How to Master Your Own Songs

A Four-Plugin Mastering Chain You Can Build for Your Own Songs

A mastering chain for indie artists works best when it is short. Four processors in a fixed order cover most of what a self-mastered track needs: a corrective EQ, a gentle compressor, a layer of saturation or tape color, and a limiter at the end. Each one has a job, a point where it starts doing damage, and a setting where you should leave it off.

Every processor you add changes level, tone or dynamics, and it becomes harder to tell which change helped. A chain of eight plugins invites you to adjust each a little, and the small adjustments pile up into a master that sounds louder and flatter than the mix you started with. Four slots keep each decision easy to hear and easy to undo.

The chain works with whatever plugins you already own. A stock EQ, a stock compressor, a tape plugin from a bundle and a free limiter will do the job if you use them well.

Why a Mastering Chain Should Be Short (and What It Can’t Fix)

Mastering makes small, global adjustments to a finished stereo file. It can fix a slightly dark top end, a bit of low-mid build-up, or a track that sits quieter than the songs around it. It cannot move the vocal forward, tame one boomy bass note, or repair a snare that is too loud.

If the problem is in one instrument, fix it in the mix. A master-bus EQ cut that helps the guitar will also take something from the vocal and the cymbals. Treat the master as the last 5 percent of the work and expect to go back to the mix when a problem is specific.

Keeping the chain short also keeps you honest about decisions. With four slots, you can ask a clear question of each one: does this make the song better once the level is matched? If the answer is no, the plugin stays bypassed, and a bypassed slot is a perfectly good result.

Before You Insert Anything: Prep Your Mix and Set Up a Fair Reference

Preparation affects the result more than any plugin setting does. Do these things first.

Remove the mix-bus limiter. If you were mixing into a limiter to hear how the track would sound loud, take it off before you export. A limiter already printed into the mix leaves the mastering limiter with less room to work.

Leave headroom. Export the mix with peaks well below 0 dBFS. A few dB of space is plenty, and the exact number matters less than the absence of clipping. Use 24-bit or 32-bit float at your session’s sample rate so you are not losing resolution before mastering begins.

Start a new session for mastering. Put the exported stereo file on one track and build the chain on that track or on the master bus. Working in a separate project keeps you from reaching back into the mix every time you hear something, and it lets you look at the song as a finished file.

Pick a reference. Choose one or two commercial tracks in a similar genre, with similar instrumentation and production, that you think sound good. Import them into the same session. Reference tracks tell you where your low end, brightness and loudness sit relative to music that people already play on the same speakers.

The reference only helps if the comparison is fair, and loudness is the main thing that makes it unfair. Louder audio usually sounds better to a listener even when it is not. Before you compare anything, turn the reference down until it matches your track’s perceived level. Use a loudness meter if you have one and your ears if you don’t, and keep the volume at a moderate listening level so your ears do not tire.

Route the reference so it bypasses your mastering chain. If the reference runs through your EQ and limiter, you are not hearing it as it really sounds.

Slot 1: Corrective EQ for Small, Broad Moves

The EQ goes first because everything after it reacts to the tonal balance it hands over. If a bass build-up reaches the compressor, the compressor responds to the boom and clamps down on more than it should.

Use this slot for small, wide moves. A boost or cut of 1 to 2 dB with a broad bandwidth changes the character of a whole track without making any one region stand out. A narrow, deep cut is a mix-stage move and a sign you should open the mix instead.

Here are the areas worth checking, with ranges that are rough guides and vary from song to song:

• Subsonic rumble. A gentle high-pass filter well below the fundamental of your lowest instrument removes energy you cannot hear but a limiter will still react to. Watch a spectrum analyzer and listen to the kick and bass after you engage it.
• Low mids. A boxy or cloudy feel often sits somewhere in the low hundreds of hertz. A small, wide dip there can clear a mix, but check that the guitars, keys and vocal body do not thin out.
• Upper mids. A fatiguing, harsh quality tends to live in the region where the ear is most sensitive, roughly 2 to 5 kHz. A fraction of a dB can matter here.
• Top end. A gentle high shelf can add air to a dull mix. If the mix is already bright, a small shelf cut can calm brittle cymbals and sibilance.

Compare against your reference after each move, with levels matched. If you boost the highs and the track gets louder, the boost will sound like an improvement even when it is not one. Lower the output by the same amount before judging.

If you need more than about 3 dB at any frequency, treat that as information about the mix. Go back, fix it at the source, and re-export. If you want more practice with how EQ decisions sound, the FabFilter beginner’s guide covers one popular EQ and its companions.

Leave this slot bypassed when your track already matches the reference tonally. Switch it in and out at matched levels. When you cannot hear a clear improvement, the right setting is off.

Slot 2: Gentle Bus Compression to Glue the Track

Bus compression on a master is much lighter than compression on a drum or a vocal. The goal is to nudge the loudest moments down slightly so the instruments move together. Many people hear this as cohesion, though the technical effect is a small reduction in dynamic range.

Start with these settings and adjust by ear:

• Ratio: low, around 1.5:1 to 2:1.
• Attack: slow enough to let drum transients through. If the attack is too fast, the punch of the kick and snare flattens.
• Release: short enough that the gain reduction recovers before the next beat, and long enough to avoid audible pumping. Many compressors have an auto release that gives a decent starting point.
• Threshold: set so the gain reduction meter shows about 1 to 2 dB at the loudest moments. Less is acceptable and more is rarely necessary.
• Makeup gain: adjust until the compressed and bypassed signals are the same loudness. This step is easy to skip, and skipping it makes the compressor seem to improve the track simply because it is louder.

If your compressor has a sidechain high-pass filter, set it so the low end does not drive the compression. Otherwise a big kick or bass note pulls the entire mix down with it.

Listen for the gain reduction meter moving in time with the song’s pulse without any audible pumping. The mix should feel a little more settled. If the vocal ducks every time the kick hits, or the cymbals swell after each snare, back the ratio off or lengthen the attack.

Bypass this slot for music with deliberate dynamics, such as a sparse acoustic song, a track that builds from a whisper to a full arrangement, or a recording where the quiet and loud sections are the point. A track that already feels cohesive also needs no compression. If the compressor is audibly changing the sound and you cannot say what is better, leave it off.

Slot 3: Saturation or Tape Color, Used Lightly

Saturation adds harmonic distortion. Tape emulations also soften transients and shift the tone in the low and high frequencies. At light settings they add thickness and a slightly smoother top end. At heavy settings they blur detail and make the track sound dull or fuzzy.

Treat this slot as a seasoning. Turn the drive up until you hear a change, then pull it back until the change is only just audible. If the plugin has a mix or blend control, use it, because a partly blended signal often keeps the clarity of the original and adds some of the color.

A few specifics to watch:

• Level shifts. Saturation often raises the output level. Match the level before judging, or the plugin will win the comparison by being louder.
• Extras you may not want. Many tape plugins add noise, wow and flutter, or stereo changes. Turn off anything you did not choose on purpose.
• Clarity. Listen to the vocal consonants and the hi-hats. If they lose definition, you have gone too far.

The order matters here too. Saturation sits before the limiter because it can raise peaks and change the harmonic content the limiter will respond to. Putting it after the limiter risks pushing the signal past your ceiling.

Plugins vary a lot in how they handle this job. A review of the Rupert Neve Designs 542 Tape Emulator shows one approach if you want to see how a dedicated tape unit is judged.

Leave this slot bypassed when the mix already has plenty of distortion and harmonic content, such as tracks with saturated drums, distorted guitars or heavy 808s. Extra saturation on top of that tends to blur the result. Bypass is also correct if you cannot hear a benefit after matching levels.

Slot 4: The Limiter, and How to Set Loudness Without Crushing the Song

The limiter goes last because its job is to control the highest peaks, and anything placed after it could push those peaks back up. It is also where most self-mastered tracks go wrong, because it is the plugin that makes the song louder and the easiest one to overuse.

Set it up in this order:

1. Set the ceiling. A true-peak ceiling a little below 0 dB, such as -1 dBTP, leaves room for the extra peaks that can appear when the file is converted to a lossy format for streaming. Turn on the limiter’s true-peak or oversampling mode if it has one.
2. Raise the input gain. Push the level up gradually while the song plays, and watch the gain reduction meter.
3. Listen as you go. Stop when the track gets as loud as the reference at matched perceived level, or earlier if the song starts to suffer.
4. Adjust the release. A fast release can add distortion on bass-heavy material. A slow release can leave the limiter stuck down after a loud hit and make the track feel squashed. Try a few settings on the loudest section.

Here are signs you have pushed too far:

• The kick loses its punch and sounds shorter.
• The vocal gets smaller and more fuzzy.
• Cymbals and hi-hats splash into a constant hiss.
• The chorus feels no bigger than the verse.
• You hear crackle or distortion on loud notes.

Loudness matters less than it used to. Streaming services use loudness normalization, which turns down tracks that are louder than their target. A track pushed far beyond that target does not play louder than a more moderate one. It plays at about the same level with less dynamic range. Check the current targets of the platforms you release on rather than relying on a number from memory, since services change them.

Genre affects this. A hip-hop or dance track usually sits louder than a folk song, so match your reference rather than a fixed number. If your reference is a dense modern track and yours is a sparse acoustic recording, do not chase its level.

Use one limiter. Stacking two of them rarely helps and often adds distortion. If a single limiter seems to need more than a few dB of reduction to match your reference, return to the earlier slots or the mix and look for peaks that are using up the headroom, such as a loud snare or a sharp vocal consonant.

Check the Master: Level-Matched A/B, Mono, and Other Speakers

A master that sounds good in the room where you made it may not hold up elsewhere. Run a few checks before you call it done.

Level-matched A/B. Compare the mastered file with the unmastered mix, with the output of your chain turned down so both play at the same loudness. The master should sound better in some way you can name: a clearer top end, a firmer low end, a more even level. If you can only say that it is louder, the chain has not helped, and you should bypass slots one at a time to find the one that is not earning its place.

Reference comparison. Play the master against your chosen reference at matched levels. Compare the low end, the vocal presence and the brightness. Notice the differences, but do not copy them. Your song has its own balance.

Mono. Collapse the track to mono and listen. Phase problems and wide effects can make instruments weaker or disappear. If the mono version loses its vocal or bass, the issue may lie in the mix’s stereo width. Our guide on making a mix sound wider without phase problems covers the mix-side causes.

Other speakers. Listen on earbuds, a phone speaker, a car stereo and a pair of headphones if you have them. The phone check shows whether the vocal and the main melody stay audible without bass. The car and headphones show how the low end behaves. Make notes, and adjust the chain only for problems you hear on more than one system.

A break. Leave the track for several hours or overnight, then listen again at a moderate volume. Ear fatigue hides problems, and a fresh listen often reveals that a setting needs backing off.

If you are releasing several songs together, play them back to back. They should sit at similar loudness and similar tonal balance, so the listener does not reach for the volume control between tracks.

When to Stop and Send the Track to a Mastering Engineer Instead

A short chain handles a lot of ordinary jobs, and some situations are better served by another person.

Consider sending the track out when:

• You have gone through the chain twice and still hear a problem you cannot name. A trained engineer in a well-treated room hears things your setup may hide.
• The release matters a great deal. For an album, a vinyl pressing or a track with a marketing push behind it, a second set of ears is worth the cost.
• The track needs more than small moves. If it needs large corrections or a lot of EQ, the problem lies in the mix, and an engineer will probably ask you to revise it.
• Your monitoring is the weak link. A room with strong bass buildup and no treatment makes it hard to judge low frequencies, and the master can end up with the opposite error.
• You are mastering a set of songs and cannot get them to sit together. An engineer who can compare them across the whole project may sort that out faster.

If you send a track, export the unmastered mix with headroom and no limiter. Include your reference tracks and a short note about what you like and what concerns you. Engineers can work with a clear description of the goal, and the unprocessed mix gives them the most to work with.

Mastering it yourself first is not wasted effort. You learn how your mixes translate, you find out which problems belong in the mix, and you send a better file to the engineer.