How to master a trap beat: LUFS, limiting and loudness for streaming

Master a trap beat for streaming by limiting the mix to around -8 to -10 LUFS integrated, capping true peaks at -1 dB, and adding a short fade at both ends of the file. Spotify turns every loud master down to about -14 LUFS anyway, so the number itself wins nothing: the real goal is perceived loudness, how loud your beat sounds after normalization has pulled it and every other track to the same measured level. That comes from the limiter, the mix underneath it, and knowing when to stop pushing.

Mastering is also more than loudness. It’s preparing the track for its destination, which for most beats means a streaming platform, a playlist, and the song that plays right after yours.

Why does my beat sound quiet on Spotify?

A beat sounds quiet on Spotify for one of two reasons: the master sits well below the platform’s -14 LUFS reference with peaks that stop it from being turned up, or the master measures -14 LUFS but carries less perceived loudness than the tracks around it. Both come from the same normalization math, so it’s worth understanding once.

Streaming platforms measure each track’s integrated loudness in LUFS and apply gain to even things out. Louder masters get turned down: upload at -7 LUFS and Spotify plays it 7 dB quieter. Quieter masters get turned up, but with a catch. The platform only adds gain until the track’s true peak reaches -1 dB. A dynamic master at -20 LUFS whose peaks already touch -5 dB only gets lifted to -16, and now it sits 2 dB under everything else on the playlist. Quiet, dynamic masters lose twice: the sparse sections sit low, and the peaks block the platform from compensating.

The second reason is subtler. Two tracks can measure an identical -14 LUFS and still sound different in level, because LUFS averages the whole track. A dense master sounds louder bar by bar than a dynamic one at the same reading. Making a beat sound loud on a normalized platform means winning that comparison, and you win it with limiting, not gain.

What LUFS should a trap beat be mastered to?

Master a trap beat to around -8 to -10 LUFS integrated. That range is competitive with commercial trap releases, and the platforms handle the rest:

Platform	Reference level	What happens to a -9 LUFS master
Spotify	-14 LUFS	Turned down about 5 dB
Apple Music	-16 LUFS	Turned down about 7 dB
YouTube	-14 LUFS	Turned down about 5 dB

Getting turned down is normal. Run a commercial trap hit through a loudness meter and you’ll find it gets pulled down 4 to 6 dB on every platform too. The reduction costs you nothing, because everything on the playlist receives the same treatment.

So why master louder than -14 at all? Two reasons. A minority of listeners turn normalization off, and there your -14 master plays several dB under every commercial release it sits next to. And within the normalized world, the denser master still wins the perceived-loudness comparison described above. Don’t chase past -8, though: at -8 and beyond the limiting starts eating the drums, and trap lives on its drums.

Make one universal master. Separate masters per platform sound tidy in theory, but they multiply your exports, and the platforms’ own gain adjustment does the per-platform work for you.

What is LUFS, and how is it different from peak level?

LUFS (loudness units relative to full scale) measures perceived loudness over time, while peak level only reports the single loudest sample. A one-shot 808 can peak at -1 dB in a track that measures -20 LUFS. Streaming platforms normalize on the LUFS number, so it’s the one that decides how your beat plays.

Three readings matter on a loudness meter:

• Integrated LUFS is the average across the entire track, and it’s the number platforms use. Measure it by playing the beat from the first bar to the last. Reading it off the loudest drop tells you nothing.
• Short-term LUFS shows the loudness of the moment. Useful for checking a hook against a verse, ignored by normalization.
• True peak (dBTP) estimates the peaks that appear between samples when the file is converted to analog or to a lossy codec. Lossy encoding overshoots: a WAV that peaks at exactly 0 dB can clip and distort once it becomes AAC or Ogg. Set the limiter’s output ceiling to -1 dBTP to be safe. I run -0.5 on my own beats and haven’t been burned, but -1 is the value the platforms themselves suggest, and the half dB of loudness you give up is inaudible after normalization.

How does a limiter make a beat louder?

A limiter sets a hard ceiling, instantly compresses anything that crosses it, and lets you raise the overall gain into the space that creates. The loud moments come down, the quiet moments come up with the added gain, and the gap between them (the dynamic range) shrinks. Less dynamic range means more average level for the same peak, which reads as a higher LUFS number and, more to the point, sounds louder.

The cost shows up at two scales. On the macro level, limiting shrinks the difference between sections, so the hook slamming in after a stripped verse hits with less contrast. On the micro level, it shrinks the difference between sounds inside the bar: the kick’s transient gets shaved toward the 808’s sustain, hats step forward, and the drums trade their crack for density.

Trap tolerates heavy limiting better than most genres. The 808 sustains through the bar, the arrangement already has modest section-to-section dynamics, and the aesthetic wants density. This is why an aggressive trap master works where the same treatment would flatten a piano ballad into wallpaper. But the kick transient is where trap keeps its punch, so it’s the first casualty to protect. My working method: push gain into the limiter until the kick stops cracking and the hats start smearing, then back off 1 dB. That point usually lands me at -9 LUFS, give or take.

How much limiting is too much?

Too much limiting is the point where the beat measures louder but hits softer. Here’s an experiment worth running on your own beat. Export two masters: one lightly limited and dynamic, one crushed hard. Turn them both down until they measure the same -14 LUFS, which is exactly what Spotify will do, and A/B them blind. The crushed master sounds louder through the sparse sections, because its quiet parts got lifted. The dynamic master hits harder when the drop lands, because it kept the contrast. Neither wins outright, and that’s the lesson: after normalization, extra limiting stops buying loudness and starts spending punch.

The balance for a trap beat sits closer to the crushed end than most genres, but it’s still a balance. Keep enough macro dynamics that the hook arrives, and enough transient on the kick that the beat knocks on a phone speaker. A limiter can’t add either one back once the master is flat.

How do you check whether a master is ready?

Check a master by comparison, because a beat alone in a studio tells you nothing about how it performs on a playlist. Before uploading, I run four checks:

1. Reference tracks, level-matched. Load two or three commercial beats in your lane next to your master and match their playback level, either with a referencing tool or by gain-staging them by hand. Unmatched comparisons are worthless: the louder track always sounds better. Compare the low end, the mids, the highs and the density, and note where yours differs.
2. Tonal balance against the genre. Metering tools can show your frequency balance against a curve averaged from thousands of commercial tracks in a genre. If your low-mids sit far under the trap curve, that finding belongs to the mix. Go back and fix it at the source; the common EQ problem frequencies covers where to look.
3. The loudness penalty. Free web tools report how many dB each platform will turn your master down. Read the number as information: if Spotify pulls you down 5 dB, you know your -1 dBTP ceiling and your density survive that reduction, and you know you sit in commercial territory.
4. A codec preview. Streaming platforms transcode your WAV to lossy formats, and the result differs from your export: the top end dulls, stereo width shifts, transients shave. Tools exist that audition the encoded result inside your session. Listen to the master through one before sending it anywhere, because the streamed version is the only version listeners hear.

What does a master need besides loudness?

A master needs clean edges and, if it ships with other tracks, a family resemblance. Loudness gets the attention, but these are the details that separate a prepared master from a loud mix:

• Fade the top and tail. Put a very short fade-in at the start of the file, even when it starts from silence, and a deliberate fade-out at the end. Transcoding can put a click or a breath of noise where you thought silence was, and a track that just stops reads as unfinished. The fade gives you control over how the beat ends instead of leaving it to the encoder.
• Master the body of work in one session. Mastering tracks one by one, in separate sessions, loses the objectivity that makes mastering work: by the time you open the next session you’ve forgotten what the last track sounded like. For a beat tape or an EP, drag every final mix into a single session, stagger them on separate tracks, and A/B across them. Match the loudness first, then the tonal balance, so five beats read as one project made by one person. Listen to each transition into the next track while you’re there; the flow between tracks is part of the release.
• Leave the mix alone. If a master needs more than broad EQ moves and limiting, the problem lives in the mix. The kick-and-808 relationship in particular decides how far a limiter can push before the low end folds; how to mix kick and 808 together covers that, and how to mix a trap beat covers the rest of the mix this post assumes is done.

Where does a loud master actually start?

A loud master starts three steps earlier, in the composition and the mix. A limiter raises the average level of whatever you feed it; feed it a mix where the 808 and kick fight, where the melody stacks mud into the low-mids, and the limiter amplifies the fight. The beats that master loudest are the ones with disciplined low end and space in the arrangement, because density you choose compresses gracefully and density you didn’t notice turns to distortion.

That’s also why I start beats in Songen. It generates a trap loop (dark trap and the neighboring trap styles are all in its 50+ genres) as four editable MIDI tracks: lead, chords, bass, drums. Because every part arrives as MIDI, the low end stays inspectable; you see exactly what the bass plays against the kick and edit notes instead of surgically EQing a rendered file. Songen doesn’t master, and nothing in this post happens inside it. The workflow is: generate the foundation, shape the beat, mix it, and on the Mac app export the full mix and stems instantly. That export is what your limiter, your loudness meter and your reference checks are waiting for.