Audiobook production has a uniquely stressful property: technical delivery specs are strict, and a single out-of-range file can delay an entire release. Audible / ACX is a common example with strict criteria that uploaded files need to fulfil before they're accepted.
That’s why we’ve added an additional loudness normalization method in Auphonic: RMS-based normalization. It’s designed specifically for workflows that need RMS targets, including Audible / ACX.
What ACX requires (loudness and peaks)
ACX asks for each file to meet a level measured in RMS and a maximum peak limit. The headline values as stated in their ACX audio submission requirements are:
- Average level: between -23 dB RMS and -18 dB RMS (per file)
- Peak level: peaks must be below -3 dB
In practice, this means you need two things: a consistent overall level, and enough headroom so transients never exceed the peak ceiling.
Recommended ACX settings in Auphonic
If you want one simple, reliable starting point, use the preset below.
It targets the middle of the ACX RMS window and applies the correct peak ceiling.
Web Interface
In the Production Form, open Loudness Normalization and set:
- Loudness Target: -20 dBRMS (Audible / ACX)
- Maximum Peak Level: -3 dBTP
- Normalization Method: RMS Levels (Audible / ACX)
Recommended Auphonic Loudness Normalization settings for Audible / ACX: -20 dBRMS, -3 dBTP, RMS Levels
Learn more about the Auphonic Loudness Normalization Settings!
API
For API users, send the parameter:
"loudnessmethod": "rms", "loudnesstarget": -20
Learn more about How the Auphonic API works!
LUFS vs RMS: why loudness measurement matters
Different measurement methods answer slightly different questions. Some aim to reflect how loud audio is perceived by a listener, others describe the average signal energy over time. Both are valid, but they are not interchangeable - and this distinction is exactly where ACX-related issues originate.
Understanding the difference helps explain why the same audio file can appear “in range” in one tool and fail a check in another.
Interested in learning more?
Here are detailed articles on
How Loudness Normalization works in Auphonic.
LUFS: standardized perceived loudness
LUFS (Loudness Units relative to Full Scale) is a standardized loudness measurement defined in international broadcast standards. Its goal is to reflect perceived loudness as consistently as possible across different types of content.
To achieve this, LUFS measurements:
- apply frequency weighting to approximate human hearing sensitivity
- measure loudness over time, not just instant levels
- use gating to reduce the influence of very quiet passages and long silences
This makes LUFS particularly robust for spoken-word content, music, and mixed material. Long pauses between sentences or chapters don’t disproportionately lower the loudness value, which means LUFS tends to align well with how loud a listener experiences the program overall.
RMS: average signal energy
RMS (Root Mean Square) is a way of calculating the signal energy that is no longer commonly used for loudness targets today. It measures the average energy of the audio signal over time, without attempting to model human perception.
In contrast to LUFS:
- RMS does not apply perceptual frequency weighting
- RMS typically does not include standardized gating
- silence and very quiet passages directly lower the average value
This means RMS values are strongly influenced by pauses, gaps between phrases, and overall dynamics. Two narration files with identical spoken content but different pause lengths can produce noticeably different RMS readings.
Why this matters for ACX
ACX specifies its loudness requirements in terms of RMS.
For audiobook production, this difference is crucial. ACX requires every individual file to fall within a strict RMS range, and even a small deviation can lead to rejection and delays in publication.
Auphonic’s LUFS normalization has been used by many producers over the years. But if your workflow is built around RMS targets, or you want a method that aligns more directly with ACX’s published wording, RMS normalization is now available.
Feedback
We’re continuously improving Auphonic based on real production workflows. If you have feedback on RMS normalization for audiobooks, ideas or wishes for other features or feeback for existing ones, please reach out via our Contact Page, email support@auphonic.com, or send us feedback on our results page.