Back to blog

The Pre-Mastering Checklist: 12 Checks Before You Send a Mix to Mastering

Before you send a mix to mastering, confirm five things: the bounce does not clip, there is no limiter or loudness maximizer on the master bus, integrated loudness sits roughly between −16 and −10 LUFS, the low end survives a mono check, and there is no accumulated mud or harshness. Then export a stereo WAV or AIFF at the session’s sample rate and at least 24-bit, with a little silence at each end. If the mix passes those checks, mastering is a short job. If it fails them, mastering cannot fix it.

Below are all twelve checks in the order worth doing them — the listening checks first, then the measurements, then the export. Each one says what to look for and what to do when it fails.

Why a checklist at all

Mastering works on one stereo file. It can lift the level, tilt the overall tone and control the peaks, but it cannot reach inside the mix to turn down one guitar or fix the phase between two kick mics. Every problem left in the bounce either survives mastering or gets louder. A checklist turns “I think it’s done” into a short list of yes-or-no answers, which is the only reliable way to judge your own mix after forty hours of hearing it.

Part 1: Listening checks

  1. Listen once, start to finish, at a quiet level. Low volume exposes balance problems that loud playback hides. If the vocal or the kick disappears when you turn it down, the balance is not done. Fix it in the mix — mastering will not separate elements that are masking each other.
  2. Check it on two other systems. Earbuds and a car or a small Bluetooth speaker are enough. A low end that vanishes on small speakers is a balance issue in the 60–150 Hz region; a top end that turns harsh only on earbuds usually points to the 2–5 kHz range. Write down what changes and fix it in the session, not on the master bus.
  3. Compare against two references, level-matched. Pick two released tracks in the same genre, turn them down until they are as loud as your mix, and compare tone, width and how the low end sits. Do not compare loudness — streaming normalization takes that away anyway. What is left is the real difference.
  4. Listen for the problems you have stopped hearing. Clicks at edit points, a breath that jumps out, a sibilant line, a reverb tail that cuts off at the end of the file. These are trivial to fix in the session and impossible to fix cleanly afterwards.

Part 2: Measurement checks

Ears get tired; meters do not. Run the bounce — not the live session — through a loudness meter, a correlation meter and a spectrum analyzer, or through a mix analyzer that reports all of them together. These are the numbers to check:

  1. No clipping, anywhere. The bounce must not contain a single clipped sample. The exact headroom number matters far less than the old “leave −6 dB” rule suggests; what matters is that nothing clips and nothing on the master bus is hiding it.
  2. No limiter or maximizer on the master bus. If you have been mixing into a limiter to hear the track loud, bypass it for the mastering bounce. A mix that only sounds right with the limiter on is a mix that is not finished. Gentle mix-bus compression or EQ that is part of the sound can stay; anything there to make it loud should go.
  3. Integrated loudness roughly −16 to −10 LUFS. This is the range Mozonic’s mix report treats as a sensible pre-master level. Much quieter and you are leaving resolution unused; much louder and the mix is almost certainly being limited already, so the mastering stage starts with no dynamics to work with.
  4. Mono compatibility: phase correlation at 0.3 or above. Sum the mix to mono and listen to the low end and the lead vocal. On a correlation meter, readings that sit at 0.3 or higher are what Mozonic’s report calls mono-compatible; readings near zero or negative mean parts of the mix cancel on a single speaker. The usual causes are stereo-widened bass, chorus on low instruments, and two mics on one source out of alignment.
  5. Stereo width between the extremes. Mozonic flags a stereo width reading below about 10 percent or above about 90 percent. Nearly mono means the mix will feel small next to its references; nearly all side signal means it will fall apart in mono and sound vague on headphones.
  6. Crest factor with some life in it. Crest factor is the gap between peak and average level. Mozonic flags a mix whose crest factor falls below about 4 dB, which usually means over-compression, or rises above about 20 dB, which usually means a few uncontrolled peaks. Mastering spends this headroom — a mix that arrives already flat gives it nothing to spend.
  7. No accumulated mud or harshness. Low-mid buildup between roughly 200 and 400 Hz and harsh energy in the upper mids are the two most common reasons a loud master still sounds amateur. They are usually the sum of several tracks each a little too full in the same band, so the fix is a few small cuts across the mix rather than one big cut on the master.

Part 3: The export

The last check is the file itself. A good mix exported badly still masters badly.

  • Format: a stereo WAV or AIFF. Do not send an MP3 to mastering if you can avoid it — lossy encoding throws away detail that the mastering chain then has to work around.
  • Sample rate: the rate the session ran at. Converting before mastering adds a step for no benefit; the mastering stage handles delivery formats.
  • Bit depth: 24-bit or 32-bit float. Avoid 16-bit at this stage — it is a delivery format, not a working one.
  • Dither: off. Dither belongs to the final bit-depth reduction, which happens after mastering.
  • Head and tail: a second or so of silence before the first transient and after the last tail has fully decayed. Never trim a reverb tail to save space.
  • Name it clearly: song title, mix version and sample rate in the file name, so the mastered file can be traced back to the exact bounce.

The checklist, short version

  • ☐ Balance holds at a quiet listening level
  • ☐ Checked on earbuds and one other small system
  • ☐ Compared against two level-matched references
  • ☐ No clicks, breaths, sibilance or cut-off tails
  • ☐ No clipped samples
  • ☐ Master-bus limiter or maximizer bypassed
  • ☐ Integrated loudness roughly −16 to −10 LUFS
  • ☐ Phase correlation 0.3 or above; low end survives mono
  • ☐ Stereo width neither near-mono nor almost all side
  • ☐ Crest factor roughly 4–20 dB
  • ☐ No low-mid mud or upper-mid harshness buildup
  • ☐ Stereo WAV/AIFF, session rate, 24-bit or float, no dither, silence at both ends

Run the measurement checks in one pass

Part 2 of this list is exactly what Mozonic’s free mix analysis measures. Upload a stereo bounce — WAV, AIFF or MP3, up to 500 MB — and the report returns integrated loudness against the −16 to −10 LUFS range, stereo width and phase correlation, crest factor, and mud and harshness readings, with a score and a specific suggested fix for each issue it finds. Free accounts get five analyses a month.

When the mix passes, Pro ($29 a month) carries it into the mastering stage in the same place: a mastering analysis with a 0–100 readiness score, guided workflows, and loudness-targeted renders to one of five profiles — Streaming default at −14 LUFS, Loud streaming at −11, Club loud at −9, Commercial loud at −8.5, and Podcast at −16.

And if you would rather the checking happened while you mix, Mozonic Studio is a free DAW for Apple silicon Macs that measures a session and proposes changes you can preview and accept or decline.

Analyze your mix free at mozonic.com before you send it anywhere — it takes a couple of minutes and tells you which of these checks the mix fails.

Pre-mastering: quick answers

How loud should my mix be before mastering?

Loud enough to use the available resolution, quiet enough that nothing is being limited. Roughly −16 to −10 LUFS integrated, with no clipping and no limiter on the master bus, is a sensible range. Reaching release loudness is the mastering stage’s job.

Should I leave −6 dB of headroom for mastering?

The specific number is a rule of thumb, not a requirement. In a 24-bit or floating-point file, what matters is that the bounce does not clip and no limiter is shaping it. Peaks at −3 dB or −6 dB are both fine.

Should I take compression and EQ off the master bus before mastering?

Take off anything whose job is loudness — limiters, clippers and maximizers. Gentle bus compression or EQ that you mixed into and that is part of the sound can stay, as long as you tell the mastering engineer it is there. Removing it at the last minute changes the mix you approved.

What file format should I send for mastering?

A stereo WAV or AIFF at the session’s sample rate, 24-bit or 32-bit float, without dither, with a little silence at the start and end. If you are using an online service, check its upload limits, but avoid sending an MP3 when a lossless file is available.

Can mastering fix a mix that fails these checks?

Mostly no. Mastering can raise the level and make broad tonal moves, but it cannot fix masking between instruments, phase problems between tracks, or a mix that arrives already limited flat. Fixing those back in the session is faster than fighting them afterwards.

Keep reading

Not sure which stage your track is stuck in? Mixing vs mastering explains what each one can and cannot fix.

On the headroom question in more depth: the −6 dB headroom myth.

If the mud check failed, why does my mix sound muddy? walks through the usual causes.

Passed everything? How to master a song at home covers the next stage step by step.