First, what "mastering" actually means
Mastering is the final pass applied to a finished stereo mix — not to individual tracks. Its three jobs: bring the song up to a competitive loudness, even out the tonal balance so it translates across speakers, and control peaks so nothing clips on the way. It's a separate stage from mixing, and that separation is the key to using GarageBand well: GarageBand is a very capable place to mix, and a limited place to master. Doing each job in the stage built for it beats forcing one tool to do both.
Step 1: Mix first — it's most of what tutorials call "mastering"
Most of the audible improvement people get from "mastering in GarageBand" tutorials actually comes from mixing moves: balancing track volumes so no element drowns another, panning parts into their own space, and taming the one drum hit or vocal transient that towers over the rest of the song. Do this before anything else — it's free, it's fully supported in GarageBand, and no later stage can substitute for it. A buried vocal or a muddy arrangement is a mix problem, and finishing the song louder just gives you a louder version of the problem (we keep an honest list of what mastering can't fix for exactly this reason).
Step 2: The master track — what Apple actually gives you
GarageBand does have a place for whole-song moves: the master track (Track > Show Master Track). Per Apple's guide to the master track, you can control the project's overall volume, watch the clipping indicators in the master level meter, add an automatic fade-out (Mix > Fade Out), and automate volume across the song; Apple's help also covers using effects on the master track. Used gently, this is a real "pre-mastering" pass: set the output level high enough to be clean but never clipping, add your fade, and make small whole-song tone adjustments if something bothers you on every speaker you own. What this stage won't tell you is the thing mastering decisions hang on — how loud your song actually is in the units that matter. Which brings us to the export.
Step 3: Export right — WAVE, and what Auto Normalize really does
When the mix is done, Share > Export Song to Disk offers AAC, MP3, AIFF, or WAVE, per Apple's documentation. If a mastering pass comes next, pick WAVE or AIFF — lossy formats throw data away for good, and processing a lossy file bakes its artifacts in deeper.
One setting worth understanding rather than fearing: Auto Normalize (in GarageBand's advanced settings, on by default). Per Apple's guide, it exports projects at "the optimum loudness — that is, the highest volume level at which no distortion occurs." That's a peak-based promise: the file is raised until its single loudest instant hits the ceiling. Perceived loudness is a different quantity — density over time, measured in LUFS — which is why an Auto-Normalized export can still sound small next to a commercial track. If your exports keep coming out quiet, that whole mystery is unpacked in GarageBand Export Too Quiet? The Cause and the 3-Step Fix.
The one step GarageBand can't do
The missing step is deliberate loudness finishing: measuring the song's integrated loudness and bringing it — with controlled peaks and intact tone — up to the level where it stands next to commercial releases. GarageBand gives you no loudness measurement to aim with: its meters read level and peaks, not LUFS, and Auto Normalize stops at the first peak that touches the ceiling. Meanwhile the platforms you're exporting for think entirely in LUFS — Spotify normalizes playback to −14 LUFS, and other services do the equivalent. Finishing a song without a loudness number is driving without a speedometer: possible, but you're guessing, and "louder sounds better" bias makes your ears an unreliable judge of your own moves.
So the realistic ways to finish, from a GarageBand starting point: (1) Ship the Auto-Normalized export as-is — legitimate for demos and drafts; know that it may sit quieter and less dense than commercial neighbors. (2) Move to a DAW with metering and mastering plug-ins — the most control and the most to learn; overkill if GarageBand is otherwise everything you need. (3) Use a dedicated mastering tool on the exported WAVE — cloud services or a local app; we compared those two routes honestly in Cloud Mastering vs. Local Mastering, and the broader menu of options in How to Master a Song on Your Mac.
Where Tonekai fits (full disclosure: it's our app)
We build Tonekai for exactly the third route. Drop the WAVE you exported from GarageBand onto it, and on-device AI analyzes the track and handles loudness, tonal balance, and peak control automatically — then lets you A/B the result against the original at matched volume, which is the only honest way to hear what changed. Processing is entirely local: your unreleased song never leaves your Mac, and re-mastering after a mix tweak costs nothing per pass. It's free to try, and Pro is a one-time unlock, not a subscription — a sensible shape for GarageBand users, who already chose a free tool over a monthly one.
Download Tonekai on the Mac App Store (free to try)
FAQ
Do I need Logic Pro to master? No. Logic Pro adds the metering and plug-in depth that route 2 describes, and it's the natural upgrade if you want hands-on control of every stage. But if your only unmet need is the finishing step, a dedicated mastering pass on the exported file gets you there without learning a new DAW.
Can't I just put a compressor on the master track and call it mastered? You can compress the whole song in GarageBand, and gentle glue on the master track is a normal pre-mastering move. The problem isn't the tool, it's the blindfold: without a loudness measurement you can't tell "denser and competitive" from "squashed and worse," especially on the speakers you've adapted to. If you go this route, at minimum A/B against a commercial reference at matched volume — our ears-only checklist shows how.
Does this apply to GarageBand on iPhone and iPad? This post is about GarageBand for Mac — the master track and export details above come from the Mac documentation. The principle carries over (a bounce from any GarageBand is a finished stereo file that hasn't been loudness-finished), but check the iOS app's own export options before assuming they match.
Should I turn Auto Normalize off before exporting for mastering? It makes no practical difference to the final result — the mastering stage reads the level actually in the file and decides loudness deliberately either way. Don't lose sleep over this setting.
Wrapping up
"Mastering in GarageBand" is really three honest steps and one honest gap. Mix until nothing needs singling out (step 1), use the master track for level, fades, and gentle whole-song moves (step 2), export as WAVE with Auto Normalize demystified (step 3) — and then give the loudness-finishing step to a stage that can actually measure it. All of it can happen on the Mac you already own, and the free parts are genuinely free: GarageBand for the music, a trial mastering pass to hear what finishing adds. Let your ears make the call at matched volume.