← All guides

What Is LUFS, and Why Do Platforms Turn Your Audio Down?

You master a track as loud as you can get it, upload it to Spotify or YouTube, and it comes out quieter than tracks that were mastered less aggressively. That's not a bug or bad luck. It's LUFS normalization, and it rewards a completely different kind of "loud" than the one most people are optimizing for.

Peak loudness and perceived loudness aren't the same measurement

A peak meter only catches one thing: the single loudest instantaneous moment in a file. LUFS (Loudness Units Full Scale) measures something different, perceived loudness averaged over time. Two tracks can have an identical peak level, both technically touching 0 dB at their loudest instant, and still sound very differently loud when you actually listen to them, because one has much higher average energy sustained throughout while the other is mostly quiet with one sharp spike. Peak tells you about a single sample. LUFS tells you what your ears actually experience across the whole track.

The targets platforms actually use

Major platforms normalize incoming audio to a specific LUFS target rather than leaving loudness up to whatever the uploader chose. Spotify normalizes to around -14 LUFS. YouTube targets a similar level, also around -14 LUFS. Podcast platforms, including Apple Podcasts, generally recommend closer to -16 LUFS. When a file arrives louder than the platform's target, the platform automatically turns it down to match. There's no way to opt out and "win" by mastering louder; the platform just pulls it back down.

Why peak-normalized audio can end up quieter, not louder

This is the trap: peak normalization (turning a track up until its single loudest moment just touches the maximum) and loudness normalization (targeting a specific average LUFS value) solve two different problems, and people often reach for the wrong one. A track that's been pushed hard with peak normalization and heavy limiting to sound as loud as possible often has a very high LUFS value, well above what the platform targets, so the platform turns it down more than a track that was mastered sensibly to begin with. The result: the "louder" master can end up sounding quieter on the platform than a track mastered with proper LUFS targets in mind, because it got turned down further to reach the same destination.

What normalizing on Kit-Bin actually does

Volume Normalize measures a file's peak level and applies a single, uniform gain so the loudest moment lands at a target peak. That's peak normalization, useful for making sure a quiet recording uses the full available range without clipping, but it isn't the same thing as hitting a specific LUFS target the way a streaming platform's loudness normalization does. If you're preparing a track specifically for a platform's loudness target, peak normalization is a reasonable first pass, not a substitute for checking the actual LUFS value with dedicated metering. If you're combining multiple clips before normalizing, merge them first, then normalize the combined result once.