics K1 and L-Acoustics K2 on large shows, Kara and Kiva as smaller arrays, and the LA12X amplified controller driving them. d&b audiotechnik is the other system you will meet everywhere, with the J and V series and the newer KSL and GSL ranges, driven by their own amplifiers -- the d&b audiotechnik D80 is the one you will see racked most often.

Meyer Sound systems are self-powered -- the amplifier lives in the box -- and appear across touring, theatre and installation, with LEO, LYON and the newer Panther on large shows and the long-serving UPA and UPJ family as fills. JBL's VTX series and Adamson's E and S series are both common on tours and festivals, and in the United States you will meet them often.

Below that tier, RCF, QSC, Electro-Voice and Yamaha cover the vast majority of small and medium venues, and a powered QSC K-series or RCF ART box on a stick is probably the most common loudspeaker in the world for small events.

• L-Acoustics K1, K2, Kara, Kiva -- with LA12X amplified controllers.
• d&b audiotechnik J, V, KSL, GSL -- with D80 and D40 amplifiers.
• Meyer Sound LEO, LYON, Panther, UPA, UPJ -- self-powered.
• JBL VTX and Adamson E and S series -- common on tours and festivals.
• RCF, QSC, Electro-Voice, Yamaha -- small and medium venues.

Measurement, processing and outboard

For measurement, Rational Acoustics Smaart is the standard by a wide margin, and "we'll Smaart it" is used as a verb on tuning days. Meyer Sound's SIM system is the other established platform. Measurement microphones come from Earthworks, DPA, NTi and Audix, among others, and they are chosen for flat response rather than character.

For system processing, Lake processing -- now found inside Lab.gruppen amplifiers and elsewhere -- is very widely used, as is Meyer's Galileo platform and the manufacturer-specific platforms already mentioned. Powersoft amplifiers with Armonia appear under a great many systems, the Powersoft X4 among the most common.

Outboard is increasingly software, but the names persist because the plugins are named after the hardware. The dbx 160 compressor, the Empirical Labs Distressor, the classic Universal Audio and Teletronix levelling amplifiers, the Waves plugin suites and the Bricasti reverbs are all spoken about by name, and a great many digital consoles ship emulations labelled after them.

For microphones, the intermediate engineer is expected to know the standards by name: the Shure SM58 and SM57 on vocals and snare, the Beta 91A in a kick drum, the Audix D6 and i5, the Sennheiser e935 and MD421, the Neumann KM184 as an overhead, and DPA's 4099 series clipped to instruments. For wireless and in-ears, the Shure Axient Digital AD4Q and ULX-D, Sennheiser's EW and Digital 6000, and Shure's PSM1000 in-ear systems are the names in constant use.

The consoles, by tier

Vocabulary, not ranking. Every one of these is a desk you may be handed.

The loudspeaker systems, by name

Rigs are referred to by these names constantly.

Microphones worth knowing by name

The standards. You will be handed these and expected to recognise them.

How to do it

1. Download the manufacturer's offline editor and run it on a laptop.
2. Build a show file for a band you know, from an empty session.
3. Patch inputs, build DCAs and groups, and set up two effects sends.
4. Create three scenes and set recall safes for gains.
5. Find the equivalent of every feature you rely on on your usual desk.
6. Arrive at the venue already knowing where things are.

What goes wrong

Practice it

Over the next month, at every show you attend or work.

Putting Intermediate Audio Together

The tools in this book are only as good as the ear directing them, and an ear is built deliberately rather than accumulated by accident. This chapter is about how the skill actually grows, what to practise, and how to protect the one piece of equipment you cannot replace.

Combine processing, mixing, and system skill into professional-sounding audio.

The complete intermediate skill set

Set out plainly, the intermediate engineer can do the following. They can set gain structure across a whole signal path and know where the headroom is. They can hear a tonal problem, name its region, confirm it by sweeping and correct it with the smallest move that works. They can set a compressor by intention rather than by preset, and say whether they are controlling or colouring.

They can place reverb and delay on sends, filter the returns, and use pre-delay to protect intelligibility. They can build a mix in a sensible order and get the balance right before reaching for processing. They can read a room, walk it, and correct for what the mix position is hiding from them.

They understand what a system tuning achieves and what it cannot, can tell a coverage problem from a tonal one, and know which problems to take to the system engineer rather than fight on a channel EQ. They can work on an unfamiliar digital console within the hour, and they can name the equipment they are working on.

That is the standard this book is written to. None of it is exotic, and all of it is ordinary competence at the level above beginner -- which is precisely why it is worth being genuinely good at.

Developing your ear and growing

The deepest intermediate skill is your ear — the trained listening that guides every processing and mixing decision. You develop it by mixing constantly, listening critically to your own mixes and to professional ones, and learning to hear what a sound or mix needs (where the mud is, what is too loud, what lacks space). Tools and knowledge matter, but audio is ultimately judged by listening, and a great mixer is above all a great listener. Deliberately developing your ear is what turns the techniques in this book into real craft.

From here, growing in audio means continuing to develop your ear and craft, mastering specific consoles and tools deeply, and moving toward advanced roles (front-of-house mixing, monitor engineering, system engineering) covered in the advanced books. The intermediate skills here — processing, mixing, system awareness, systems fluency — are the platform those advanced specialties build on. Keep mixing, keep listening, and keep refining, and you will continue from solid intermediate audio toward true mastery of the craft.

Practical sign-off

Intermediate audio is built mixing real material. The learner uses EQ, dynamics, and effects deliberately, builds a balanced polished mix, works with deeper routing/scenes, and develops critical listening. A mentor signs off when the learner mixes at a professional intermediate level.

• Used EQ deliberately to shape tone and solve problems
• Applied compression appropriately for control and consistency
• Built a balanced, clear, polished mix with tasteful effects
• Worked with deeper routing/scenes and demonstrated critical listening

Practising deliberately

Mixing more shows does not automatically make you better at mixing. It makes you faster and more comfortable, which is valuable, but improvement comes from practice that has a specific target and immediate feedback -- and a live show has neither.

The most effective single practice is frequency identification. Take a piece of music, apply a strong narrow boost at a known frequency, listen, then try to identify it without looking. There are training applications that do this systematically. Twenty minutes a day for a few weeks changes what you can hear permanently, and it is the difference between "something is wrong in the low mids" and "there is a ring at 250".

The second is virtual soundcheck. Record a multitrack of a show, then replay it through the desk and mix it repeatedly. You get the real material, the real console and unlimited attempts, which is exactly what a live show denies you. Most modern consoles will record and replay multitrack over their network with no extra equipment.

The third is analysis. Take a mix you admire and work out what was done: where the vocal sits, how much reverb is really there, what is panned and what is not. Then try to reproduce it. You will fail in specific and instructive ways.

Reference tracks, and knowing what right sounds like

An engineer without a reference is calibrating against memory, and memory for sound is unreliable over minutes, let alone between venues. A reference track is a recording you know intimately, played through the system you are about to mix on, to tell you what that system and that room are doing before you form any opinion about the band.

Choose three or four tracks and use the same ones for years. They should be recordings you have heard on hundreds of systems, ideally covering different territory: something with a clear vocal and acoustic instruments, something with heavy low end, something dense and loud. What matters is not that they are audiophile recordings but that you know them so well that any deviation is obvious.

Play them at show level, not quietly. A system behaves differently at volume, and a room's problems appear at volume. Then walk the room while they play.

Hearing protection, and the working life of an ear

Every skill in this book depends on hearing that you cannot replace, repair or insure. Noise-induced hearing loss is permanent, cumulative and completely preventable, and it is the most common reason audio careers end early.

Exposure is a product of level and time. A moderately loud afternoon is harmless; the same level for eight hours a day for twenty years is not. The damage is done long before you notice it, because the first losses are at frequencies you do not rely on for speech, and by the time conversation becomes difficult a great deal has already gone.

Wear proper protection. Not foam plugs, which remove the high end and leave you mixing a sound you cannot hear -- filtered musicians' plugs, which reduce level across the spectrum evenly and let you work normally. Custom-moulded ones are better again and are among the best money an audio professional spends. Wear them during load-in, during soundcheck when the drummer is alone, and during the support act.

Take quiet breaks. Your hearing recovers from moderate exposure given time, and a genuinely quiet twenty minutes in the middle of a long day does more than most people expect. Tinnitus after a show is a warning, not a badge.

Practice that actually improves the ear

Specific target, immediate feedback, repeated often.

Hearing protection

The instrument you cannot replace.

How to do it

1. Twenty minutes of frequency identification training, most days.
2. One virtual soundcheck session on real multitrack material.
3. Listen to one mix you admire and write down three specific decisions in it.
4. Play your reference tracks on any unfamiliar system before forming opinions.
5. Walk the room at every show you work, during playback or a support act.
6. Note one thing you would do differently, and why.

What goes wrong

Practice it

Starting this week.

What Goes Wrong at This Level

Beginner faults are loud and obvious: no signal, feedback, a hum. The faults that trouble an intermediate engineer are subtler, and they are frequently mistaken for taste or for equipment. This chapter is a diagnostic method and a catalogue of the specific things that go wrong once the basics are working.

A method for subtle faults

The beginner's troubleshooting method is to follow the signal path until the signal stops. That works beautifully for no-signal faults and not at all for the faults in this chapter, because the signal is present everywhere -- it is simply wrong.

The method that does work is to narrow the domain before narrowing the location. Ask first: is this a level problem, a tonal problem, a timing problem or a coverage problem? Those four have almost no overlap in their solutions, and identifying which one you are in eliminates three quarters of the possibilities immediately.

Then ask whether it changes when you move. A fault that follows you around the room is in the system or the mix. A fault that appears at one position and vanishes at another is the room, and is almost never fixable from the console.

Then ask whether it is constant or intermittent, and whether it correlates with anything -- a particular note, a particular performer, a particular moment in the song. Intermittent faults that correlate with musical events are nearly always dynamics or resonance; intermittent faults that correlate with nothing are nearly always connections.

• Which domain: level, tone, time or coverage?
• Does it change when you move? Then it is the room.
• Constant or intermittent?
• Does it correlate with a note, a performer or a moment?
• Does it survive a bypass of everything you have done?

The faults that look like taste

Several common problems present as "it just does not sound very good", which sends engineers on long unproductive hunts through their own EQ settings. They are worth recognising as faults.

A mix that is loud, tiring and unclear is almost always masking rather than a processing failure. A mix where the vocal will not stay audible however far it is pushed is masking too. A low end that is loud but has no definition is usually a time alignment problem between subs and mains, not an EQ problem. A system that sounds thin at the mix position and boomy elsewhere is a mix position problem.

A source that sounds hollow or phasey rather than merely wrong is usually comb filtering, from two microphones hearing the same source at different distances or from a signal summed with a delayed copy of itself. No EQ setting fixes comb filtering; moving a microphone does.

And a mix that sounded fine yesterday and wrong today, on the same system with the same band, is very often the room -- different audience, different temperature, different humidity, different door positions.

The faults that come from processing

Processing introduces its own failures, and an intermediate engineer needs to recognise their own handiwork.

Over-compression sounds flat, dense and lifeless, and it is seductive because it feels controlled. The test is to bypass everything and listen: if the raw version has more life, you have gone too far. Over-gating sounds mechanical, with sources appearing and vanishing; the fix is range rather than full closure. Over-equalising sounds hollow and small, usually the result of many narrow cuts that individually seemed reasonable.

Too much reverb sounds distant rather than spacious, and it is the most common single fault in the work of newly intermediate engineers. The test is honest and quick: mute all effect returns, and notice whether the mix improves. Frequently it does.

Phase and polarity problems from processing are subtler. Any signal that passes through a device and is summed with a version that did not will comb filter. This is why parallel processing should be done on a properly delay-compensated bus, and why routing a channel both directly and through an outboard unit is a trap.

When the fault is not yours

Some of what an intermediate engineer spends the evening fighting is not theirs to fix, and recognising that quickly is a skill in itself.

Coverage problems, array shading, wrong presets, amplifier channels in protection, a system asked for more level than it can produce -- all of these belong to the system engineer. A stage too loud for the room, a backline turned up beyond what the vocal can compete with, a drummer louder than the PA in the first ten rows: these belong to the stage and to a conversation, not to your EQ.

Raising these things is not passing the buck; it is accurate diagnosis. The engineer who spends an hour equalising around a problem that a two-minute conversation would have solved has wasted the hour and still has the problem.

The way to raise it is factual and specific: what you observe, where, and what you think it is. "The right hang sounds different from the left from about halfway back" is useful. "The PA sounds bad" is not.

Narrowing the domain

Identify which of the four you are in before looking for a location.

Intermediate faults and their real causes

These are the ones most often misdiagnosed as taste or equipment.

How to do it

1. Decide which domain it is in -- level, tone, time or coverage.
2. Walk twenty metres and decide whether it changed.
3. Bypass all your own processing and decide whether it is still there.
4. If it survives the bypass, it is the system, the stage or the room.
5. Check whether it correlates with a note, a performer or a moment.
6. State it factually to whoever owns that part of the system.

What goes wrong

Practice it

At the next three shows you work or attend.

Glossary

What the gear actually looks like

Every device named in this book, in one place. Recognition is the whole point of Chapter 4 — use this page to attach a picture to each name before you meet it on a dock at six in the morning.

Audio

Dynamic microphone

Condenser microphone

Wireless handheld

Radio rack

Mixing console

Stage box / snake

Power amplifier

Main loudspeaker

Subwoofer

Stage monitor wedge

In-ear monitors

DI box

Video

Camera

Video switcher

Scaler / processor

Projector

LED wall panel

Confidence monitor

Lighting & control

Conventional fixture

Moving-head fixture

Lighting console

Dimmer

Show controller

Media server

Networking & infrastructure

Network switch

Patch bay

Equipment rack

Power distribution

Cable tester

A properly coiled cable

Connectors, close up

XLR

TRS / TS ¼-inch

SDI on BNC

HDMI

Cat5e / Cat6 (RJ45)

Fiber optic

DMX (5-pin XLR)

AC power — not signal