Loudness Normalization: Hitting the Right LUFS Target

Loudness normalization is the process of measuring how loud a track feels, then adjusting it toward a target the platform expects. The unit is LUFS, short for loudness units relative to full scale. Streaming services read that number and turn loud uploads down to a common level.

That last part is why you care. If your master is much louder than the target, the platform simply reduces it, and any distortion you baked in comes along for the ride. Over-compressing for volume buys you nothing on playback.

What the numbers mean

Integrated loudness measures the whole program. Short-term loudness tracks a rolling window and shows where a section drifts loud or quiet. True peak catches inter-sample spikes that a regular meter misses, which matters because those spikes clip after conversion.

A podcast built for -16 LUFS integrated with a true peak near -1 dBTP will sit comfortably on most platforms. Video-first channels often aim a little lower. Check what your destination asks for and hit that number rather than guessing.

Loudness targets are a contract with the platform. Hit the number and your mix plays back the way you heard it in the room.

Normalize for feel, then check the peaks

I set loudness first, listening for a consistent feel across the whole episode, then I confirm true peak afterward. Loudness and flavor are two separate jobs, and mixing them together is how people end up with a track that measures right but sounds pinched.

If a section reads hotter than the rest, fix it in the mix rather than the limiter. A quiet passage that sits five LUFS under the rest is a mixing problem, and a limiter will only smash the loud parts while leaving the gap intact.

Common ways the target goes wrong

• Measuring only the loudest section and assuming the rest matches, when integrated loudness is an average across the whole file.
• Ignoring true peak until after export, when inter-sample clipping is already locked into the file.
• Slapping a limiter on the master to chase a number instead of balancing the mix underneath it.
• Matching loudness but not tone, so the track hits the target and still sounds thin on small speakers.

Key takeaways

• Learn your platform target and treat it as a fixed goal, not a suggestion.
• Measure integrated loudness across the full program, not just your favorite minute.
• Keep true peak below -1 dBTP so conversion does not introduce clips.
• Fix loudness drift in the mix, because a limiter cannot repair a bad balance.