ACX audio requirements at a glance
| Requirement | Target | Why it matters |
|---|---|---|
| RMS loudness | -23 to -18 dB | Consistent listening volume across books |
| Peak level | -3 dB or lower | Headroom so audio never distorts |
| Noise floor | -60 dB or lower | Quiet background between words |
| Format | MP3, 192 kbps+, CBR | The format ACX ingests |
| Sample rate | 44.1 kHz | Standard for audiobook delivery |
| Channels | All mono or all stereo | Consistent playback between chapters |
| File length | 120 minutes or less | One chapter per file, of manageable size |
| Room tone | 1–5 s at the start and at the end | Natural transitions between chapters |
Understanding the three levels
RMS loudness
RMS is the average loudness of the whole file. Narration that sits around −20 dB RMS sounds full without being tiring. Too quiet and listeners turn the volume up, then get blasted by the next book; too loud and the audio sounds squashed.
Peak level
The peak is the single loudest sample. Keeping peaks at or below −3 dB leaves room for the encoding and playback chain. Plosives (“p” and “b” sounds), laughs and shouted lines are the usual culprits; the level view marks every place in red.
Noise floor
The noise floor is the level of the quietest half-second: your room tone. Computers, fans, traffic and electrical hum raise it. A floor at −60 dB or lower means the pauses between sentences sound truly quiet.
A good order of operations
- Record in the quietest space you can and check the noise floor first: it is the hardest thing to fix later.
- Edit, then raise the level so the RMS sits near −20 dB.
- Add a limiter with a −3.5 dB ceiling to tame peaks.
- Export MP3 at 192 kbps CBR, 44.1 kHz, then run this checker on every chapter.