What LUFS should your podcast be?

What LUFS actually measures

LUFS is loudness as a person experiences it, measured over time. That is the important part. Peak level — the number most editing software shows biggest — only tells you the single loudest instant in the file, which is almost never what a listener notices.

Two files can peak at exactly the same level and be wildly different in perceived loudness. LUFS is the measurement that captures the difference.

The numbers people use

Most spoken-word podcasts are delivered around −16 LUFS in stereo, or about −19 in mono. Broadcast targets are often lower, and music streaming platforms publish their own numbers.

Do not treat any of these as law. Check the platform you actually publish to — they say so on their own documentation — and be consistent across your own back catalogue, which matters more than hitting a number exactly.

Why the platform changes it anyway

Most platforms normalise on the way in. If you deliver far louder than their target, they turn it down — and everything you did to make it loud is undone, except the damage. A track squashed to sound competitive and then turned down is quieter *and* flatter than one delivered correctly.

That is the real argument for hitting the target: not compliance, but making sure what the listener hears is what you mixed.

Normalising is not compressing

Normalisation moves the whole file to a target. Compression squeezes the distance between quiet and loud within it. They are different jobs and people confuse them constantly.

If your episode swings wildly in level, normalisation will not fix that — it will just move the swing somewhere else. Fix the source first: the noise floor, the dead air, and the distance between the speaker and the microphone.

Do it last

Loudness is the final pass, after noise and silence are dealt with. Normalising a file that still contains a noise floor means normalising the noise along with the voice, and removing it afterwards leaves you somewhere other than your target.

In short: Aim for about −16 LUFS stereo for talk, check your platform, be consistent across episodes, and normalise last.