An ACX rejection email is the automated response ACX sends when your audiobook fails one of the eight required specs. The email arrives in your inbox with a list of technical failures that block your project from distribution.
These notifications lack context. They tell you a metric failed without explaining how that metric relates to your audio software. Fortunately, every automated specification is fully deterministic. When you change the required number, the file passes. There is no guesswork. The platform evaluates files against hard rules for RMS, true peak, noise floor, format, and silence spacing.[1]
If you adjust your files directly to the platform target, the automated analysis approves them instantly. Before you spend hours re-processing, measure your rejected chapter using the Free ACX Check. The tool examines the file locally in your browser and maps the exact parameters you need to fix.
Why did my file fail for being too loud?
The Excerpt: "Your file has an RMS level of -16.2 dB. Audible requires RMS to be between -23 dB and -18 dB."
What ACX is measuring: The platform measures RMS (Root Mean Square), which represents the average volume of your narration across the entire file.[1] The closer the number gets to zero, the louder the file is. A measurement of -16 dBFS is significantly louder than the required ceiling of -18 dBFS.
The most common root cause: Heavy normalisation paired with aggressive compression. Many narrators normalise their files to 0 dBFS or -1 dBFS, pushing the entire vocal track up until the average energy exceeds the safe streaming range. ACX requires lower overall levels to prevent listener fatigue during hours-long books.
The deterministic fix: You must lower the overall gain. If you are using a DAW, apply a gain reduction plugin or use the normalisation tool to target a lower average level. Do not aim for the exact edge of the limit (-18 dBFS).
The spec target: Master the file to target -20 dBFS. This drops the volume below the ceiling but keeps it comfortably above the negative floor, satisfying the automated measurement. See the detailed process breakdown in the ACX audio requirements guide.
Why did my file fail for being too quiet?
The Excerpt: "The RMS level of your file is -26.5 dB. The submitted file must have an RMS between -23 dB MS and -18 dB."
What ACX is measuring: The system sees the average vocal energy falling below the minimum volume threshold. -26 dBFS is far quieter than the -23 dBFS floor.[1] If the file is too quiet, listeners walking down a busy street will struggle to hear the dialogue over their environment.
The most common root cause: Submitting raw, unprocessed audio. Microphones record signals at safe, quiet levels to avoid distortion. Unmastered narration usually sits around -35 dBFS to -28 dBFS. If you skip the dynamic compression step, the audio remains too quiet to pass.
The deterministic fix: Apply dynamic range compression to reduce the distance between your loud words and quiet words. After compressing the file, normalise the RMS level upward.
The spec target: Target -20 dBFS. Note that raising the volume of the narration also raises the volume of the room tone. Keep an eye on your noise metrics while pushing the gain.
Why did my file fail for true peak clipping?
The Excerpt: "Your submitted audio has peaks that exceed -3 dB. Please ensure peak levels read no higher than -3 dB."
What ACX is measuring: True peak evaluates the loudest single transient in your file. The platform uses an oversampling formula to estimate where the analog soundwaves hit between the digital samples. If the audio touches -2.8 dBFS, it triggers a rejection.[1]
The most common root cause: Using a standard peak limiter instead of a true-peak-aware limiter. Basic limiters stop the digital sample at -3 dBFS. However, when the file is converted into MP3 or played through speakers, the curve of the audio wave briefly pushes past that sample point. ACX detects this inter-sample clipping.[2]
The deterministic fix: Replace your standard limiter with a true peak limiter. Set the limiter ceiling lower than the required threshold to build in a digital safety margin.
The spec target: Set your true peak ceiling to -3.5 dBFS. This hard boundary leaves enough room for the platform's compression systems to encode the audiobook for cellular delivery without crushing the vocal transients.
Why did my file fail for high noise floor?
The Excerpt: "There is excessive background noise in your file. The noise floor must read at or below -60 dB."
What ACX is measuring: The system analyses the quiet gaps between your words to determine the volume of your room tone. A noise floor of -55 dBFS is unacceptable because listeners will hear a noticeable hiss whenever you pause speaking.[1]
The most common root cause: A noisy recording environment magnified by volume adjustments. When you boost your RMS loudness to pass the volume requirements, you amplify the background static by the exact same amount. If your raw noise floor was -62 dBFS and you add 8 decibels of gain, your new noise floor is -54 dBFS.
The deterministic fix: You must separate the noise from the dialogue before boosting. Apply a gentle downward expander or a speech-aware noise gate. This attenuates the idle static during silences while leaving the spoken words untouched. Avoid heavy broadband noise reduction, as it causes metallic artifacts.
The spec target: Achieve a noise floor below -60 dBFS. For an in-depth walkthrough on managing static and hiss, read the audiobook noise floor guide.
Why did my file fail for format or bitrate?
The Excerpt: "Your file was submitted at 128 kbps. Audible requires all files to be presented at 192 kbps or higher in CBR."
What ACX is measuring: The platform scans the metadata of the submitted file to ensure it aligns with the Audible delivery infrastructure. The system rejects loose WAV files, low-quality bitrates, and Variable Bit Rate (VBR) encoding schemas.[1]
The most common root cause: Incorrect DAW export settings. Many audio editing programs dynamically alter bitrate to save hard drive space. Submitting mono files at 128 kbps sounds fine to the human ear, but it violates the platform archiving standard.
The deterministic fix: Delete the exported file, return to your session, and re-export the project. Open the advanced export properties block. Select MP3 format. Choose Constant Bit Rate (CBR). Force the bitrate higher.
The spec target: Export as 192 kbps MP3 using Constant Bit Rate. No other format is accepted.
Why did my file fail for silence padding?
The Excerpt: "Your file does not contain the required amount of room tone. Ensure 0.5 to 1 second at the head and 1 to 5 seconds at the tail."
What ACX is measuring: The system measures the length of non-spoken audio at the beginning and end of the track. If a file begins talking after 0.2 seconds, or if the tail drags for 8 seconds, the chapter fails.[1]
The most common root cause: Generating digital silence instead of pasting recorded room tone. If you use a software tool to create absolute zero silence at the start of the file, the transition to your natural room ambience causes an audible pop. ACX rejects absolute digital silence.
The deterministic fix: Highlight a section of quiet room tone from inside your recording booth. Copy it. Trim the start of your dialogue file. Paste exactly 0.5 seconds of the room tone at the beginning. Paste 3 seconds of the tone at the end.
The spec target: 0.5 to 1 second head room tone. 1 to 5 seconds tail room tone. Every file in the submission package (including the retail sample) must meet this standard.
How do I resubmit my files after fixing them?
Once you apply the deterministic fixes to your audio files, return to the ACX dashboard. You do not have to initiate a brand-new project. Navigate to the rejected title in your library view.
Locate the specific files noted in the rejection email. Use the interface to upload the corrected MP3 files, overwriting the faulty tracks. Replace every file that failed the automated parameters. It is highly recommended that you run a full batch export using a unified set of tools so consistency carries across all chapters.
When you replace the files and click submit, the project re-enters the automated analysis queue. Review the ACX submission checklist to verify your delivery shape before taking that final step.
Common Questions
Do I have to wait weeks again for the review after fixing the files?
No. Automated failures bounce back swiftly, sometimes in minutes. If you pass the automated analysis on a resubmission, you enter the human quality review queue immediately. You cannot circumvent the 10 to 14 day waiting period for the quality review.
The email says I failed for multiple issues. Does the order matter?
Yes. Always fix your format and channel count first. Then handle noise mitigation. Then compress and normalise for loudness. Finally, employ your true peak limiter. If you adjust loudness before gating the noise, you permanently bake the loud hiss into your file. See the ACX rejection fix guide for workflow routing.
Does ACX reject the retail sample for the same reasons?
Absolutely. The retail sample is subjected to the identical testing parameters as the main chapter files. It must maintain the precise RMS window, true peak limit, and head/tail silence.
Why didn't my DAW warn me about a true peak failure?
Most standard audio meters display sample peaks, which measure the loudest point among the digital dots. True peak reconstructs the analog wave between the dots. You need a dedicated true peak meter to see the overshoot that ACX detects.
What if the rejection email says the pacing is too fast or there are audible mouth clicks?
That is a human quality review failure, not an automated spec failure. A human heard the audio and found the performance lacking or the editing sloppy. You cannot standardise these issues away with a limiter. You must manually edit the tracks and re-record the rushed segments.
If spending an afternoon translating measurement formulas in your DAW sounds frustrating, let software handle it. ChapterPass masters audiobook files directly in your browser. It solves loudness deviations, catches inter-sample peaks, manages background noise, and shapes the correct output format without compromising your performance. Discover how easy submission can be at our pricing page.