ZEROPINK
Production aids · Axis: ZERO

Loudness targets

Enter your master's integrated loudness and true peak, and read what each destination will actually do to it — streaming, CD, and the K-System monitoring references.

Streaming

PlatformTargetYour deltaWhat it does
Spotify -14 LUFS on target Plays as mastered — you are at the target. (−2 dBTP if your master is louder than −14 LUFS)
Apple Music -16 LUFS +2.0 LU Turns you down 2.0 dB. The limiting stays; the loudness does not. (Sound Check; normalises per track by default)
YouTube -14 LUFS on target Plays as mastered — you are at the target.
Tidal -14 LUFS on target Plays as mastered — you are at the target. (album normalisation applied throughout)
Amazon Music -14 LUFS on target Plays as mastered — you are at the target. (the strictest ceiling of the set)
Deezer -15 LUFS +1.0 LU Turns you down 1.0 dB. The limiting stays; the loudness does not.
SoundCloud -14 LUFS on target Plays as mastered — you are at the target.
True peak −1.0 dBTP Within reach of most platforms, but above Amazon Music’s −2 dBTP. One master at −2 clears the whole table.

Figures verified 2026-08 · re-checked quarterly · sources listed below

CD

CD has no normalisation and no loudness target. The player plays exactly what is on the disc — which is why the loudness war happened there, and why it cannot happen on streaming. The Red Book standard specifies only one hard limit: sample peaks must not exceed 0 dBFS.

ReferenceValueWhy
Peak ceiling −0.3 dBFS The hard limit is 0 dBFS, but a small margin protects older players and DACs whose reconstruction overshoots between samples.
Loudness your call No standard exists. Loudness-war masters ran −9 to −6 LUFS; dynamic modern masters sit nearer −14. The disc will play either.

If the same master also goes to streaming, make one master to the streaming numbers rather than a separate loud one for CD — the loud version buys nothing on the disc and gets turned down everywhere else.

The K-System

Bob Katz's K-System predates LUFS and solves the same problem from the monitoring side: pin the meter's 0 to a calibrated listening level, and let the number in the name be the headroom above it. Mix to 0 on the meter and the headroom takes care of itself.

Scale0 on the meterFor
K-20 −20 dBFS RMS Film, classical, and anything with real dynamics — the most headroom, the least limiting.
K-14 −14 dBFS RMS Pop and rock — the general mastering scale, and no accident that it lands beside the −14 LUFS streaming targets above.
K-12 −12 dBFS RMS Broadcast — the least headroom of the three, for chains that expect a dense signal.

The K-System reads RMS where streaming reads LUFS — close cousins, not identical. Its idea is the ZEROPINK idea: one reference, stated out loud, that everything is measured against. The suite's 0 VU = −18 dBFS sits between K-20 and K-14 — mixing territory rather than mastering territory.

How to read this

Every platform in the table measures your track's integrated loudness and applies a fixed gain offset at playback. That single fact undoes the argument for mastering loud. A master pushed to −9 LUFS is turned down by roughly 5 dB on Spotify, and every decibel of limiting you paid for that level comes along with it — at the same playback volume as everyone else. You keep the damage and gain nothing.

The reverse is not symmetric, which is the part most pages get wrong. Turning down is trivial; turning up is not, because a boost can push peaks into the ceiling. Spotify and Apple Music will lift a quiet master, but only as far as its own headroom allows. YouTube, Amazon and Tidal do not boost at all — a sparse, dynamic record at −18 LUFS plays at −18 LUFS there, quieter than everything around it.

So the useful reading of this table is not "hit this number". It is: know which direction each platform can move you, and leave it the room to do so. The true-peak ceilings are not stylistic advice — they are the headroom a lossy encoder needs. Transcoding to AAC or Ogg Vorbis can push inter-sample peaks above where they sat in your WAV, so a master resting at 0 dBTP can distort after encoding even though the file itself never clipped.

One more honest caveat: normalisation is a setting, not a law. Listeners can switch it off, and Spotify offers Quiet, Normal and Loud playback modes. These are defaults, not destinations — which is exactly why understanding them beats chasing them.

Sources

  • Spotify for Artists — Loudness normalization on Spotify: the −14 LUFS Normal target, the Loud / Normal / Quiet playback modes, and the −1 / −2 dBTP upload guidance.
  • Apple — Apple Digital Masters and Sound Check documentation: the −16 LUFS playback reference.
  • Where a platform publishes no figure of its own, the value is the settled industry consensus and is marked as such rather than presented as official.
Typing a number in is the rehearsal. Loudness reads it off the bus as you work — integrated, true peak and crest, against the target you are delivering to. ZEROPINK | Loudness →