What is Loudness Normalization?
Loudness normalization adjusts a finished audio file so its average perceived volume hits a target level, which is how platforms make different uploads play back at a consistent loudness.
Updated · Editorial policy
Unlike peak normalization, which only scales the file so the single loudest sample reaches a ceiling, loudness normalization measures how loud the whole programme actually sounds and shifts it to a target. Streaming platforms apply this automatically on playback, turning loud uploads down and quiet ones up.
This is why over-compressing a mix to make it sound louder no longer works. If you master a podcast episode far above the platform target, the platform simply attenuates it, and you are left with a squashed, lifeless version of the same loudness everyone else got. Matching the target instead preserves your dynamics.
Measure with a loudness meter, aim for the platform's integrated target, and keep true peaks a little below zero so <a href="/software/later">later</a> encoding does not clip. Around -1 dBTP is a safe ceiling. Normalize the finished master, not individual clips, since the measurement describes the whole programme.
Where Loudness Normalization shows up
You will meet Loudness Normalization most often while comparing audio editing software - Riverside, Podcastle, Cleanvoice AI, and Descript all deal with it, though not in the same way or at the same price.
Related terms: LUFS.