A pragmatic loudness target for podcasts

Music streaming long ago settled on -14 LUFS as the loudness war ceasefire. Podcasts are a different category of audio with a different listening context, and the right target reflects that.

The number

For spoken-word podcasts, the working target is:

• -16 LUFS integrated, with a true peak ceiling of -1 dBTP, and a loudness range (LRA) under 8 LU.

That’s the Apple Podcasts recommendation, it’s also what the BBC, NPR, and most established podcast networks land on. It’s the right default unless you have a reason to deviate.

Why -16 and not -14

Podcasts get listened to in different environments than music: walking around with imperfect earbuds, in a car, on a kitchen speaker, in a noisy commute. In those environments, raw loudness matters less than intelligibility. A podcast hitting -14 LUFS feels aggressively loud in the same scenarios where music at -14 feels lively.

-16 LUFS gives you about 2 dB of breathing room, which is what spoken word needs. The dialogue sits comfortably, the listener can turn it up to taste, and the loud moments don’t blow ears.

Why LRA matters more than peak

Spoken-word audio with a wide loudness range (say, 12+ LU) is exhausting to listen to. The listener constantly adjusts the volume — louder during the host’s quieter sections, quieter when a guest gets animated. A target LRA under 8 LU is what makes a podcast feel “produced” instead of “captured.”

This is the place where conservative compression pays off. A 2:1 or 3:1 ratio on the master bus with a slow attack and medium release will pull your LRA down to a manageable range without flattening the dynamics that make speech sound natural.

True peak ceiling

-1 dBTP is the right ceiling because:

• Most podcast hosts (Spotify, Apple, Castbox, Pocket Casts) apply some normalisation.
• Lossy codecs (AAC, MP3) can push intersample peaks above 0 dBFS, causing audible distortion on playback even when your sample peaks are below 0.

True peak metering (any modern limiter or meter — Waves WLM, Youlean, FabFilter Pro-L2) accounts for this. Set your limiter ceiling to -1 dBTP and let it do its job.

What about music in podcasts?

The hard case: a podcast that uses music beds, sound effects, or scripted theme music. The conventional approach is to mix everything to the spoken-word target, but that often makes the music feel quiet and lifeless.

Two pragmatic options:

1. Duck the music: keep it 8–12 dB below the dialogue during voice, then bring it up between sections. Most modern podcast tools have ducking built in.
2. Mix sections separately: theme music gets its own loudness pass at -14 LUFS, voice sections at -16 LUFS. The transition between them is the perceived “drop” or “lift” that listeners interpret as the production telling them something is happening.

Don’t try to hit a single number across the whole episode if music is structurally part of the show.

Tools

For QA, any of these work:

• Waves WLM (paid, industry standard)
• Youlean Loudness Meter (free version is plenty)
• FabFilter Pro-L2 (limiter + meter combined)
• VoiceLab QA (this site) — for fast QA of an existing recording, not for live monitoring

For mastering, the BBC R128 plug-in (or your DAW’s loudness target preset) automates a lot of the painful manual work.

Related

• Why filler density matters more than filler count
• Estimating room echo without RT60