What is audio mastering, and how is it different from mixing?
updated · November 2, 2026 · by Alex Pritsert
Audio mastering is the last stage of making a record. It takes a finished stereo mix and adjusts its tone, dynamics, stereo image and loudness so it holds up on any system and meets delivery specs. Mixing balances the individual tracks; mastering works on the whole song, usually from one stereo file.
Short answer
Mastering happens after the mix is finished. A mix decides how loud the vocal is against the drums. A master takes that balance as given and prepares it for the outside world: it corrects the overall tone, controls the peaks, sets the loudness and produces the files you upload.
It can't rebuild the balance, because it usually works from a single stereo file. If the snare is too loud, fix it in the mix. Streaming has also changed the goal. Spotify turns masters louder than -14 LUFS down to that level, so extreme loudness no longer wins. A good master now aims to sound right on any speaker and survive lossy encoding. You can hire an engineer, use an online service, or use a desktop app such as Tunr.
Mastering vs mixing: what each one changes
This article is for anyone finishing their first few releases at home and wondering whether the word mastering hides something they are missing. The difference comes down to what each stage can reach. Mixing has every track open, so it can turn down one hi-hat. Mastering usually receives one stereo file, so any change affects everything playing at that moment.
That limit explains most bad masters. When a kick is buried, boosting the low end in mastering also lifts the bass and the low notes of the vocal. The fix belongs in the mix, and part of a mastering engineer's job is telling you so.
| Question | Mixing | Mastering |
|---|---|---|
| Starts from | Individual tracks or stems | One stereo mix |
| Main concern | How the parts sit together | How the whole song travels |
| Typical tools | Faders, panning, per-track EQ, reverb, automation | Broad EQ, compression, stereo width, limiter, dither |
| Can it fix one loud snare? | Yes | Rarely, and with side effects |
| What comes out | A stereo mix | Release files with metadata |
What happens during mastering, step by step
Engineers order their tools in different ways. The fixed chain we use in Tunr makes a useful map, because each stage answers one question: gain, restoration, match EQ, parametric EQ, dynamic EQ, compressor, multiband, saturation, stereo, limiter.
Repair comes first. Gain sets a sensible working level, and restoration deals with noise, clicks and hum, because every later stage would make them louder. Tone comes next: match EQ moves the track toward a reference song, parametric EQ makes fixed corrections, and dynamic EQ acts only when a frequency gets out of hand, such as a harsh vocal note.
Dynamics follow. A compressor holds the overall level together, and multiband compression does the same for separate frequency ranges. Saturation adds harmonic density and the stereo stage adjusts width. The limiter goes last, since it is the stage that keeps peaks under a ceiling, and anything placed after it could push them back over.
A real master rarely uses all ten. If a track already sounds balanced, the right amount of EQ may be none.
How loud should a master be?
Loud enough for the music, and no further than your platform rewards. Spotify normalizes playback to -14 LUFS and reduces louder masters to that level. A master pushed to -7 LUFS comes out at the same playback level as one at -14, with less dynamic range to show for it.
Quiet masters don't always get the full lift. Spotify's own example is a track at -20 LUFS with a true peak of -5 dBFS, which is raised only to -16 LUFS because of its peak limit. Spotify also normalizes full albums as one unit, so the quiet song on your record stays quieter than the loud one.
For music, the AES streaming recommendation TD1008 suggests an overall -16 LUFS with the loudest song at -14. Our LUFS targets guide covers more platforms and club use.
What you should have at the end
A finished master is a set of files: a high-resolution WAV or AIFF for distribution, plus any MP3 or AAC copies you need. 16-bit files should be dithered, and the true peak should be checked again after any sample-rate conversion, since that step can change it. Distributors also ask for metadata such as an ISRC, the code that identifies a recording.
It helps to keep the numbers too. Integrated loudness, loudness range and true peak tell you whether the delivered file matches what you intended, and they settle arguments later with a label or a client.
Three ways to get a song mastered
A mastering engineer brings trained ears, a treated room and honest feedback about your mix. Online services take an upload and return a master quickly. Desktop software runs on your own computer. Our comparison of AI mastering and engineers looks at when each makes sense.
Tunr is the third kind. It is free, runs offline on macOS and Windows, and measures your track before it changes anything. You see the findings with their numbers, then a proposed chain with a plain-language reason for every step, and you can bypass or adjust any of it. The loudest 20 seconds render in about a second, so you hear the result before committing to a full export.
Frequently asked
Is mastering necessary for streaming?
Streaming platforms will play an unmastered mix, but they only adjust its overall level. They don't correct tone, control peaks or check how it survives encoding. If the song is going out under your name, a master, or at least a careful check of those things, is worth doing.
Can mastering fix a bad mix?
Only partly. Mastering can correct broad tone and control dynamics, but it works on the whole stereo file, so it can't turn down one instrument without affecting others. Balance problems such as a buried vocal or a loud snare are cheaper and cleaner to fix in the mix.
What is the difference between mixing and mastering in one sentence?
Mixing balances the separate parts of a song against each other, while mastering takes the finished stereo mix and prepares the whole song for release by adjusting its overall tone, dynamics and loudness and producing the delivery files that streaming services and distributors ask for.
Can I master my own music?
Yes, and many independent artists do. The risk is that you have heard the song hundreds of times, so you may miss problems. Comparing against a reference track at matched loudness and taking a day away before the final export both help. Measurements catch what tired ears miss.