What Is Mastering in Music? Mixing vs Mastering, Explained

Published October 20, 2026

The short answer. Mixing combines the separate tracks of a recording into one stereo file: levels, panning, EQ, effects. Mastering is the last step after that: it works on the finished stereo mix as a whole, adjusting overall tone and loudness so it plays well on every system, and prepares the final files for release. On streaming, extra loudness matters less than it used to, because most services turn loud tracks down.

"Mixed and mastered" travels as one phrase, so it is easy to think of it as one job. It is two jobs, done in order, on different material, with different aims. Here is each one in plain words, where they meet, and the question most beginners ask next: how loud a finished song should be.

What is mixing?

Mixing is the stage where many separate recordings become one. A song recorded in a DAW (a digital audio workstation, the recording software on your computer) is usually a stack of tracks: a vocal, a second vocal, drums split across several microphones, bass, guitars, keys. Mixing is, in the words of the Wikipedia article on audio mixing, "optimizing and combining multitrack recordings" into a final mono, stereo or surround product.

In practice, the person mixing does four kinds of work on each track:

• Balance. How loud each part is against the others. This is most of the job.
• Placement. Panning each part left, right or center in the stereo field, so parts that compete for the same space can sit side by side.
• Tone. EQ (equalization, turning particular frequency ranges up or down) and compression (evening out the gap between a part's quiet and loud moments).
• Space. Reverb, delay and other effects that make a dry recording sound as if it happened in a room.

The key fact about mixing: you can still reach every individual part. If the snare is too loud, you turn the snare down. Nothing else moves.

What is mastering?

Mastering starts where mixing ends. It takes the finished stereo mix, a single file, and prepares it for release. The Wikipedia article on mastering describes it as the step that transfers audio from a final mix to a master, the source for every copy that follows, and says mastering engineers use EQ and compression to "optimize sound translation on all playback systems". Berklee College of Music's careers page calls it "the final creative step before release", when a record's overall dynamic and spectral balance is "sculpted and polished", with the aim of a release that works everywhere from vinyl and streams to earbuds.

The key fact about mastering is the opposite of mixing: you can no longer reach the parts. If the snare is too loud at the mastering stage, there is no snare fader. Every move changes the whole song at once. That is why mastering moves tend to be small and broad, and why a problem that lives inside the mix is best fixed in the mix.

Mixing vs mastering at a glance

Why mastering exists at all

Mastering began as a technical necessity, not a polish. According to the Wikipedia history of mastering, early recordings were cut straight into wax or soft-metal discs, and electrical cutting lathes arrived in the mid-1920s. Once magnetic tape spread after the Second World War, the session and the disc cutting could happen separately, and a specialist engineer began taking the final mix tape, adjusting it, often with EQ, and cutting the master disc from it.

The format still matters today. The same article notes that a vinyl master may need its dynamics reduced or parts of it folded down to mono. A CD needs its start, end and index points defined. That is the part of mastering that is about the medium, and it explains why one mix can produce more than one master.

What happens in a mastering session, in order

The Wikipedia article, citing a study of mastering practice, lists the common tasks. Here they are, grouped in a typical order:

1. Clean up. Edit out clicks, noise or small flaws.
2. Shape the tone. Broad EQ moves: a little less low mud, a little more air.
3. Shape the dynamics. Gentle compression or expansion, and a peak limiter that stops the loudest moments from going over the top of the digital scale.
4. Adjust stereo width where it helps.
5. Sequence the release. For an EP or album: the track order, the silence between songs, fades in and out, and matching the level of each song so none jumps out.
6. Embed the codes. ISRCs (the codes that identify each recording; see our guide to ISRC codes) and, for a CD, CD-Text.
7. Dither and deliver. Dither is a small amount of deliberate noise that masks the distortion created when audio is reduced to a lower bit depth, such as 16-bit for CD, which is why it is often one of the last steps. Then the files go out in the format each destination needs.

Step 5 is the one beginners underrate. For an EP, it is much of the reason to master: songs mixed on different days rarely match in level or tone until someone makes them.

How loud should a master be for streaming?

First, how loudness is measured

Peak level (how high the single loudest moment goes) says little about how loud a song feels. Streaming services instead measure integrated loudness, an average over the whole track weighted toward how the ear hears, in units called LUFS (loudness units relative to full scale). The LUFS article traces the method to the ITU-R BS.1770 standard, whose revision BS.1770-5 was released in November 2023. The numbers are negative: -8 LUFS is much louder than -14 LUFS.

What Spotify says it does

Spotify publishes its rules on its loudness normalization page for artists. As checked in October 2026, it says:

• The standard playback level is -14 LUFS. Premium listeners can choose Loud (-11), Normal (-14) or Quiet (-19).
• Louder masters get negative gain so they play at -14. Softer masters get positive gain, but only as far as the track's peaks allow: its own example is a track at -20 LUFS with a true peak of -5 dB, which is lifted only to -16 LUFS.
• Whole albums are adjusted as one, so a quiet song stays quieter than the loud one beside it. Tracks are adjusted one by one in shuffle and playlists.
• The files are not changed. Loudness is measured at upload and the adjustment happens during playback. The web player and third-party devices such as TVs and speakers do not normalize.
• Its advice to artists: aim for -14 LUFS integrated, keep true peaks below -1 dB, and below -2 dB if the master is louder than -14 LUFS, because loud tracks distort more when they are encoded to compressed formats.

A worked example

Take two masters of the same song. Master A is pushed hard with a limiter to -8 LUFS. Master B is left at -14 LUFS with more room between its quiet and loud moments. On Spotify's Normal setting, by the rule above, Master A is turned down by about 6 dB (from -8 to -14) and Master B plays as it is. They now play at the same average loudness. The difference a listener hears is no longer that A is louder. It is that A has had its peaks flattened to get there, and B has not.

That is why the "louder is better" instinct has weakened. The Wikipedia article on the loudness war describes the decades in which records were mastered ever louder, sometimes to the point of audible distortion; Metallica's Death Magnetic (2008) drew wide complaints for exactly that. It also notes analysis suggesting the trend peaked around 2005, and that many engineers still master loud because they assume some playback will not be normalized. Both things are true at once.

Other services

Not every platform publishes its numbers on a page we could read in this research. The loudness war article compiles reference levels from the services' own documentation and from mastering engineers. Treat these as a guide, not a spec, and check each service's current documentation:

The same article reports that Amazon, Tidal and YouTube turn loud tracks down but do not turn quiet ones up. So a very quiet master can play quieter than its neighbours there.

Is -14 LUFS a rule?

No. It is Spotify's playback level and the target Spotify suggests, not a law of mastering. The Audio Engineering Society's streaming recommendation, TD1008, sets a "distribution loudness" for music of -16 LUFS. Production Advice, the site run by mastering engineer Ian Shepherd, explained in October 2021 that the term was renamed from "target loudness" precisely because people read it as a mastering target, which it is not: it is aimed at streaming services, not at the people making records. Mastering louder than -14 is a choice of taste. What you give up is dynamic range, and on a normalized service you no longer get extra volume in return.

How to prepare a mix for mastering

Whether you hand the mix to an engineer or master it yourself, the file you start from decides how much is possible. Two published guides, one from plugin maker Waves (February 2024) and one from Chicago Mastering Service, agree on most of it:

• Leave headroom. Do not let the mix peak at the top of the scale. Waves suggests peaks between -2 and -4 dBFS for its own online mastering. Chicago Mastering Service asks for at least 2 dB of peak headroom. Engineers differ on the exact figure, so ask yours.
• Take the limiter off the mix bus. Waves advises removing any EQ, compression or limiting you added only to make the rough mix sound finished, because that is the mastering stage's job. Light "glue" processing that is part of the mix's sound can stay.
• Export at the session's own settings. Chicago Mastering Service prefers a 24-bit stereo WAV at the native sample rate of the project, and asks you not to resample. Waves says the same about matching the project unless your engineer asks otherwise.
• Do not dither. Waves notes dithering belongs to mastering.
• Do not trim the ends. Leave a few seconds before and after the song, so reverb tails and fades are not cut. Chicago Mastering Service asks for 2 to 3 seconds either side.
• Send notes. Known problems, the track order and gaps for a multi-song release, and one or two reference records you like the sound of.

Do you need a mastering engineer?

It depends on what you are releasing and what you can hear. A few questions settle it for most people:

• Is it one song or a collection? A single song to streaming needs the least. An EP or album gains most from someone matching levels and tone across the set.
• Is a physical format involved? Vinyl and CD bring format-specific work, described above, that is easy to get wrong without experience.
• Can you trust your room? Mastering decisions are small, and they are only as good as the speakers and room you make them in. The Wikipedia article lists low-distortion monitors and an acoustically optimized room among a mastering engineer's usual equipment for a reason.
• Have you heard it too many times? After weeks of mixing, it is hard to hear your own song fresh. A second set of ears, even a friend's, is part of what a mastering engineer offers.

Mastering your own first release is a sensible way to learn. So is paying for one song and listening closely to what changed. Neither guarantees a "professional" sound, and no master rescues a buried vocal. That fix belongs in the mix.

Common mistakes

• Judging changes at different volumes. A version that is a little louder almost always sounds "better" at first. Waves advises matching levels before comparing your mix with a reference track.
• Clipping the true peaks. Spotify's guidance is to keep true peaks below -1 dB, because peaks can grow when audio is encoded.
• Listening on one system only. Waves suggests checking on earbuds, a car stereo and a small Bluetooth speaker, the kinds of places most people will hear it.

Common questions

What is a premaster?

The finished stereo mix you send to mastering. It has headroom, no final limiter and no dither.

Can a master be too quiet?

On Spotify, a softer master is turned up, but only as far as its peaks allow, so a very quiet, very peaky master may still play below -14 LUFS. On services that do not turn quiet tracks up, it will play quieter than its neighbours.

When your song is mixed, mastered and out, you can submit it at For Artists. We check only that it is genuine, yours to share and safe, never whether it is good.

Send us your music

Send your song to Life of Art. We never judge the music: once a quick check confirms it is genuine and yours to share, it goes out to listeners every morning, and you see how many it reached and what share of them liked it. See how submitting works.