When Is Your Master Done: 5 Signs to Stop Mastering

That moment hits every engineer differently. Some feel it as relief. Most feel a quiet, nagging suspicion that one more pass would fix something. It won't. And learning to tell the difference between productive revision and compulsive tweaking is the actual skill nobody teaches you in mastering courses.

We've sat at the bus enough times to know this: the fear of being done is real, it's irrational, and it will cost you hours if you let it run unchecked. This guide is a framework for calling it. Not a philosophical meditation. A decision process with checkpoints.

The technical side is simpler than people make it. The psychological side is harder than anyone admits. We'll cover both.

What Does "Done" Actually Mean in Mastering?

Mastering has two jobs. Make it sound right. Make it meet specifications. Both need to be satisfied before you bounce the final file.

"Sound right" means your master holds up against references, translates across playback systems, and serves the mix without fighting it. "Meet specifications" means your LUFS targets, true peak limits, sample rate, bit depth, and file format match what the distribution platform demands.

Streaming platforms have clear, public requirements. Spotify normalises to -14 LUFS integrated. Apple Music targets -16 LUFS. YouTube sits around -14 LUFS. Tidal varies by tier. If your master is hitting -9 LUFS integrated and you're delivering to Spotify, the platform will turn it down anyway. Loud isn't done. Correct is done.

True peak should land at -1 dBTP or lower for most streaming platforms. If your limiter is hitting higher than that, you're not finished. This isn't aesthetic. It's a technical pass/fail.

The Distinction Between Aesthetic and Technical Completion

Technical completion is binary. Either the file meets spec or it doesn't. Check this with a metering tool that reads LUFS, true peak, and dynamic range. Youlean Loudness Meter is free and accurate. HOFA IQ-Analyser and Nugen Audio MasterCheck give you more detail if you're delivering to broadcast.

Aesthetic completion is where the psychology creeps in. You can always find something to adjust. The question is whether that adjustment makes the master better for a listener, or just different for you.

We use a simple test: write down exactly what you'd change and why, in technical terms. "Bring up the 3kHz shelf by 0.5dB because the vocal presence is sitting behind the guitars during the chorus" is a legitimate revision note. "It just feels like something's off" is not. If you can't write it down specifically, leave it alone.

How Long Should Mastering Take Before You Stop?

Professional mastering engineers spend 30 to 90 minutes per track on a well-prepared mix. A poorly prepared mix can push that to several hours. Stem mastering takes longer, often 2 to 4 hours per song.

Here's the number that matters more: if you're spending more than 4 to 5 times the song's runtime actively adjusting the master, something is wrong. Either the mix needed more work before it reached you, or you're in a revision loop that isn't serving the music.

Full album mastering runs about a week for a professional working through 10 to 12 songs with cohesion passes. A 5-track EP is 3 to 4 days. A single should be done in under 48 hours from a clean mix.

These aren't arbitrary timelines. They're calibration tools. If you're on hour six of mastering a 3.5-minute track, stop. Step away. Come back tomorrow with fresh ears.

The Diminishing Returns Window

We call it the diminishing returns window. It opens about 45 minutes into a mastering session and closes when you start making things worse.

Here's what that looks like in practice: we were mastering a drum-heavy electronic track last year. The low end felt tight at the 40-minute mark. By hour two, we'd pulled so much 60Hz out trying to "tighten" the kick that the whole bottom end went thin. We had to reload the session from 45 minutes in. Two hours of work, net effect: negative.

Ear fatigue is real and it compounds. Your brain stops hearing what's there and starts hearing what you're afraid is there. Take a break before you hit that window, not after.

How Do You Build an Objective Checklist?

Before you start any mastering session, write down what you're trying to achieve. Not vaguely. Specifically. This list becomes your done condition.

A usable goals list looks like: "Bring the integrated LUFS to -14. Tame the 200Hz buildup that muddies the low mids. Increase the perceived loudness of the chorus without squashing the transients. Match the overall brightness to the reference track." Four goals. When all four are addressed, you're done.

Without that list, you'll drift. Every pass reveals something new to fix. That's not progress. That's anxiety in disguise.

The Three-Reference Method

Pick three reference tracks, not one. One that represents the brightness you're aiming for. One that represents the low-end weight. One that represents the dynamic range and loudness level.

They don't all have to be in the same genre. They just have to isolate the characteristic you need to verify. Load them into your DAW and level-match before comparing. This is non-negotiable. Our brains consistently perceive louder as better. Compensate for it or your A/B test is meaningless.

In Pro Tools, you can use a gain plugin before the master fader to match levels manually. In Reaper, volume trim on the track works fine. Nugen Audio's SigMod handles this cleanly if you want a dedicated solution. Get them within 0.5 dB of each other and then compare.

If your master holds up across all three reference dimensions, it's done. If it fails on one, you know exactly where to work.

What Are the Emotional Indicators That You're Finished?

This is the part nobody talks about seriously. There are emotional signals that tell you the master is done, and they're as reliable as LUFS meters once you know how to read them.

The first is listening without trying to fix anything. You play the track through and you're just listening. Not monitoring. Not scanning for problems. When that happens naturally, your brain has accepted the master as a finished object.

The second is boredom with revision. Not boredom with the song. Boredom with the mastering process itself. That's your nervous system telling you there's nothing left to productively do.

The third is the frustrating one: when every revision you try sounds like a lateral move. You boost 8kHz by 1dB and the air comes up. You undo it and the air goes back. Neither version is better. That's done. Pick one and move on.

The Psychology of Perfectionism in Mastering

Perfectionism in mastering isn't about quality. It's about fear. Fear of releasing something imperfect. Fear of criticism. Sometimes, fear of being finished because finished means it can be judged.

We've seen this wreck careers, slowly. Engineers who never release because the master is "almost there." Producers who pay for three mastering revisions of the same song over two months and still sit on it.

The honest truth: no commercially released master is perfect. Play any chart-topping track through your reference monitors and you'll find something to critique in 10 minutes. The Rolling Stones' catalogue has mixes that would fail a modern A/B test on low end. Nobody cares. The music matters. Get close, get it right technically, and release it.

When Should You Push Back on Client Revision Requests?

If you're mastering for clients, this gets more complicated. Clients request changes for two reasons: the master genuinely needs adjustment, or they're feeling the same anxiety you'd feel if it were your own music.

Standard professional practice is two rounds of revisions included, a third at additional cost. That's not arbitrary. It's a psychological structure that makes clients commit to their feedback. Open-ended revisions are a loving gift to indecision.

When a client asks for a revision, ask them to be specific. "It needs to be louder" is actionable. Push the integrated LUFS up by 1 to 1.5dB and check true peak. "It doesn't feel right" is not actionable. Ask what doesn't feel right and in which section. Make them locate the problem in time and frequency before you touch anything.

When the Engineer Should Say No

Push back when a client request would damage the technical integrity of the master. If they want -7 LUFS on a Spotify release and you're already at -12 with limiting artefacts starting to show, explain what they're asking for in plain terms: "If I push this louder, the transients will smear and Spotify will turn it down to -14 anyway. You'll have a worse-sounding master at the same playback volume." Then let them decide with full information.

A good client will trust that. A difficult client might not. But your job is to tell them the truth once, clearly, and then honour their decision if they still want to proceed. You've done your job when you've given them accurate information.

Does AI Mastering Change When You're Done?

AI mastering tools like LANDR, eMastered, and Matchering 3.0 produce a finished file quickly. Sometimes in under a minute. That speed makes the "when am I done" question feel answered before it's really been asked.

AI mastering is done when the output file meets platform spec and serves the song. The same criteria apply. Don't assume it's done because the algorithm decided it was. Run the output through Youlean. Check the true peak. A/B it against a reference. AI tools are fast, but they're not infallible and they can't hear your creative intent.

Matchering 3.0 is worth knowing: it's open-source, lets you supply your own reference track, and the mastered output often gives you a useful starting point for hearing where your mix sits tonally relative to the reference. It's not a replacement for ears, but it's a satisfying diagnostic tool.

Summary

Your master is done when it passes both a technical checklist and a creative one. Technical: LUFS target met, true peak at -1 dBTP or lower, correct sample rate and bit depth for the platform. Creative: it holds up against level-matched references, every remaining change is lateral rather than additive, and you can listen through without reaching for a knob. Write your goals before you start. Work from a three-reference system. Respect your ears' fatigue. Stop when the list is checked, not when the anxiety quiets. Those are different things.