Audio Mastering Guide · Updated for 2026
How to Master a Song: A Step-by-Step Guide for 2026
Mastering is the final quality-control and finishing stage between your mix and release. This guide gives you a repeatable process for improving tonal balance, controlling dynamics, setting final loudness and exporting a master that translates across headphones, cars, phones and streaming services.
How do you master a song?
Export a clean stereo mix, open it in a fresh mastering session, compare it with a level-matched reference, correct broad tonal problems with EQ, control movement with light compression or dynamic processing, check stereo and mono compatibility, raise the level with a true-peak limiter, then quality-check and export a high-resolution WAV. Every processor should solve a problem you can hear; none is compulsory.
What is mastering—and how is it different from mixing?
Mixing balances the individual tracks inside a song: vocals, drums, bass, instruments and effects. Mastering works on the finished stereo mix and prepares that complete recording for release. A mastering move affects the entire song, which is why small decisions matter.
| Stage | Material | Main job | Typical tools |
|---|---|---|---|
| Mixing | Individual tracks and buses | Balance, separation, depth and emotion | Faders, panning, EQ, compression, reverb, delay |
| Mastering | Finished stereo mix | Translation, tonal balance, dynamics, final level and delivery | Metering, EQ, compression, saturation, stereo control, limiting |
Mastering cannot reliably rescue a buried vocal, a clipping 808 or reverb printed too loudly. If the problem belongs to one instrument, fix the mix whenever possible. For a vocal-specific workflow, use the rap vocal mixing and mastering guide.
Before mastering: prepare the mix correctly
A good master starts with a finished mix. Mastering should refine the record, not become a second mixing session.
Export lossless audio
Use WAV or AIFF at the mix session’s native sample rate and bit depth. Do not convert to MP3 before mastering if the lossless mix is available.
Prevent clipping
The stereo mix should not clip. There is no mandatory “minus 6 dB” rule; the requirement is clean headroom and a mix bus that is not being damaged.
Keep intentional processing
If mix-bus compression or saturation defines the sound, keep it. If a limiter was added only for loudness, export an un-limited premaster and keep the limited version as a reference.
Print clean starts and endings
Remove accidental count-ins, clicks and unused silence. Preserve natural reverb and delay tails; final fades can be refined during mastering.
Take a listening break before you begin. Fresh ears make broad tonal problems easier to hear and reduce the urge to process a mix simply because plugins are available.
How to master a song in 10 steps
1. Create a fresh mastering session
Import the stereo mix into a new project at the same sample rate. Leave the track at its original tempo unless tempo-synchronised processing requires it. Route the song through a dedicated mastering chain and place your meters after the final processor so they report the delivered result.
2. Choose one or two reference tracks
Use professionally released tracks with a similar genre, arrangement and low-end profile. Route references around your processing chain. Match their playback loudness to your work before judging bass, brightness, punch or width; louder audio can seem better even when it is not.
Do not copy a reference curve blindly. Use references to expose large differences and to decide whether those differences help or hurt your song.
3. Listen all the way through and take notes
Before touching a control, listen from the first transient to the final tail. Note moments where the vocal disappears, bass blooms, cymbals become sharp, the chorus loses impact or the stereo image changes unexpectedly.
Check meters, but let the song define the work. A spectrum analyser can confirm excessive energy; it cannot tell you whether the emotion is right.
4. Set gain and clean the edit
Use clip gain or a trim plugin to feed the chain at a sensible level. Repair obvious clicks and edit noise. Set the opening space, ending and fades deliberately.
Do not use the limiter as an input-gain repair tool. A clean gain structure makes every later decision easier to hear.
5. Correct broad tonal problems with EQ
Start with wide, restrained moves. Reduce genuine rumble, boxiness, harshness or excessive brightness only when you hear it. If the entire master needs extreme EQ, return to the mix.
Use narrow cuts for specific resonances and broad curves for overall colour. Repeatedly bypass the EQ at matched loudness. The processed version should feel clearer or better balanced—not merely different.
6. Control dynamics only as much as needed
A mastering compressor can gently connect the mix, restrain peaks or shape groove. Use a low ratio, choose attack and release by ear, and watch what happens to the kick, snare and vocal consonants.
If compression makes the chorus smaller, the snare duller or the groove pump unintentionally, back off. Dynamic EQ or multiband compression can control a frequency area that becomes excessive only at certain moments, but both can create an unnatural master when overused.
7. Add colour only when the song benefits
Subtle saturation, tape or clipping can add density and tame brief peaks before the limiter. This is optional. Listen for smeared transients, brittle high frequencies and low-end fuzz.
Colour is a creative choice, not a requirement for a “professional” master. Clean digital processing is valid when it serves the music.
8. Check width, phase and mono compatibility
Compare stereo and mono playback. Listen for disappearing vocals, weak bass, hollow drums or unstable high frequencies. Use a correlation or phase meter as supporting evidence, not as the only judge.
Keep low frequencies stable and be conservative with widening. A slightly narrower master that survives phones, clubs and mono playback is stronger than a huge stereo image that collapses.
9. Set final loudness with a true-peak limiter
Place a true-peak limiter near the end of the chain and raise level until the song reaches the intended balance of loudness, punch and clarity. Watch integrated LUFS and true peak, but listen for flattened drums, pumping, crackle, vocal harshness and bass distortion.
There is no universal LUFS target for every release. Loudness is a creative and genre-dependent decision, while streaming services may change playback level through normalisation. If the limiter must work brutally hard to reach a number, fix the mix or accept a more dynamic master.
10. Export, encode-check and quality-control the file
Export a stereo WAV at the native sample rate and highest useful bit depth required by your distributor. Spotify currently accepts a native 24-bit master at 44.1 kHz or higher and advises delivering the file exactly as mastered. Do not upsample a lower-resolution mix simply to create a larger number.
Apply dither only when reducing bit depth, and place it at the final output stage. Do not repeatedly dither the same master. Audition the exported file, not only the live DAW session.
Listen on studio monitors or headphones, ordinary earbuds, a phone, a car and a small speaker if available. Check the beginning, transitions, loudest section, quietest section and final tail. Confirm the filename, sample rate, bit depth, channel format and metadata before delivery.
A simple mastering chain for beginners
This order is a useful starting point, not a rule. Skip any stage that does not solve an audible problem.
The safest default is fewer processors and smaller moves. If bypassing the complete chain makes the music feel more open, more emotional or more powerful after you match the volume, the chain is not helping yet.
How loud should a mastered song be for streaming?
Do not master every song to one universal number. Integrated LUFS describes average programme loudness; it does not measure punch, tonal balance or whether the master sounds good. The right result depends on the music, the arrangement and how much transient movement you want to preserve.
Spotify’s published guidance
Spotify currently states a target of −14 dB integrated LUFS and a maximum of −1 dB true peak. For masters louder than −14 LUFS, Spotify recommends keeping true peak below −2 dB to reduce the risk of extra distortion during lossy encoding.
That is platform guidance, not a command to force every genre to −14 LUFS. Spotify applies loudness normalisation during playback. A louder master can be turned down, but dynamics removed by heavy limiting are not restored.
For Apple playback, auditioning the encoded result is more useful than chasing a supposed universal Apple LUFS target. Apple provides Digital Masters tools that let engineers preview the effect of AAC encoding. Codec previewing matters because a clean WAV can reveal inter-sample overs or high-frequency distortion after encoding.
For a deeper explanation of integrated loudness, short-term loudness and true peak, read the complete LUFS guide.
Mastering priorities by genre
Genre references help you understand the expected relationship between low end, vocal, percussion and density. They should guide decisions—not lock you into a preset.
| Style | Protect | Watch for |
|---|---|---|
| Rap / Hip-hop | Vocal authority, kick impact and intelligibility | Harsh vocal presence, dull snares and low-mid congestion |
| Trap / Drill | 808 note definition, slides and sharp drum transients | Limiter distortion, uncontrolled sub notes and piercing hats |
| R&B / Soul | Vocal detail, depth and smooth dynamics | Flattened emotion, sibilance and excessive brightness |
| Afrobeats / Amapiano | Percussion movement, bass rhythm and open mids | Crowded transients, muddy log drums and narrowed groove |
| Electronic / Club | Low-end stability, energy and section-to-section impact | Fatigue, brittle drops and bass that disappears in mono |
If your focus is rap, trap or R&B, continue with the rap mastering settings guide for a more genre-specific workflow.
Should you master manually, use AI or hire an engineer?
| Route | Best when | Main strength | Main limitation |
|---|---|---|---|
| Manual DIY | You want control and are willing to learn | Every decision can follow your intent | Monitoring and experience affect judgement |
| AI mastering | You need fast, repeatable results or a comparison master | Speed and accessibility | It receives the stereo mix, so mix problems remain limiting |
| Mastering engineer | The release is important and independent judgement matters | Experienced ears, room, monitoring and feedback | Requires budget, communication and turnaround time |
These routes are not enemies. You can use an AI master to test a mix, build your own version for learning, and send the strongest mix to a mastering engineer for a major release. The decision should follow the value of the song and the consequences of getting it wrong.
Common mastering mistakes
- Chasing loudness before balance: the limiter exposes problems; it does not solve them.
- Comparing at different volumes: level-match bypasses and references before deciding which sounds better.
- Using every processor: an empty slot is better than unnecessary EQ, compression or widening.
- Trying to fix one instrument in the stereo file: request a mix revision when the vocal, kick or bass is the real problem.
- Mastering only on headphones: use several playback systems and check mono compatibility.
- Ignoring codec behaviour: lossy encoding can expose peaks and distortion not obvious in the original WAV.
- Exporting without listening: always audition the actual delivered file from beginning to end.
Before release
Final mastering checklist
- The mix is final, lossless and not clipping.
- References were compared at matched loudness.
- The full song was checked, not only the loudest section.
- EQ and dynamics processing improve a problem you can identify.
- Kick, bass, vocal and percussion remain clear after limiting.
- Stereo width survives a mono check.
- Integrated loudness and true peak were measured on the final output.
- The encoded preview does not introduce audible distortion.
- Dither was used only if bit depth was reduced.
- The exported WAV was replayed and verified.
Frequently asked questions
Can I master my own song?
Yes. You need a clean mix, reliable monitoring, references, accurate metering and the discipline to make small decisions. An independent engineer may still add value by hearing problems you have become used to.
What plugins do I need to master a song?
At minimum, use dependable metering, EQ and a true-peak limiter. Compression, dynamic EQ, saturation, clipping and stereo tools are optional. Stock DAW processors can be enough when you understand what each move is doing.
Should I master to −14 LUFS?
Not automatically. Spotify publishes −14 dB integrated LUFS as its playback reference and mastering guidance, but the right production master still depends on the song and genre. Prioritise sound, dynamics and true-peak control instead of forcing every record to one number.
How much headroom should I leave before mastering?
There is no mandatory fixed number. Export a clean stereo mix that does not clip and is not crushed by a loudness limiter. A mastering engineer can reduce level with gain; clipped audio cannot be restored reliably.
Should I export WAV or MP3 for mastering?
Export WAV or AIFF at the project’s native sample rate and bit depth. Use MP3 only as a listening preview when a lossless mix is unavailable.
Does mastering make a bad mix sound professional?
Mastering can improve the presentation of a mix, but it cannot independently rebalance every instrument inside a stereo file. Fix arrangement, recording and mixing problems at their source whenever possible.
From finished mix to release-ready master
Master your song online
Upload your stereo mix to BTR AI Mastering, hear the finished result and compare it with your own chain before you release.
Try AI Mastering