ACX Production Platform Specs — A Practical Guide
If you’re researching ACX production platform specs, you’re likely trying to answer one practical question: Can I produce an audiobook here without failing quality control? The honest answer is yes—but only if you understand the technical requirements before you start recording, not after you’ve already burned hours in the booth.
ACX—Audiobook Creation Exchange—is Amazon’s platform connecting rights holders with narrators and producers. It’s also the pipeline that gets finished audiobooks onto Audible, Amazon, and iTunes. The specs aren’t optional suggestions; they’re automated pass/fail gates that reject files before a human ever listens.
Here’s what the specs actually mean for your workflow, where newcomers stumble, and how to verify your setup before you waste a single take.
The Audio Specs That Decide Pass or Fail
ACX enforces strict technical standards on every submission. These are non-negotiable, and the automated checker catches violations instantly.
File format: Deliver MP3 files at 192 kbps or higher, with a 44.1 kHz sample rate. This matches standard professional audio production, so anyone with basic digital audio experience will recognize it.
Bit depth: Record source files at 24-bit. This provides more headroom during editing and mastering, which matters because ACX’s quality control checks for clipping and distortion. Recording at 16-bit works in a pinch, but you’ll have less room to fix problems in post.
RMS levels: This is where most new producers fail. ACX requires an average RMS (root mean square) of -23 dBFS to -18 dBFS, with a target of -20 dBFS. For reference, most music sits around -14 dBFS to -12 dBFS—audiobook standards are noticeably quieter. If your audio averages louder or quieter than this window, the automated checker rejects it.
Peak levels: Peaks must not exceed -3 dBFS. This leaves headroom and prevents distortion on playback systems, especially in cars and on phone speakers where clipping becomes painfully obvious.
Noise floor: ACX requires a noise floor of -60 dBFS or lower. A humming AC unit, a noisy computer fan, or an untreated room with noticeable room tone will fail this check. A quiet room typically measures around -70 dBFS to -80 dBFS, so you have some margin—but not much.
Consistency: ACX measures RMS levels across the entire recording. If chapter 3 averages -18 dBFS and chapter 7 averages -22 dBFS, the submission gets flagged. This is the most common rejection reason for experienced narrators who record across multiple sessions.
A concrete example: Steven Pacey, who narrates Joe Abercrombie’s First Law series, is celebrated for his dynamic range. But even his performances sit within ACX’s RMS window because mastering normalizes levels across the full recording. Dynamic performance doesn’t mean inconsistent levels—it means emotional variation within a consistent technical frame.
How to Verify Your Setup Before You Record
Before you commit hours to recording, run a quick test to confirm your chain works. This takes fifteen minutes and saves you from discovering problems after recording ten chapters.
Step 1: Record a 60-second test clip in your actual recording space, using your actual microphone and interface. Read a passage from the book you plan to narrate. Don’t use your “good” voice—use your normal narration volume and energy.
Step 2: Run ACX Check. ACX provides a free tool called ACX Check that analyzes audio files against their specs before submission. Upload your test clip and review the results. It checks RMS, peak levels, noise floor, and file format compliance. It’s not a guarantee of approval—human reviewers can still flag issues—but it catches the most common problems.
Step 3: Listen critically on multiple devices. Play your test clip on your phone speaker, in your car, and through headphones. If you hear background hum or hiss on any of them, your noise floor is too high. If the audio distorts when you get animated, your peaks are too hot.
Step 4: Check your room tone. Record 10 seconds of silence in your recording space. Listen to it at a moderate volume. If you hear anything—a refrigerator hum, traffic, an HVAC system—you need to address it before recording. A Portable Vocal Booth (around $100) is a budget-friendly option for narrators recording in untreated rooms, but acoustic foam panels and bass traps work better for permanent setups.
The practical implication: if your test clip fails ACX Check, don’t submit it anyway hoping for leniency. The automated checker will reject it, and you’ll receive a report detailing the issue. Fix the problem before recording the full book, not after.
The Production Workflow: From Raw Audio to Approved Audiobook
Understanding the specs is one thing. Understanding how they fit into the production process is another. Here’s the typical workflow:
Step 1: Audition or hire. Rights holders post projects, narrators audition with a sample, and the two parties agree on terms. You can also produce your own audiobook if you hold the rights.
Step 2: Record. Narrators record the manuscript in sections, usually chapter by chapter. Most professional narrators use a home studio with a high-quality microphone, an audio interface, and acoustic treatment. The Rode NT1 and Shure SM7B are common choices, but the mic matters less than the room—a $200 mic in a treated room beats a $2,000 mic in an echoey kitchen.
Step 3: Edit and master. This is where the specs come into play. You’ll remove mistakes, breaths that are too loud, and background noise. Then you’ll master the audio to hit ACX’s RMS and peak targets. Software like Audacity (free), Reaper (inexpensive), or Adobe Audition (subscription) all work.
Step 4: Upload and QC. ACX runs an automated quality check on your files. It measures RMS, peak levels, noise floor, and file format compliance. If your files pass, they go to a human reviewer. If they fail, you get a report telling you what to fix.
Step 5: Approval and distribution. Once approved, your audiobook goes live on Audible, Amazon, and iTunes. The timeline varies, but expect 2-4 weeks for the full process after submission.
The key takeaway: the specs aren’t bureaucratic hurdles. They exist because audiobook listeners are sensitive to inconsistent audio quality. A listener switching from a loud chapter to a quiet one at 1.5x speed on their commute will notice the difference immediately.
What Gets Flagged in Quality Control—and What Actually Happens
ACX’s QC process is more thorough than most people expect. Here are the most common rejection reasons and what they mean for your project:
Clipping or distortion. If your audio peaks above -3 dBFS, you risk distortion. This often happens when narrators get animated during intense scenes and push the microphone too hard. The fix is usually backing off the mic during loud passages or using a limiter during mastering.
Inconsistent levels. ACX measures the RMS level of each chapter. If one chapter averages -18 dBFS and another averages -22 dBFS, that’s a fail. The fix is usually a combination of compression and normalization during mastering. This is especially common for narrators who record across multiple days—your energy level naturally varies, and the audio reflects it.
Background noise. This includes hums, hiss, and room tone. The spec requires a noise floor of -60 dBFS or lower. If you’re recording in a space with noticeable room tone, you’ll need to use noise reduction tools like iZotope RX or Audacity’s noise reduction effect. But be careful: aggressive noise reduction can create a “swirly” or underwater sound that human reviewers will flag.
Mouth noises. This is a subjective check, but ACX reviewers do flag excessive lip smacks, tongue clicks, and saliva sounds. Some of this is normal—human mouths make sounds—but if it’s distracting, you’ll need to edit them out or adjust your recording technique. Drinking water between takes and keeping your mouth hydrated helps more than you’d think.
Mispronunciations. ACX doesn’t check this automatically, but the rights holder might. If you’re narrating a book with character names like Kvothe (from Patrick Rothfuss’s The Name of the Wind) or Daenerys (from George R.R. Martin’s A Song of Ice and Fire), you need to be consistent. Listeners will notice if you pronounce a name differently in chapter 12 than you did in chapter 3.
Here’s a useful data point: The Audie Awards, which recognize excellence in audiobook production, consistently honor narrators who demonstrate technical consistency alongside performance skill. The 2023 winner for Best Male Narrator, Edoardo Ballerini, is known for his precise pacing and clean production—both of which require attention to the same specs ACX enforces.
The Real Trade-Offs: What the Specs Don’t Tell You
The specs are clear, but they don’t tell you where the real difficulties hide. Here are the trade-offs that experienced producers learn the hard way.
The loudness trap. New narrators often record too quietly, thinking they’ll fix levels in post. But if your raw recording averages -30 dBFS, you’ll need significant gain during mastering, which amplifies background noise. The better approach is to record at a healthy level—aim for peaks around -6 dBFS during recording—so you have room to work without introducing noise.
The noise reduction trade-off. When your noise floor is borderline, you might be tempted to apply aggressive noise reduction. But heavy noise reduction creates an unnatural, “swirly” artifact that human reviewers can flag. The safer fix is to address the room itself—turn off the AC, move away from the computer fan, or add acoustic treatment. Noise reduction should be a last resort, not a first step.
The consistency problem. If you’re recording a 12-hour book over several weeks, your voice will naturally change. You might be slightly congested one day, or more energetic after a good night’s sleep. This is normal, but it means you need to match levels across sessions. The fix is to keep a reference track—a short clip of your target levels—and check it before each recording session.
The distribution conflict. ACX’s specs are specific to their platform. Other distributors like Findaway Voices and Authors Republic have their own requirements, which are often similar but not identical. If you plan to distribute through multiple channels, you’ll need to master your audio to meet each platform’s requirements. This is a common surprise for authors who assume one master file works everywhere.
A concrete example: Julia Whelan, who reads The Invisible Life of Addie LaRue and countless other titles, records from a home studio. Her production quality is consistent across dozens of audiobooks, which is exactly what ACX’s specs are designed to ensure. But that consistency comes from disciplined monitoring—checking levels, matching room tone, and reviewing each chapter before moving on.
Common ACX Spec Questions, Answered
Can I submit WAV files instead of MP3?
No. ACX requires MP3 files at 192 kbps or higher. You can master from WAV files, but your final submission must be MP3.
What happens if my audio fails QC?
You’ll receive a report detailing the issue. You can fix the files and resubmit. There’s no penalty for resubmission, but repeated failures can delay your project significantly. Each resubmission goes back into the queue, so a project that fails three times can take weeks longer than expected.
Do I need to match ACX specs if I’m publishing through another distributor?
No. Other distributors like Findaway Voices and Authors Republic have their own specs, which are often similar but not identical. If you plan to distribute through multiple channels, you’ll need to master your audio to meet each platform’s requirements.
Can I use AI narration on ACX?
ACX has specific policies around AI narration. As of 2024, ACX requires disclosure if AI narration is used, and some projects explicitly require human narration. Check the project listing before you audition.
What’s the fastest way to check if my files meet specs?
Use ACX Check before submitting. It’s free, it’s fast, and it catches the most common problems. It won’t catch subjective issues like mouth noises or mispronunciations, but it will catch format, RMS, peak, and noise floor violations.
What These Specs Mean for Your Next Decision
If you’re deciding whether to produce through ACX, here’s the practical takeaway: the specs are learnable, and the tools to meet them are affordable. A decent microphone, a quiet room, and free software like Audacity are enough to pass QC for most projects. The bigger challenge is consistency—matching levels across sessions, keeping your noise floor low, and delivering a clean performance.
If you’re already producing and hitting rejections, the fix is usually one of three things: your room needs acoustic treatment, your levels need more careful mastering, or you’re recording too hot and clipping. Run ACX Check on your files, review the report, and address the specific issue rather than guessing.
The specs aren’t designed to be difficult—they’re designed to ensure consistency across a massive catalog of audiobooks. When you listen to a professional production like Andy Serkis narrating The Hobbit or Moira Quirk reading Gideon the Ninth, you’re hearing the result of careful technical work that meets these exact standards.
If you’re just starting out, focus on three things: a quiet recording space, consistent microphone technique, and mastering to ACX’s RMS and peak targets. The tools and specs are learnable. The performance—your voice, your pacing, your interpretation—is what will set you apart.
Try Audible Free for 30 Days — Start your free trial on Amazon and get two free audiobooks.
Whether you’re producing your own audiobook or just curious about what goes into the listening experience, understanding these specs gives you a deeper appreciation for the craft. The next time you hear a perfectly produced audiobook, you’ll know exactly how much work went into making it sound effortless.
<!– cluster-navigation –>
Explore This Topic
- Back to Specs & Manuals
- Back to Time-Pressed Multitasker
Related guides in this cluster: