Loudness Targets and Mastering for Streaming Platforms
Let’s talk about loudness targets. You know, those magical numbers like -14 LUFS on Spotify or -16 LUFS on Apple Music that people act like are the holy grail of mastering. The truth is, chasing a number too hard can make your song sound like a smoothie someone forgot to stir. Sure, the meter looks pretty, but your mix loses energy, dynamics, and, frankly, its soul. It becomes technically compliant and emotionally undercooked. Consider reading my article on loudness, dynamics, and translation in mastering for a deeper look at how these factors interact.
Tl;dr: The song comes first. The loudness number comes second. Always.
What Loudness Nomralization Means
Most streaming platforms adjust playback levels so one track doesn’t blow out your ears while the next barely whispers. This is called loudness normalization.
Tracks louder than the platform’s target get turned down
Tracks quieter might get boosted
The kicker is that your carefully crafted peak and dynamics might shift depending on the listener’s device or app settings. That’s right, someone could be listening on a Bluetooth speaker in a coffee shop or turn normalization off entirely, and your mix could play at the level you mastered it anyway.
True story: I’ve spent hours trying to hit the “perfect” LUFS number and then realized it doesn’t matter for half of the listeners. Lesson learned. Meters do not attend the listening party. Humans do.
Normalization primarily comes into play when tracks from multiple sources are combined in playlists or shuffled streams. For a single release or a carefully sequenced album, your master’s relative dynamics and level relationships are preserved, so the listener will hear the song as you intended, even though minor differences can occur depending on the platform’s processing.
What You Should Focus On The Song, Not The Target
Each song has a personality. Some tracks thrive with big, wide dynamics, and some need controlled energy to hit hard. If you squash everything just to hit a number, you kill that personality.
By mastering for the song first, you make sure:
Tonal balance stays intact
Dynamics feel musical
Punch and emotion survive any platform processing
If the meters end up happy, that’s a bonus. If they don’t, that’s okay too. The listener doesn’t care if it’s -13 or -14 LUFS, they care if it feels right. No one ever listened to Stairway to Heaven and thought, “Great track, but I wish it averaged 0.8 LU louder.”
Practical Tips Without Focusing On Numbers
Leave a little headroom, such as -1 dBTP, to avoid nasty clipping surprises on streaming platforms.
Focus on listening to a sound system you know best rather than switching between multiple setups.
Use meters as guides, not rules. If the song sounds off but the LUFS are perfect, trust your ears.
Reference tracks are your friends. Pick songs in the same vibe and energy as yours and compare tonal balance and loudness.
Remember, the goal of mastering is to serve the music, not just hit a number on a meter.
Normalization Is A Moving Target
Normalization varies between streaming platforms and devices. Spotify may normalize aggressively while Apple Music barely adjusts levels, and some users may disable normalization entirely. The specifications can change at any time, as platforms frequently update their targets. Trying to optimize for every scenario is futile. Focus on making your track sound solid on the system you know best and trust the platform to handle the rest.
Summary
Loudness targets exist, but the song comes first
Loudness normalization ensures playback consistency, but you cannot control every listener
Master for dynamics, tonal balance, and musical emotion first, numbers second
Meters and LUFS are helpful, but your ears are the final judge
In the end, mastering is like tending to a tiny waveform. You can try to make everything neat and safe, but peaks and inter-sample gremlins can sneak past 0 dBFS. Focus on the song first, and the loudness targets will take care of themselves. For more on true peak and inter-sample peaks, see True Peak vs Inter-Sample Peaks.
