The default move for an independent artist with a finished mix is to send it to a mastering engineer. The default has been the default for so long that most artists never question it. They should.
This is the case for owning the master chain — what changes when you do, what stays the same, and the actual technical setup that makes it work in 2026 on a regular Mac.
The Outsource Reflex
Mastering is the last step before release. It's the step where small changes — broadband EQ, glue compression, a final limiter — make the difference between a track that competes on a streaming service and a track that sounds like a demo.
For most of the last twenty years, the right play for an independent artist was to send the mix to a mastering engineer. The room mattered. The monitors mattered. The engineer's ears mattered. The combination of those three things was hard to replicate at home.
That argument is weaker now than it has ever been.
The room still matters, but room correction software (Sonarworks, ARC, the Apple-native equivalents that shipped in 2025) flattens the room enough for serious work. The monitors still matter, but a pair of $1,500 nearfields plus a pair of headphones for cross-reference is a one-time spend that rounds to about ten outsourced masters. The engineer's ears still matter — and that is the part this post is about.
What I Stopped Outsourcing
I stopped outsourcing the mastering decisions. The decisions are: how much loudness, how much glue, how much top-end air, what the low-end shelf does, and whether the track sits in front of, behind, or alongside the rest of my catalog.
Those decisions should not be made by someone who has not heard the rest of my catalog and does not know what release this track sits between in the calendar. They should be made by the artist who built the catalog and is going to live with the result for the next twenty years.
A great mastering engineer makes a great track. The artist who masters their own catalog makes a great catalog. Those are different optimization problems.
The Chain I Actually Run
Names matter. Here is the actual chain on the Mac:
Reference Loop. Three reference tracks loaded in Logic, level-matched at -14 LUFS short-term. The first reference is something I love from another artist. The second is the best track in my own catalog from the last 90 days. The third is a "do-not-sound-worse-than" track from a peer. Every mastering decision is ear-checked against these three on a one-button toggle.
Pre-Master Bus. Light broadband EQ with a high-shelf air boost above 12 kHz, a tilt toward the upper-mids if the mix landed dark, and a subtle low-end shelf adjustment if the kick is fighting the bass. Nothing aggressive. The mix should already be 90% there before this stage.
Glue Compressor. Slow attack, fast-ish release, 1.5 dB to 2.5 dB of gain reduction max, ratio in the 2:1 to 3:1 zone. The point is not to compress. The point is to make every element sound like it lives in the same room.
Saturation Stage. A tape-style saturator with mix knob around 30%. This is the secret sauce — analog warmth without a single piece of analog hardware.
Final Limiter. Loudness target of -10 LUFS integrated for streaming masters, -6 LUFS for radio, -14 LUFS for an audiophile bandcamp variant. I cut three masters per release at three different loudness targets. Streaming services normalize anyway, but the transient response of the masters at different loudness levels is dramatically different — pick the one that suits the track.
True Peak Ceiling. -1.0 dBFS for streaming. Hard cap. Codec encoders distort above this and the distortion is brutal on Spotify mobile playback.
That entire chain runs on a single MacBook with $200 worth of plugins, a pair of $400 nearfields, and a $200 pair of reference headphones. Total spend: less than three professionally outsourced masters.
The DAJAI Mastering Helper
Beyond the chain, I built a small Python helper that automates the parts of mastering that should be automated.
It loads my reference tracks. It level-matches them. It runs the candidate master through a battery of analysis (LUFS integrated, LUFS short-term, LUFS momentary, true peak, RMS, crest factor, frequency-band energy distribution). It compares those numbers against the reference tracks. It flags every dimension where the master is meaningfully off.
It does not make decisions. It surfaces the data. The data is what's hard for an unaided human to compute in the time pressure of a master session.
The helper turns a 90-minute master session into a 25-minute one without changing the quality bar. It also turns consistency across the catalog from an aspiration into a measurable property.
What Going Self-Mastered Bought Me
Speed. A master happens the same day the mix is done. No waiting on an engineer's queue. No follow-up emails. No "can you push the vocal up 0.5 dB and re-cut?"
Iteration. Small changes are free. If a track gets traction and needs a louder streaming master, I cut it. If a remix collaboration wants a stem-friendly variant, I cut it. The marginal cost of a new master is twenty minutes.
Consistency. The helper enforces a measurable consistency across the catalog. Track 47 from last month sounds like it lives in the same world as track 12 from a year ago because the same chain processed both, with the same reference tracks anchoring the decisions.
Sovereignty. This is the part that matters most and that nobody talks about. Owning the master chain means I don't have a dependency. I don't have a vendor relationship I have to maintain. I don't have a quality variance from one engineer to the next. The chain is mine, the references are mine, the helper is mine, the result is mine.
What I Gave Up
A great mastering engineer hears things I do not. There are tracks where I bring an outside ear in for a final pass — usually the lead single from a project, where the stakes justify the round-trip.
The right framing: I master 95% of my catalog. The 5% that needs an outside ear gets one. That is a complete inversion of how most artists work, where they outsource 100% by reflex and never develop their own ear.
The result is that the 5% I outsource benefits from the 95% I did myself. I know the chain. I know the references. I know the catalog. The outside engineer is improving on a starting point I built — not making decisions in a vacuum.
When You Should Not Master Your Own
Be honest with yourself.
If your mixes are not consistently good, your masters will not be consistently good. Mastering does not save a bad mix. Outsource the mix, learn from the mix, then graduate to mastering.
If you are not willing to A/B test on three different playback systems before committing a master, do not master. The single biggest mastering mistake is overconfidence after a good listening session on one system.
If you genuinely do not have the time, outsource. There is no shame in it. But understand what you are buying — speed and a second opinion. You are not buying necessary expertise. The expertise is learnable.
The Actual Move For Most Independent Artists
Master your next ten releases yourself. Buy a basic chain. Use the references. Be honest about the results. After ten masters, look back at the catalog and ask: would an outside engineer have made these meaningfully better?
For most independent artists doing serious volume, the honest answer is no. The outside engineer would have made one or two of them better, made several of them about the same, and made one or two of them worse — because the engineer did not know the catalog the way you do.
That ratio is the case for owning the chain. It is also the case for owning every other part of your career that you have been outsourcing by reflex. The chain is the chain. The principle is universal.
FAQ
What plugins do you actually use?
Stock Logic Pro plugins do most of the work. The pre-master bus is Logic's Channel EQ. The glue is Logic's Compressor in vintage VCA mode. Saturation is FabFilter Saturn 2 with a tape preset. Final limiter is FabFilter Pro-L 2 with the modern preset. Total non-stock spend is about $300 lifetime.
What about Dolby Atmos masters?
Streaming services pay a higher per-stream rate for Atmos masters, but the mastering decisions are different — spatial panning, height channel use, binaural rendering for headphone playback. Atmos is its own discipline. I outsource Atmos for now and master stereo myself.
Do streaming services need different masters?
Spotify, Apple Music, and Amazon all normalize to a target loudness, so absolute loudness is less important than it was. Codec efficiency and transient response matter more. Cut your master so it sounds best after normalization, not loudest before it.
How long should a mastering session take?
If it takes more than two hours per song, the mix needs to go back to the producer. Mastering is the last 10% of a song's quality. If you are making 30% changes at the master stage, the mix is wrong.