In this lesson
Streaming loudness myths waste masters
One forum says master to -14 LUFS, the next says go much louder. Spotify documents how its normalization works, and the rules settle both arguments.
Updated · 6 min read · Lesson 33 of 44 in Mixing & Mastering
In short
- On its default setting Spotify turns louder masters down to -14 LUFS with a plain gain change, so the extra level is cancelled and the density stays.
- Quieter masters are only raised as far as their peaks allow, so a very dynamic master with high peaks can play quieter than its neighbours.
- Normalization is not everywhere: the web player, some devices and listeners who switch it off hear your file at its own level.
In this lesson (8)
Two rules that cannot both be right
One thread tells you to master at exactly -14 , because anything louder gets wrecked on Spotify. The next one says -14 sounds weak and real releases sit far louder. You bounce both versions, trust neither, and lose a week to a number.
Both claims start from a real feature, , and both get it wrong. Spotify publishes how its normalization works in detail. Once you know the rules, the decision goes back to where it belongs: how the song should sound.
Why it matters
The myth sets your limiter
Believe the -14 rule and you may leave a dense rap or club master under-limited, chasing a level the platform would have reached for you anyway. Believe the loud rule and you crush a ballad to win a level race that normalization cancels. Either way, a rumour sets the instead of your ears.
On its Normal setting, the default, Spotify turns a master louder than -14 LUFS integrated down until it measures -14. That is a plain volume change: it adds no distortion, no compression, and it removes none of the density you built. A master quieter than -14 is turned up, but only as far as its peaks allow, because Spotify leaves 1 dB of for lossy encoding.
Listen
Hear what normalization does to a loud master
A synth loop or a real mix, as a dynamic master and a loud, clipped one. Turn on streaming-style normalization and compare them at the same loudness.
Start with your speakers or headphones low.
The science
One gain, set from one number
Normalization is the simplest process in the chain. The service measures integrated loudness with ITU-R BS.1770, then applies one fixed gain to the whole track:
A -8 LUFS master on a -14 LUFS target gets dB. Every sample moves by the same 6 dB, so the distance from peaks to average, the punch, the stereo image and any distortion stay exactly as you mastered them. When would be positive, Spotify stops raising the track once its peaks are 1 dB below full scale. A -20 LUFS track that peaks at -5 can rise 4 dB, so it plays at -16 LUFS.
Where and how someone listens changes the result:
- Listener settings. Premium listeners choose Quiet (-19 LUFS), Normal (-14) or Loud (-11). On Loud, Spotify sets the level regardless of and runs a limiter on soft, dynamic tracks. It engages at -1 dB in sample values, with a 5 ms attack and a 100 ms decay. That is the one case where your master gets extra dynamics processing.
- Albums. When someone plays an album, Spotify normalizes the whole album at once, so the gain does not change between tracks. In shuffle or a playlist, each track is adjusted on its own.
- Places without it. Spotify says the web player and third-party devices such as TVs and speakers do not use normalization, and listeners can switch it off.
Apple's Sound Check works on the same principle: it measures each track, stores the result as metadata and raises or lowers playback, per song or per album (Apple, 2021). Apple's guide gives no target number, so do not master to a rumoured one.
Try it
Hear normalization before the platform does
0 of 6 steps done
If the louder master still wins at matched level, its density is doing real musical work and you can keep it. If it only wins with normalization off, the extra limiting buys level and nothing else.
Common mistake
Assuming quiet masters get turned up
A dynamic master at -18 LUFS with peaks near 0 dBTP has no room to be raised on the Normal setting, because the boost stops when the peaks reach -1 dB. It plays at -18 LUFS, quieter than the songs around it in a playlist. If you want a very dynamic master, accept that, or control the peaks so the boost has room.
The other mistake is forgetting where normalization does not happen. The web player, many TVs and speakers, and listeners who switch it off hear your file at its own level. "Master to -14 because everything ends up at -14" leaves those listeners out.
Takeaway
Master the song, then check the rules
Choose the level by matched-level listening. Then check it against the rules: louder than -14 LUFS means it will be turned down and should peak below -2 dBTP; quieter means it will only be raised as far as its peaks allow. Neither number grades your master. They tell you what each listener will hear, and that is all you need from them. Why heavy limiting tends to lose at matched level is covered in the lesson on loud masters after normalization.
LUFS
Loudness Units relative to Full Scale. A loudness measurement weighted toward how the ear hears. Streaming services use integrated LUFS, an average over the whole track, to match playback level.
Loudness normalization
A streaming service turning each track up or down so everything plays back at a similar loudness. A louder master is simply turned down further.
Limiter
A very fast compressor with a very high ratio. It stops peaks from going over a set ceiling.
Headroom
The space between the loudest peak in a signal and the level where it would clip.
dBFS
Decibels relative to full scale, the largest value a fixed-point file or converter can hold, so levels below it are negative numbers. Inside a floating-point mix a signal can read above 0 dBFS and stay intact; it clips where it leaves float or passes through a plugin that clips. dBTP is the same scale measured as true peak, and it can read above 0.
True peak
The highest level a waveform reaches between samples once it is converted to analog or decoded from a lossy file. It can sit above the highest sample value, so meters estimate it by oversampling.
3 questions
Check yourself
Sources
Show all 4Hide sources
- Apple. (2021). Apple Digital Masters [Technology brief]. www.apple.com/apple-music/apple-digital-masters/docs/apple-digital-masters.pdf
- International Telecommunication Union. (2023). Recommendation ITU-R BS.1770-5: Algorithms to measure audio programme loudness and true-peak audio level. ITU. www.itu.int/rec/R-REC-BS.1770/
- Spotify for Artists. Loudness normalization on Spotify. support.spotify.com/artists/article/loudness-normalization/
- Spotify Support. Volume normalization. support.spotify.com/article/volume-normalization/
Next lesson · 34 of 44
Loud masters can shrink after matchingNormalization plays a crushed master and a dynamic one at the same loudness. Then the only difference left is how far the drums rise above the mix.
Written by Virzy Guns
Songwriter and producer, founder of Virzy Guns Production. Now building HealingWave.