How do you master a song?

Start from a mix with real headroom (peaks around -6 dB, no limiter on the mix bus). Then, gently: corrective EQ for tonal balance, light compression for glue (a dB or two), and a limiter last, raising loudness until it hurts the punch and then backing off. Reference against level-matched commercial tracks in your genre, keep true peaks around -1 dBTP, take a day off, and listen again on multiple systems before exporting with dither, applied once.

Is it OK to master your own music?

Coming from someone who masters for a living: yes, sometimes. DIY mastering is genuinely fine for demos, SoundCloud material, testing songs, and learning the craft. It costs you on releases with stakes, ad spend, playlist pitching, sync submissions, or an album that needs track-to-track cohesion, where fresh professional ears and a calibrated room earn their fee. The line is the release's stakes, not your ability.

What LUFS should I master to?

Not -14. That number is a playback normalization level, not a mastering target, and commercial releases ignore it: in our measurement of 314,876 tracks, the overall median was around -9.5 LUFS and 90% of releases were louder than -14. Real-world ranges run roughly -6 to -9 for pop, hip-hop and EDM, -8 to -10 for rock, and -12 to -16 for acoustic and jazz. Master to your genre's actual range with dynamics intact, not to a platform myth.

What can't mastering fix?

Anything that lives in the mix: a buried vocal, clashing arrangement, bad vocal tuning, clipped recordings, mono-incompatible stereo bass, sibilance baked across the mix. Mastering polishes a finished mix; it cannot reach inside it. If a mastering engineer tells you to fix the mix first, that's the good ones, the alternative is paying to make the problems louder.

How to Master a Song (the Honest Version)

Every guide ranking for this question is selling something that shapes its advice: the plugin company concludes you can master (with its plugins), the AI service concludes DIY is hard (use the AI), the $49 mastering shop concludes you need fresh ears (theirs). I master records for a living, which gives me the one position nobody else on this page has: I'm the person you'd otherwise pay, telling you when not to pay me. Here's the whole thing, the chain, the honest decision line, and the numbers the myth-repeaters get wrong.

The Chain, In Plain Steps

Start with a proper pre-master: your mix bounced with peaks around -6 dB and no limiter on the mix bus. Then, in order, all of it gentle: (1) corrective EQ, small moves for tonal balance against a reference; (2) glue compression, a dB or two of gain reduction, slow enough to breathe; (3) limiter last, raise loudness until the punch starts to die, then back off a notch; keep the true-peak ceiling near -1 dBTP so streaming codecs don't clip it. Reference against level-matched commercial tracks in your lane the whole way (unmatched comparisons always flatter the louder file). Sleep on it, listen on your monitors, earbuds, phone, and car, then export at delivery spec with dither applied exactly once. That's the craft's skeleton, every ranking guide agrees on it, and it's genuinely learnable.

The -14 LUFS Myth, Killed With Data

The most repeated advice in DIY mastering is "master to -14 LUFS because Spotify normalizes there." It confuses a playback level with a mastering target, and the actual releases you compete with ignore it completely. We measured 314,876 commercial tracks: the overall median sits around -9.5 LUFS, and 90.4% of releases are louder than -14. Real ranges by lane: pop/hip-hop/EDM around -6 to -9, rock -8 to -10, acoustic and jazz -12 to -16, the full ten-year, genre-by-genre study is here. Master to -14 and your track plays back sounding polite next to a playlist of records that didn't. That doesn't mean crush it, normalization does punish the slammed masters, it means your target is your genre's real range with the dynamics alive, not a platform's playback constant.

What I Actually Hear in Submitted DIY Masters

The guides list process mistakes; here are the audible ones that arrive at my studio weekly: the limiter slammed until the transients are flat and the chorus pumps; low end that either swallows the mix on monitors or vanishes on phones; a harsh 2–5kHz glare from additive EQ chasing "clarity"; stereo-widened bass that folds to mush in mono; and, most common of all, mastering deployed to fix a mix problem, which just makes the problem louder and shinier. If you learn to hear those five in your own work, you're ahead of most of the DIY field.

What Mastering Cannot Do

A buried vocal. A clashing arrangement. Vocal tuning. Clipped recordings. Sibilance baked through the mix. None of that is reachable from the master bus, mastering polishes the mix it's given, and an engineer who says "fix the mix first" is doing you the biggest favor in this business. Check before you polish: the analyzer measures your track against those 314,876 records, loudness, balance, the works, free, and tells you whether you're mastering a finished mix or laminating a problem.

The Shorter Version

If you asked me in the room instead of on a search page, the answer is shorter than that chain. Get it up to level. Leave some true peak, minus one dB. Turn it up until it's hitting your limiter, and once it sounds right, that's probably a wrap.

I don't know. I'm just playing, most of the time. Different songs need different tools, and I have to sit with a record and play with it before I know which ones this one wants. The chain above is the skeleton and it's the same for everybody. What hangs on it is not, and that part I can't write down for you.

Which is the real decision line, and it's simpler than the one below: if you can't get it to sound the way you need it to sound, pay someone else.

The Honest Decision Line

DIY is genuinely fine for: demos, SoundCloud material, testing songs with your audience, and learning, do it, you'll get better at hearing your own mixes. Pay a professional when the release has stakes: ad spend behind it, playlist pitching, sync submissions, or an album that needs cohesion across tracks, fresh calibrated ears are what you're buying, and here's what that honestly costs. I make part of my living on the paid side of that line, so weigh my bias, but notice the line cuts both ways, and I just told you where most of your projects fall on the free side of it.