The Vocal Producer's Real Job: Coaching, Not Patching

There is a version of vocal production that lives entirely downstream of the performance. The singer sings, you keep what’s usable, and everything after that is repair: comp the best syllables, tune the drifting notes, ride the levels, patch the breaths. It works. It produces competent records. And it is the reason so many technically clean vocals land on a listener as nothing at all.

The job has two halves, and the industry only teaches one. The first half happens before the take exists — it is coaching, and it is the half that decides whether there is anything worth editing. The second half is editing, which has become non-negotiable for anyone calling themselves a producer. Full-stack producers now carry both, and most self-taught producers carry neither well, because the available instruction is almost entirely about the signal path.

Half one: the coaching layer

Coaching a vocal is not motivational. It is a set of specific, repeatable interventions that change what comes out of a person — and almost all of them are physical or contextual rather than musical.

Build the headphone mix like it’s the finished record. The first and most under-rated intervention, and it happens before anyone sings a note. A singer hearing a rough, thin, unbalanced mix in their headphones is being asked to imagine the record while performing it — which costs attention they need for the performance. Working vocal producers treat this as foundational: the cue mix should sound finished — the track balanced, the vocal sitting in it the way it will sit, enough space and enough energy that the singer is performing into a record rather than over a demo.

This is not politeness. It changes what comes out. Confidence is audible, and the cheapest source of confidence available to you is a singer hearing something that already sounds good with them in it.

(This is standard practice among working vocal producers — Simone Torres demonstrates the principle directly in her session breakdowns: build comfort with the mix, then spend your own attention on performance rather than on the singer’s doubt.)

Fix the room’s psychology first. A singer who is being evaluated sings differently than a singer who is working. The single most reliable performance improvement available to a producer costs nothing: stop talking about whether takes are good. Talk about what you’re trying next. “Let’s try one where you’re telling it to someone in the room” produces better results than “that was close, try again” — because the second sentence puts the singer in a job interview and the first puts them in a song.

Give physical direction, not adjectives. “More emotion” is unusable. What a singer can actually act on:

• Distance and proximity — step back for the loud line, lean in for the intimate one. This changes tone and dynamics simultaneously, at the source, in ways no plugin reproduces.
• Posture and body — sitting versus standing changes support; a hand on the chest changes awareness of it.
• Breath placement — telling a singer where to breathe reshapes a whole phrase, because breath position determines where the line’s natural stress lands.
• Consonant timing — asking for a consonant slightly early or late does more for a phrase’s feel than any timing edit you’d apply later.
• Syllable shaping — reshaping how a word is pronounced on a specific note, not just when. This is the same lever the pop vocal architecture → pulls at the writing stage with open vowels, applied live in the booth by a producer who can hear that a closed vowel is costing the note its size.

Performance first, technique second. The order matters and most producers invert it. Practitioners who do this professionally — the vocal-production teaching that comes out of rooms like Blackbird is explicit about it — describe chasing an honest, connected performance first and correcting technical detail after, using small adjustments to reconnect the singer to the story rather than to fix a note. A technically clean take from a disconnected singer is a repair job you’ll be doing for the rest of the record. A connected take with a flaw in it is a record with a flaw in it, which is a category listeners have never minded.

Direct the story, not the notes. The highest-leverage question in vocal production is not “can you hit that note better.” It’s “who are you talking to in this line, and what do you want them to do?” Specificity produces performance: not “sing it sad,” but “sing it like you already know they’re not coming back and you’re being polite about it.” That instruction changes vowel shapes, timing, and dynamics in one pass, without the singer thinking about any of them.

Manage the physical instrument. Warm water, a break before the belt takes, the hard lines tracked while the voice is fresh rather than at the end of a four-hour session. And the arrangement decision most producers never consider: change the key. A song sitting one semitone too high produces a strained take that the entire rest of the chain will spend hours failing to fix. Try it down a half step before you try anything else.

Know when to stop. After a certain number of passes, most singers get worse — more careful, less connected, tighter in the throat. The take you want is frequently take four, and the reason you don’t have it is that you kept going to take eleven. Print the keeper when you hear it and move on.

The layer nobody teaches: the first EQ happens before the signal

As the Craft Desk’s vocal-chain piece argues →, the physical world performs the first EQ. A singer’s throat, hydration, posture, distance, and word choice shape the sound before any processing exists — and the corollary for coaching is that most of what producers try to fix with plugins was decided in the room.

A thin vocal is often a distance problem. A harsh one is often a tension problem. A vocal that won’t sit is often on the wrong key or the wrong vowel. Reaching for an EQ is the correct move only after you’ve established that the source is what you actually want, and the difference between producers who work quickly and producers who fight their mixes usually traces to that habit.

Half two: the editing layer — the part that’s no longer optional

Editing used to be someone else’s job. Today, a producer who can’t edit vocals is missing a core competency, and the gap shows in the work immediately.

Comping is a musical decision, not a technical one. The best comp is rarely the collection of best individual moments; a comp assembled from twelve sources frequently sounds like twelve people. Prefer long stretches from a single pass, because the internal consistency of one performance — its breath rhythm, its intent, its arc — is a real quality that gets destroyed by micro-comping. Take the whole verse from take four even if take seven had a better word.

Tuning is repair, then effect — in that order. The pop vocal architecture → runs a two-tier system for a reason. Correct pitch manually first, preserving natural attack transitions and intentional drift; then, if the record wants the modern snap, apply the harder correction as an aesthetic layer on top. Reverse those and you get artifacts. Skip the first and you get the sound everyone complains about.

Timing: align, don’t quantize. Doubles and harmonies should sit tight against the lead — but tight to the lead’s feel, not to the grid. The lead vocal’s timing is the record’s timing. Aligning stacks to the grid instead of to the lead is how a stack starts sounding mechanical while measuring perfectly.

Breaths are performance, not noise. Deleting every breath produces an uncanny, airless vocal. The professional practice is to manage them: attenuate rather than remove, delete only on dense background stacks where twenty simultaneous breaths become a wash, and keep the ones that carry intent — the gasp before the big line is part of the take.

Sibilance is edited before it’s processed. On stacked backgrounds, cut or reduce sibilant consonants by hand so only the lead carries them. This single habit does more for a stack’s clarity than any de-esser, because the problem is phase-smearing across multiple sources rather than excess high frequency in any one of them.

Why these two halves became one job

Three forces collapsed the roles, and none of them are reversing.

Budgets. The sessions that once carried a producer, an engineer, a vocal producer, and an editor now frequently carry one person. That person is you.

Tools. The editing that required a specialist and a room now happens on a laptop — which democratized the capability and, in the same motion, removed the excuse for not having it.

Expectation. Contemporary vocal production is audible: listeners have absorbed thirty years of comped, tuned, stacked, aligned vocals and now hear an unedited one as unfinished, whatever their stated preference for authenticity. You don’t have to like that. You do have to work in it.

The consequence for a working producer: the two halves compound. A producer who coaches well needs less editing, and a producer who edits well knows exactly what to coach for — because nothing teaches you what a usable take sounds like faster than spending a year fixing unusable ones.

The honest counterpoint

Not every record wants this. A live band capture, a jazz vocal, a documentary-style folk record — these are damaged by the same treatment that makes a pop vocal enormous. The comping and tuning discipline described here serves a specific aesthetic, and applying it indiscriminately strips exactly the humanity that some records are made of. Know which record you’re making before you open the editor.

Coaching has a limit, and it’s an ethical one. You are directing a person’s body and emotional state to get a result. The line between skillful direction and manipulation is real, and singers know instantly which side you’re on. The producers people work with twice are the ones who treat a session as collaboration rather than extraction — and no take is worth being the other kind.

And the uncomfortable one: the editing layer is precisely where automated tools are advancing fastest. Comping, tuning, alignment, and breath management are all being automated competently right now. Which is exactly why the coaching half is the durable investment. A model can align your stacks. It cannot walk into a room, notice that a singer is holding their breath in their shoulders, and get them to tell the truth on take four.

Sources: Simone Torres on vocal production psychology and the cue mix (Audiopunks, ~0:15–0:45) · Inside a Real Vocal Production Session With Kim Wood Sandusky (Inside Blackbird, ~0:15–0:35). Both cited as documented professional practice by named working vocal producers. Last verified 2026-08-11.