How loud should you master for streaming?
LUFS, normalization and the -14 myth
There is no single level. Every platform normalizes playback loudness: Spotify, YouTube and Tidal around -14 LUFS, Apple Music at -16. Aiming for -14 doesn't make you sound louder. Normalization turns down masters that are too loud, but it never gives back lost dynamics. What matters is dynamics and a controlled true peak, not a number.
What is loudness normalization?
On playback, every major platform brings all tracks to a shared target loudness. A master louder than the target is turned down; a quieter one is turned up. So two tracks reach the same perceived level whatever the file's LUFS. Pushing loudness at the mastering stage gives no loudness advantage on the platform.
Targets, platform by platform
| Platform | Target (LUFS) | True peak |
|---|---|---|
| Spotify | -14 | -1 dBTP |
| Apple Music | -16 | -1 dBTP |
| YouTube | -14 | -1 dBTP |
| Tidal | -14 | -1 dBTP |
| Amazon Music | -14 | -2 dBTP |
| Deezer | -15 | -1 dBTP |
| SoundCloud, Bandcamp | No normalization | -1 dBTP |
Reference values as of mid-2026. A single good master, around -14 LUFS and under -1 dBTP true peak, travels well everywhere: no need for one file per platform.
Should you aim for -14 LUFS?
It's the most stubborn misconception. No: -14 LUFS isn't a requirement, and locking to it often weakens the sound. Normalization equalizes loudness; it doesn't restore dynamics.
A master crushed to -6 LUFS, once turned down to -14 by the platform, will sound smaller than a -14 LUFS master that kept its transients. Normalization rewards dynamics, not the file's loudness.
So what is the right level?
The one the track asks for. Pop, hip-hop or electronic music live loud (often -7 to -9 LUFS); classical, jazz or ambient need to breathe (sometimes -16 LUFS or lower). The right instinct isn't to hit a number, but to serve the music and let normalization set the level.
What about true peak?
Above the level sits the ceiling. True peak measures the signal reconstructed between samples, where inter-sample peaks can exceed what your software shows. When the platform converts your file to a compressed format (AAC, Ogg, Opus), those peaks can clip. Hence the standard: stay under -1 dBTP, and under -2 dBTP for Amazon. Keeping that margin avoids distortion you don't hear at home but that appears on streaming.
What about social media?
TikTok, Reels and Stories don't always apply the same normalization, and listening often happens on a small speaker. A slightly more present master (around -9 to -12 LUFS) can carry better there. It isn't the same use as music streaming: if it matters to you, plan a dedicated version.
Key takeaways
- Every major platform normalizes loudness: louder doesn't win.
- -14 LUFS isn't a target to reach, it's a playback reference.
- After normalization, dynamics make the difference, not the file's loudness.
- Master for the track, not for a number.
- Keep true peak under -1 dBTP (-2 for Amazon) to avoid distortion on streaming.
- A single good master is enough for every platform.
A master made for your track
Lexpi adapts the level and holds the true peak for you. Listen to the result before you pay.