Mastering is the final processing stage between a finished mix and a release. It sets overall loudness, tonal balance, and stereo width so a track holds up on earbuds, car speakers, and club systems. It polishes and translates a mix, and it is not a dependable rescue for one that was not working.
Mastering is the most misunderstood stage in production. Some artists skip it, figuring a good mix is enough. Others treat it as a rescue service for a broken mix. This mastering guide covers the middle ground: what the process does to a finished stereo file, the order the tools go in, how to think about loudness targets, and when paying someone beats doing it yourself.
A master takes a finished stereo mix and prepares it for playback outside your room. That means broad EQ moves so the tonal balance travels across speaker systems, dynamic range control so the track sits at a competitive level, and a last pass for technical problems that survived mixing. If you are still sorting out which stage does what, Mixing vs Mastering draws the line. For the production workflow that leads here, see Music Production Basics.
What Mastering Does to a Mix
Mastering is a short list of processing moves applied to one stereo file. Each move is small. The stack of them is not.
EQ (tonal balance)
A mastering EQ makes broad adjustments across the frequency spectrum. Where a mix engineer EQs individual tracks, a mastering engineer shapes the whole thing at once. A 1 dB cut in the low mids can clear up mud, and a small lift above 10 kHz can add air. These are rarely dramatic moves, and when a mastering EQ has to do heavy lifting, the mix usually was not finished.
Compression (dynamic range control)
Mastering compression evens out the dynamic range of the full mix. Something near 1.5:1 to 2:1 with a slow attack and auto release is a common starting point rather than a rule. The goal is cohesion: loud sections and quiet sections should feel like the same record. Too much flattens the energy, too little leaves the track unfocused.
Stereo width
Mastering can widen or narrow the stereo image. Mid-side processing lets you treat the center and the sides separately. Lifting the highs on the sides adds width and sparkle. Keeping low frequencies centered tightens the bottom end on most playback systems.
Limiting (loudness)
The limiter usually sits last in the dynamics chain, with dither after it if you are reducing bit depth. It sets a ceiling and raises the overall level toward a loudness target.
Which target depends on where the track is going, and platform guidance changes over time. Numbers repeated across mastering forums often drift from what the services actually publish, so check a figure at the source when a decision rests on it.
Spotify's artist support documentation, as of August 2026, says it normalizes tracks to -14 dB LUFS. The same page recommends keeping true peak below -1 dB TP, and below -2 dB TP when your master runs louder than -14 LUFS. That second figure rarely makes it into general mastering advice. For the platform-by-platform breakdown of loudness and format requirements, see Mastering for Streaming.
Pushing a limiter hard for maximum loudness can cost transients, and drums and vocals take the hit first. Where the destination normalizes playback, some or all of the extra level gets turned back down while the damage stays.
The Mastering Signal Chain
A typical chain runs in the order below. Typical is not mandatory. Plenty of masters use two of these six steps and nothing else.
Step | Tool | What it is for | Skip it when |
|---|---|---|---|
1 | Linear-phase EQ | Broad tonal corrections | The mix already translates across systems |
2 | Multiband compression | Taming one frequency range that moves | Nothing is misbehaving by band |
3 | Stereo bus compression | Glue and cohesion | The mix bus already did this |
4 | Mid-side EQ | Stereo field adjustments | The image is already sitting right |
5 | Limiter | Ceiling and final level | Seldom skipped, it sets the final level |
6 | Dither | Noise shaping on bit-depth reduction | You are not reducing bit depth |
Order matters more than tool choice. EQ before compression means the compressor reacts to the corrected balance instead of the problem you just fixed. Dither goes last, because any processing after it defeats the point. For a walkthrough of an actual session in this order, see How to Master a Song.
DIY, Automated Service, or an Engineer
Three routes, and the right one tends to change per release rather than per artist. Budget is the obvious input. Monitoring and objectivity are the ones people skip.
Your situation | Usually the better fit | Why | Cost shape |
|---|---|---|---|
First release, untreated room, borrowed headphones | Engineer or automated service | Hard to fix what you cannot hear | Per track, or a subscription |
Regular release schedule, treated room, monitors you trust | Yourself, with a reference check | Iteration beats revision turnaround | Plugins once, then nothing per track |
Album or EP that has to hold together | Engineer | Cross-track consistency is its own craft | Per track, usually with album rates |
Demo, sync pitch, or a rough for a label | Automated service | Speed matters more than nuance here | Per track, or a subscription |
A mix you have been living inside for weeks | Engineer, or a full day away first | Ear fatigue hides what fresh ears catch | Time, or a fee |
Mastering it yourself
You can master your own tracks with iZotope Ozone, FabFilter Pro-L 2, or stock DAW processing. After a one-time plugin purchase, there may be no per-track software fee, and you can iterate without waiting on a revision queue.
The risk is your own ears. You just spent hours inside this mix and your hearing has adapted to its balance. A problem that is obvious to someone else can be invisible to you. If you master your own work, leave a day between the final mix and the mastering session, then check the result on more than one system before you commit.
Hiring an engineer
An engineer brings fresh ears, a room that has been treated and measured, and monitoring most home setups cannot match. Rates vary by engineer, turnaround, and how many revisions are included, and album bookings usually price below the single-track rate. Ask for a current rate card rather than budgeting from a number in an article. For what to send and what to expect back, see Working with Mastering Engineers.
Automated services
LANDR and eMastered analyze an uploaded mix with automated processing and can return a master within minutes. Their plans differ and can change, so check each service's current pricing page.
These services analyze each uploaded track and build or select processing for that track, but the result still depends on whether the automated choices fit the artist's intent. Use the output as a reference and compare it at matched loudness before committing.
Preparing Your Mix
Whether you master it yourself or send it out, the file you start from decides how much room the process has.
Leave headroom. Engineers commonly ask for peaks somewhere around -3 to -6 dBFS, unnormalized, with the master bus limiter off. The exact number matters less than the principle: leave the next stage something to work with. If your engineer publishes a spec, theirs beats any general advice.
Export at session resolution. If you recorded at 24-bit/48 kHz, export at 24-bit/48 kHz. Downsampling before mastering throws away resolution the process could have used, and format conversion is usually handled at the end.
Decide about master bus processing. If you mixed into a bus compressor and the mix depends on its character, leave it on and say so. If you put a limiter there to make the mix feel louder while you worked, take it off.
Listen for problems first. Solo each track and listen for clicks, pops, hum, and noise buried under the full mix. Compression and limiting bring that material forward, and it is far cheaper to fix in the session than after.
Albums and EPs
Single mastering and album mastering are different jobs. An album master needs consistency across tracks, so a listener is not reaching for the volume between songs, and the tonal character holds even when the songs sit in different genres.
That means working in sequence and setting each track's level and EQ relative to its neighbors instead of in isolation. It also means deciding the gaps deliberately. Two seconds is a common default, and some transitions want less or more.
If you are running releases yourself, or coordinating them for an artist you manage, an album is the strongest argument for hiring someone. Cross-track consistency is hard to hear when you have been living inside the material for months.
Where Masters Go Wrong
Mastering a mix that was not ready. Mastering often brings existing problems forward. A buried vocal or muddy low end is usually better fixed in the mix, although some mastering tools can make limited balance corrections from a stereo file.
Chasing loudness for its own sake. Slamming a limiter to reach a less-negative LUFS figure can cost transients. Where a destination applies loudness normalization, some or all of the added playback level may be turned down while the lost dynamics remain.
Skipping the A/B. Compare the master against the unmastered mix at matched loudness, and against a commercial reference in the same genre. Matching the loudness matters, because louder tends to sound better for the first ten seconds. If the master does not hold up in that comparison, it is worth another pass.
Frequently Asked Questions
How loud should my master be?
It depends on the destination. Spotify's documentation, as of August 2026, says it normalizes to -14 dB LUFS and asks for true peak under -1 dB TP. Check each platform's own current guidance.
Can I master my own music?
Yes, if your monitoring is trustworthy and you can hear the mix the way a stranger would. Leave a day between finishing the mix and mastering it, then check the result on more than one system.
What file format should I deliver?
Usually your highest-quality native master, a 24-bit WAV at your session sample rate. Check your distributor's current specifications before converting, since downsampling by default throws away resolution the platform can use.
Does mastering make a song sound better?
A good master usually sounds more polished, more consistent across playback systems, and louder at the same volume setting. Some balance problems can be reduced during mastering, but a mix revision usually provides more direct control.
Read Next
How to Plan a Music Release is the next step once the master is signed off and the release needs a date, a timeline, and a plan behind it.