Eleven things decide whether a record works. The mix is one of them, it is the one everybody blames, and it is almost never the one at fault — so the useful skill is telling them apart.
A session goes technically right. Clean gain, no distortion, a balance you would defend, everything where you meant to put it. You play it back and it does not sound good, and the reason it does not sound good is that the song underneath it was not there. Almost nobody outside the room will be able to say that. They will just know something is off, and the word they will reach for is the mix.
That gap is the whole reason for this lesson. Most people do not separate the parts that make up a recorded piece of music — they hear one thing and judge one thing. Engineers do separate them, which is both the job and the reason engineers are worth hiring. Eleven distinct aspects decide whether a recording works. The mix is one of them. It is the one that gets blamed, and it is very rarely the one at fault.
Worth saying plainly, because it is a career fact rather than a technical one: this work happens whether or not anybody calls it production. Where there is no producer — most sessions — the engineer takes the role by default, and where there is one, they lean heavily on the engineer's critique. The engineers who get good at it tend to end up being called producers, and paid accordingly.
Every aspect has to clear a minimum. Not be excellent — clear a minimum. A single badly flawed one shows through as the weak link in the whole recording no matter how well everything else went, because a listener experiences the record as one object and hears the worst part of it.
Getting all eleven simultaneously excellent is close to impossible, which is a large part of why records that manage it are rare. But chasing that is not what the list is for. The list is a diagnostic. When something is not landing, it tells you where to look — instead of reaching for whichever aspect you understand best, which for most engineers means reaching for the mix because the mix is the part with knobs on it.
What counts as "good" in any of these is genuinely subjective and it moves, so each aspect below ends with the values people commonly hold for and against it — not as a standard, but as a map of where opinions tend to sit. Reading a particular band's position on that map is the subject of the next lesson.
The concept is the combination of the other ten. Call it the mood, the flow, the aura — it is the feeling conveyed most consistently and most strongly across the largest number of aspects. Some records have one so strong it survives a bad mix. Others have none at all, and that absence is itself the problem.
It is the only aspect you cannot work on directly. There is no concept fader. You change it by changing the others, which means the useful move is to notice when something does not fit — a bright modern reverb on a record whose entire identity is one room and one ribbon microphone, or a sound somebody loved on a different song imported into this one where it belongs to nothing. Pointing that out is the engineer's job even though nobody asked.
Working through how each aspect contributes to the whole surfaces inconsistencies, and quite often produces an idea nobody had. Valued: that a concept exists at all, that it is cohesive, that it has colour or complexity. Against: rote, uncreative, predictable.
A progressive succession of single tones of varying pitch over time — the linear aspect of music, as against harmony, which is the vertical one. Melody and rhythm are the two irreducible elements; you cannot have music without them. Harmony you can do without, and for a long time Western music did: monophonic writing was the principal form before roughly 1000 AD, and it survives in folk music and across many non-Western traditions. From about 1000 AD melodies were combined with other melodies, and that polyphony dominated composition until around 1600, when homophonic music — melody supported by harmony — developed in Italy and spread across Europe over the following century.
Practically, three things come up. If a melody is identical to another song's, say so — that is a copyright problem, not an aesthetic one. If a note sounds wrong, mention it, accepting it may be deliberate. And when a band improvises around the melody from the first chorus, suggest they play it straight once first, so the listener has something to recognise before it starts being decorated.
Valued: catchy, hummable, beautiful, interesting. Against: busy, banal, simplistic, annoying, chaotic.
The temporal element: duration, and stress or accent, whether irregular or organised into regular patterns. Most rhythm is metrical — its values are multiples of a basic unit, the beat. Free rhythm, common in a good deal of Asian music, has no meter at all, meaning its durations are not derived from any underlying unit. Polyrhythm, central to African music and to a number of twentieth-century composers, runs several patterns at once whose accents deliberately do not coincide. Rhythmic tension is one of the most direct routes to an emotional response there is.
You can critique rhythm without knowing any of that. The question that always works is: too busy or too simple? If you are falling asleep it needs spicing up. If you cannot keep up, something needs weeding out. Both judgements need ears rather than training, and both are worth saying out loud.
Beyond that: parts that step on each other, and variation between sections. If the guitar and the keyboard are fighting for the same rhythmic space, say so. If the drum pattern is identical in every section, ask whether that is intentional. And on tempo there is a reliable tell — listen to the vocal. The singer has to fit words into the time available, so a tempo that is fighting the song shows up in the phrasing before it shows up anywhere else.
Two or more tones sounding simultaneously, and the study of chords and their relations. Harmony was the last of the basic elements to develop — melody, rhythm and timbre all preceded it. Superimposed horizontal melodic lines prevailed until the 16th century, when vertical chordal construction was established, and in 1722 Rameau proposed that different groupings of the same notes are inversions of a single chord — the idea that turned a collection of sonorities into a system. Tonality became general through the 18th century, modulation followed from exploring how triads relate, and Liszt, Wagner and Strauss greatly expanded the vocabulary before some 20th-century composers discarded tonality altogether.
None of which you need in order to be useful. Most people can hear that a harmony part is wrong without being able to name what is wrong with it, and saying so is enough — the band can find the fix. Where an engineer can add something specific is in the number of parts and where they end up. A three-part harmony sung once and placed in the centre is one thing; the same three parts recorded twice and spread hard left and hard right is a different and much larger thing, made from exactly the same performance.
That principle scales further than most bands expect — a three-part harmony recorded repeatedly across many tracks produces a choir out of three singers, and it works because no two takes are identical, so the small timing and pitch differences between them create width no processing generates from one take. Even the minimum is worth doing: double the background vocals rather than recording them once.
Valued: multiple parts, an unusual chord structure. Against: too simple, too full, or an arrangement of chords that fights the song.
The aspect to tread most carefully around, for two separate reasons. The obvious one is that telling somebody their lyrics are weak damages a working relationship fast, and a remarkable number of successful songs have lyrics that will not survive being read aloud.
The less obvious one is legal. Contributing even a small amount of text to a song can make you a co-writer of it, which is why songwriters can be wary of accepting suggestions and why an engineer should be careful about offering them. Nothing about that is unreasonable — publishing is how songwriters get paid — but it means the safe and professional move is to raise the problem and let the band write the solution.
There is one observation you can nearly always make without going near authorship: do the lyrics sit rhythmically? Too many syllables crammed into a phrase, or too few so the line sags, is a question about fit rather than about writing. In some styles, rap most obviously, this is not a detail but the whole craft. Asking whether they are happy with how the words land against the music is a question, not a rewrite.
Valued: heartfelt, sincere, thoughtful, rhythmic, poetic. Against: banal, clichéd, corny, nonsensical.
Here arrangement means density — how many sounds are present at any single moment, and how many of them are in each frequency range. Not which instruments, which is the next aspect, and not what order the sections come in, which is the one after that.
It fails in both directions. Too thin is the easier problem: double or triple track a part, double it with a different instrument, ask for a busier part, record something in stereo with two microphones, or add time-based effects — delay, chorus, reverb — which create the impression of more without more actually being played.
Too full is the far more common problem. Some bands will record forty-eight tracks purely because forty-eight tracks are available, and then want all forty-eight audible. This is worth understanding properly because of what it implies: two parts fighting for the same frequency range at the same moment is an arrangement fact, and EQ can only referee it. Muting one of them for that section does what no processing can, and costs nothing. A large share of what people call mix problems are solved with the mute button, and engineers reach for the plugin first because the mute button feels like somebody else's job.
Whether each individual sound is any good, independent of what is played on it. The band chooses the instruments; the engineer is responsible for how they sound, which puts you in the awkward position of being accountable for decisions you did not make.
So the leverage is early. Recognise a bad source before it is recorded, when it can still be replaced. Rent a different kit rather than trying to rescue this one. New heads. Check the intonation on every guitar. There is no reason for one guitar and one amplifier to cover an entire album, and a genuinely useful trick is splitting one guitar into two different amplifiers with a Y cable and recording both to separate tracks — you get a sound that is not available any other way, and you get to decide the blend later. Learn the studio's synthesisers well enough to steer a band to the right bank instead of watching them scroll through five thousand presets.
When you cannot replace a bad source, say so anyway. Nothing in the control room fixes a drum head held together with tape, and if nobody says that out loud at the time, the shortfall becomes your mix's problem three weeks later. Naming it is not passing blame; it is making sure the right thing gets fixed.
The engineer is frequently the person in the building who knows most about what instruments and sounds exist, which is why producers lean on engineers here. Valued: unique, unusual, new. Against: the same thing again.
The order and the length of the sections — intro, verse, chorus, lead break, bridge, vamp. Normally not the engineer's place to comment unless you know the writer well, and normally not where a record is won or lost.
But if something concrete is wrong, ask. A five-minute lead break is a real observation. An introduction longer than about ten seconds is a real observation about how anyone deciding whether to keep listening actually behaves. The move that avoids the argument entirely is to suggest two versions — keep the one the band wants and edit a shorter one from the final mix — because nobody has to concede anything and the comparison makes the case by itself.
Valued: different from the norm. Against: too simple, or too normal.
The largest aspect, and the one the engineer is most often responsible for critiquing whether or not there is a producer in the room. Five separate things live inside it, and separating them is what turns "that take felt off" into something actionable.
Two levels of pitch perception. Perfect pitch is recognising the exact note or frequency of a sound cold — some people will tell you a tone is 440 Hz — and it is rare and, honestly, not the one this job needs. Relative pitch is hearing whether a sound is in tune with the sounds around it. That is common, it can be learned by anyone who has not got it, and it is what tuning actually means in practice.
Nearly everyone drawn to this work already has usable pitch; what they lack is speed and granularity. The skill to build is hearing the pitch of every note in a fast string of them, and then, harder, hearing the beginning, middle and end of a single sustained note as separate events — because a note can start in tune and drift. It comes down to concentrating on finer and finer moments, and concentration of that kind genuinely does get easier with practice.
The hard part is not hearing that something is wrong. It is remembering which note it was across a whole phrase. Being able to say whether it was flat or sharp is a bonus; knowing which note is enough to act on.
A few people hear tempo instability instinctively. Most of us have to work at it, and nearly everybody develops some physical method — patting a leg, a finger moving in the air, a slight tilt of the head. Whatever the method, hearing small tempo variation takes real concentration.
There are places it reliably happens. Drummers commonly speed up through a tom roll, so listen hardest at exactly the moment the fill starts. And bands rush the ending as they get going, which is why the last chorus of a live-tracked song is so often faster than the first.
Whether any of this matters depends on the band, and finding that out early is part of the job. Some are fanatical about stable tempo. Some do not care and should be gently pushed to care a bit more. And there is a fact about click tracks worth knowing before you offer one as the fix: a drummer who has not practised with a click for weeks cannot use one in a session. It will fight them. A genuinely useful move is to have them play to one for a while and then take it out of their headphones when it is time to record, which leaves their attention on their own timing rather than on a machine's.
The physical mechanics specific to each instrument, which change with the style being played. Nobody can be expected to know the right thing to say about every instrument, and the way you acquire it is by paying attention when an experienced musician says something that visibly works, and keeping it.
Concrete versions: a kick drum wants to be popped with the foot; some styles want the snare genuinely hit rather than played. Guitarists have a list of small things worth watching — string buzz, and the squeak of fingers moving along a wound string, which is inaudible in a room and enormous in a close-miked overdub. Singers respond to "sing out" or "project" when it arrives at the right moment and with some care about how it is said, and often what actually helps is getting them to support from the diaphragm rather than pushing from the throat. On bigger sessions a vocal coach in the room is a normal expense and not an admission of anything.
The general form of the skill: you do not need to solve it, you need to notice it and name it.
Two kinds, and they are worth checking separately. Volume dynamics — does the level move in a way that suits the song? Too much variation is distracting; too little makes a human performance sound like a drum machine, which is a real fault and not merely a stylistic preference.
Then emotional intensity, moment by moment. This one is checked less often and matters more. The classic failure is a singer who sings the first verse at the intensity the last chorus needs, leaving nowhere to go — the performance is committed and unmoving at the same time. The opposite happens too, where a take is technically fine and emotionally absent from the start. Dynamics is a shape across the song, not a level, and hearing it that way lets you place the peak deliberately rather than discovering where it landed.
The goosebump factor, and the one with no technical definition at all. Every box on this list can be ticked and a take can still be inert. The working rule is simple and hard to follow when everyone is tired: do not let a performance go until it does something to you. If every individual performance clears that bar, the whole record usually does.
Common values here are sincerity, heartfelt feeling, and emotion, but the specific ones vary by person and by genre — which is exactly the subject of the next lesson.
The problem with perfection is that it has no ceiling. Reach what you thought was the target and you immediately hear how it could be better. Somebody has to decide when to stop, and on most sessions that somebody is the engineer. What decides it is not the music — it is budget, deadline, purpose and the people involved, which is the whole content of the lesson that follows this one.
The engineer's own aspect, and the only one on this list that is unambiguously yours. Equipment quality means the recording chain rather than the instruments, which were the seventh aspect. And the part that actually matters is less about how expensive it is than whether it is in good working order — a well-maintained older console beats a new one with an intermittent channel every time.
Recording quality means sensible levels going in, microphone technique that suits the source, and an absence of distortion and noise that nobody asked for. All squarely the engineer's responsibility.
This is the aspect nobody praises and everybody notices. No listener in history has admired somebody's gain staging. Every listener hears hum, hiss, a clipped transient or a muddy capture, and forms an opinion about the whole record from it. It is judged exclusively in the negative, which is worth making peace with early. Valued: present and clean. Against: noisy, distorted, unclear.
One small part of what makes a recording work, and simultaneously the most powerful of the eleven — because the mix is the only aspect that can hide weaknesses in the other ten. A thin arrangement can be made to feel bigger. A performance that drifts can be tightened enough to pass. A dull instrument sound can be given some character it never had.
Which is exactly why it takes the blame. It is the aspect that could theoretically have concealed the problem, so when the problem shows through, the mix looks like the failure. Understanding that is worth more to your working life than any technique in this course: you will be told the mix is wrong, frequently, by people who are correctly identifying that something is wrong and incorrectly identifying what.
What drives the decisions in a mix varies song to song, though most often it is the lyrics and the performance — you build around what the record is trying to say and who is saying it. Arguably the best mixes are the ones where every aspect gets some say and none wins outright.
So the question worth carrying out of this lesson: which of the other ten should be determining the eleventh? Answer it and most mix decisions stop being preferences and start being consequences.
Before reaching for a plugin, name which of the eleven is actually failing. A mix can hide a weak aspect and it can never replace one — which is exactly why the mix gets blamed for problems it did not cause.
Students pay for getting unstuck: ask a concept question, routing issue, DAW confusion, or mix decision tied to this lesson.
Free readers can learn from the public Q&A archive. Paid students can ask their own lesson-specific questions and get mentor replies.