LUFS explained: measure loudness, understand streaming
LUFS stands for Loudness Units relative to Full Scale: a weighted measure of loudness over time. A negative value closer to zero means a louder measured signal: −10 LUFS is louder than −14 LUFS. Use integrated LUFS for a whole song and check peaks separately with true peak. No single reading tells you whether a master sounds good.
Integrated, short-term and momentary: which reading matters?
Measurement duration is part of the result. A chorus can be much louder than the complete track. Integrated measurement also uses gating, so extended silence does not simply pull the result down without limit.
| Reading | Time span | Use |
|---|---|---|
| Integrated (LUFS-I) | Complete measured track | Compare tracks and exports |
| Short-term (LUFS-S) | 3 seconds | Inspect verses and choruses |
| Momentary (LUFS-M) | 400 milliseconds | Inspect brief loudness changes |
LUFS, dBFS and dBTP are not the same
Sample peak in dBFS describes digital sample peaks. True peak in dBTP also estimates peaks between samples. LUFS instead describes weighted loudness. Two songs can reach the same peak yet sound very different in loudness.
A limiter can control peaks and allow greater average loudness. Excessive limiting can also weaken attacks and movement. Compare at matched loudness rather than simply chasing a higher number.
Does my master have to be exactly −14 LUFS?
No. Spotify specifies −14 LUFS for normal loudness normalization; this is a playback reference, not a universal delivery requirement. Settings, playback modes and platforms can differ. Spotify can turn quieter tracks up where available peak headroom allows.
Choose dynamics and loudness to suit the music. A dense club track and a quiet ballad do not need identical processing. Streaming normalization cannot repair distortion or excessive compression.
Worked example: −9 LUFS against a −14 LUFS reference
With gain reduction alone, a song measured at −9 LUFS would be played about 5 dB lower to reach −14 LUFS. Previously clipped attacks do not return.
Conversely, a −18 LUFS track would mathematically need 4 dB of gain. If its true peak is already −2 dBTP, only 1 dB is available before an assumed −1 dBTP ceiling without additional dynamics processing. This illustrates gain and headroom, not guaranteed platform playback.
How to check your finished export
Measure the actual exported file across the whole song. Compare the same section before and after processing at similar loudness. A short excerpt cannot replace a complete measurement.
- Check integrated LUFS and true peak together, not just a chorus.
- Use the streaming preview to inspect estimated gain changes; it is a simulation.
- If you hear pumping, distortion or weak attacks, ease the limiting and compare again.
Try it with your own audio.
These free tools help you apply the guide. Measurements are starting points, not automatic mixing decisions.
Further reading: Spotify: loudness normalization. Playback behaviour depends on settings; the streaming preview is an estimate, not the platform player.
Want to hear the difference before uploading your own track? Compare real original mixes and masters.
Use LUFS alongside true peak and listening. A master that serves the music matters more than hitting one exact number.