Audiobook Formatting and ACX Standards

Quick Answer: ACX requires audio files in 192 kbps constant bit rate MP3 format, at 44.1 kHz sample rate, with RMS loudness between -23 and -18 dBFS, peak levels at or below -3 dBFS, and a noise floor at or below -60 dBFS. Each chapter must be a separate file. Opening and closing credits are separate files. Files that don't meet any of these specifications are rejected automatically.

ACX (Audiobook Creation Exchange) is the gateway to distribution on Audible, Amazon, and iTunes. Its technical requirements are precise and enforced by automated quality checks before any human reviewer listens to your submission. Understanding these specs prevents rejection and delays.

The Technical Specifications

RMS (Root Mean Square) measures average loudness. Files that are too quiet (below -23) or too loud (above -18) are rejected. This range ensures consistent listening volume across all audiobooks on the platform.

Peak level is the loudest single point in the file. Keeping it at or below -3 dBFS prevents digital clipping (distortion caused by the signal exceeding the digital maximum). A -3 dBFS ceiling provides headroom for playback across different devices.

Noise floor is the level of background sound when nobody is speaking. Below -60 dBFS means the room and equipment are quiet enough that background noise is essentially inaudible. Recording in an untreated room with ambient noise often produces a noise floor above -60, which triggers rejection.

File Structure

One file per chapter. Each chapter of your audiobook must be a separate MP3 file. This allows listeners to navigate by chapter. File names should follow a simple convention: the book title or abbreviation followed by the chapter number (for example, "BookTitle_001.mp3," "BookTitle_002.mp3").

Opening credits. A separate file at the beginning containing the book title, author name, and narrator name. Example: "[Book Title], written by [Author Name], narrated by [Narrator Name]."

Closing credits. A separate file at the end containing the same information plus optional production credits and copyright notice. Example: "[Book Title], written by [Author Name], narrated by [Narrator Name]. Copyright [Year] [Author Name]. Production by [Producer, if applicable]."

Retail audio sample. A separate file, typically one to five minutes of narration from the beginning of the book (without opening credits), used as the preview sample on the Audible product page.

Room tone. Each chapter file should begin and end with 0.5 to 1 second of room tone (silence recorded in the same environment as the narration). This prevents the audio from starting or ending abruptly.

Common Rejection Reasons

Noise floor too high. The most common technical rejection. Recording in a room with air conditioning, computer fans, or street noise often produces a noise floor above -60 dBFS. Solution: record in a treated space or use noise reduction software (iZotope RX is the standard tool).

Inconsistent RMS across chapters. If Chapter 1 has an RMS of -19 and Chapter 5 has an RMS of -22, the overall submission may be flagged. Mastering should normalize all chapters to a consistent level within the -23 to -18 range.

Clipping or peaks above -3 dBFS. Usually caused by the narrator getting louder during an emotional passage or leaning into the microphone. A limiter applied during mastering prevents this.

Incorrect file format. Submitting WAV files instead of MP3, or variable bit rate instead of constant bit rate. Always export as CBR MP3 at 192 kbps.

Missing room tone at chapter start/end. Files that begin or end with a hard cut sound unnatural and may be flagged.

What Good ACX Formatting Looks Like vs What Does Not

Works well: - All files pass automated QC on the first submission. RMS, peaks, and noise floor are within spec across every chapter. File structure follows ACX conventions. - The audio sounds consistent from chapter to chapter. Volume, tone, and room sound match throughout the entire book. - The opening and closing credits match the metadata entered on the ACX platform.

Red flags: - The author submits files without checking them against ACX specs. Free tools like ACX Check (a plug-in for Audacity) measure RMS, peak, and noise floor in seconds. Skipping this check guarantees avoidable rejections. - Chapters vary noticeably in volume or background noise, indicating they were recorded in different sessions without consistent mastering. - Credits on the audio don't match the credits in the ACX listing (different narrator name spelling, missing subtitle). Any mismatch can delay approval.

The Takeaway

ACX's technical requirements are strict but predictable. Mastering your audio to meet the five core specs (format, sample rate, RMS, peak, noise floor) and structuring your files correctly (one per chapter, separate credits, room tone) prevents rejection. Test every file before submitting. For the full audiobook process, see the Audiobook Creation guide.

Q: What are ACX audio requirements? A: 192 kbps CBR MP3, 44.1 kHz sample rate, RMS between -23 and -18 dBFS, peaks at or below -3 dBFS, noise floor at or below -60 dBFS. Each chapter is a separate file. See What Are ACX Audio Requirements? for the full spec breakdown.

Q: How do you check if your audio meets ACX specs? A: Use the ACX Check plug-in for Audacity (free) or Auphonic (online mastering tool). These analyze your files and report whether they pass each specification.

Q: Can you submit AI-narrated audiobooks to ACX? A: As of 2026, ACX accepts AI-narrated audiobooks, but you must contact ACX support directly before submitting. Do not use the standard submission flow for AI-narrated titles. The same technical specifications apply regardless of narration method.