How to Fix Every ACX Rejection Issue: The Complete Troubleshooting Guide

ACX rejections fall into two categories: technical failures (measurable specs your file didn't meet) and quality failures (recording issues caught by human review). Technical failures are fixable with reprocessing. Quality failures usually require re-recording or careful editing. This guide is the troubleshooting manual that takes you from the rejection email to the specific fix, covering every common rejection reason, ranked by frequency, with exact settings and the correct processing order.[1]

If your email mentions multiple issues, fix all of them before resubmitting. Don't fix one and hope the others slide through. They won't. And the order you fix them matters. Fixing things in the wrong sequence means re-doing earlier steps. The correct fix order is covered below.

How Do You Read the ACX Rejection Email?

ACX rejection emails typically include one or more of these indicators:

• "Audio levels do not meet our requirements" → RMS loudness, true peak, or both
• "Noise floor exceeds acceptable levels" → Noise floor above -60 dBFS
• "Audio does not meet format requirements" → Sample rate, channels, bitrate, or encoding mode
• "Inconsistent audio quality" → Chapter-to-chapter variation
• "Room tone / background noise" → Noise floor or environmental noise
• "Recording quality issues" → Human review caught artefacts
• "Extraneous sounds" → Clicks, pops, mouth noise, or environmental sounds

Find your rejection reason below and follow the fix.

How Do You Fix an RMS Loudness Rejection?

RMS loudness outside the -23 to -18 dBFS range is the single most common ACX rejection.[1] RMS is a measurement of average signal power across your entire chapter file. It represents how loud the chapter sounds over its full duration. Most home studio recordings come in too quiet, typically around -30 to -40 dBFS, because narrators record conservatively to avoid clipping. For a focused deep-dive on this specific rejection, see the RMS loudness rejection guide.

Diagnosis: Measure your file's RMS loudness. In Audacity, select all audio → Analyze → Contrast → read "RMS of selection." The value must be between -23 and -18 dBFS. For details on what RMS means and how ACX measures it, see the ACX audio requirements guide.

If RMS Is Below -23 dBFS (Too Quiet)

This is the most common rejection. Your file is too quiet. For a step-by-step walkthrough of fixing loudness specifically, see the loudness fix guide.

Fix, Compression + Gain:

1. Apply gentle compression: 2:1 ratio, -20 dBFS threshold, 15–20ms attack, 100–200ms release
2. Measure the new RMS
3. If still below -23, apply positive gain (Effect → Amplify in Audacity) to bring RMS to approximately -20 dBFS
4. Re-check that peaks haven't exceeded -3 dBFS (compression and gain both raise peaks)

Fix, Loudness Normalisation:

1. Use a loudness normalisation tool to target -20 dBFS RMS
2. Audacity v3.x supports both LUFS and RMS normalisation modes. Use the RMS mode for ACX
3. Verify both RMS and peaks after normalisation

After fixing: Check noise floor. Boosting gain to fix RMS also boosts background noise by exactly the same amount. Your noise floor may now exceed -60 dBFS even if it was fine before. A raw recording at -65 dBFS noise floor boosted by 10 dB becomes -55 dBFS, above the limit.

If RMS Is Above -18 dBFS (Too Loud)

Less common. Your file is over-compressed or over-amplified. Stacking multiple compressors in a chain often pushes the average past -18 dBFS.

Fix:

1. Reduce gain (Effect → Amplify with a negative value)
2. If you used compression, reduce the ratio or raise the threshold. If using multiple compressors in series, remove one.
3. Re-verify RMS lands between -23 and -18 dBFS. Target -20 dBFS for safe margin.

How Do You Fix a True Peak Rejection?

True peak is the maximum amplitude of the reconstructed analog waveform between digital samples.[2] ACX requires all true peaks to stay below -3 dBFS. Standard sample-peak meters in most DAWs can miss inter-sample peaks, which exceed the highest sample value by 0.5–1.5 dB. That's the difference between sample peak and true peak measurement. For a complete explanation of true peak and how to measure it, see the true peak guide.

Diagnosis: Measure true peak using a true-peak meter (not a standard sample-peak meter). Youlean Loudness Meter (free) or FabFilter Pro-L 2 can measure this.

If True Peak Is Above -3 dBFS

Fix:

1. Apply a true-peak-aware brickwall limiter
2. Set the ceiling to -3.1 dBFS (leaving margin for measurement precision)
3. If using Audacity's built-in limiter (which is sample-peak only), set the ceiling to -3.5 dBFS to compensate for inter-sample peaks
4. After limiting, re-check RMS, since limiting can reduce loudness, potentially pushing RMS below -23 dBFS

Critical step: Verify the final MP3, not the WAV. MP3 encoding can raise true peak levels by 0.5–1 dB. A WAV at -3.0 dBFS may become -2.5 dBFS after encoding. This is the most common reason files pass as WAV but fail as MP3.

Why Do Files Pass in the DAW but Fail ACX's Check?

You're probably measuring sample peaks, not true peaks. Inter-sample peaks can exceed the highest sample value by 1–3 dB. Switch to a true-peak meter and re-measure. See the ACX audio requirements guide for detailed information on true peak measurement.

How Do You Fix a Noise Floor Rejection?

The noise floor is the level of background sound present when nobody is speaking.[1] ACX requires it to stay below -60 dBFS, measured as the RMS of silent sections in your final MP3. At this level, background noise is inaudible to most listeners on consumer devices at normal volumes. For a focused deep-dive on this specific rejection, see the noise floor rejection guide.

Diagnosis: Find a silent section in your recording (no speech, at least 1 second). Select it and measure its RMS. Must be below -60 dBFS. For a deep dive into noise floor specifically, see the complete noise floor guide.

If Noise Floor Is -55 to -60 dBFS (Marginal)

Fix, Noise Gate:

1. Apply a noise gate to silence audio during pauses
2. Settings: threshold just above your room tone (typically -55 to -45 dBFS), level reduction -60 dBFS, attack 1–5ms, hold 200–500ms, decay 100–200ms
3. This replaces room tone in pauses with near-silence, which lowers the measured noise floor

Fix, Light Noise Reduction:

1. Profile a noise sample (select pure room tone, at least 0.5 seconds)
2. Apply noise reduction at 6–8 dB per pass
3. Preview before committing. Listen for artefacts
4. Re-measure noise floor

If Noise Floor Is -50 to -55 dBFS (Moderate)

Fix, Spectral Noise Reduction:

1. Use iZotope RX (paid) or Audacity's built-in noise reduction (free)
2. Profile the noise from a silent section
3. Apply 6–8 dB reduction per pass
4. Apply in two passes rather than one aggressive pass for better quality. Two passes of 6 dB sounds better than one pass of 12 dB
5. Preview carefully. This level of reduction often introduces watery or metallic artefacts

If Noise Floor Is Above -50 dBFS (Severe)

Heavy noise reduction at this level almost always creates audible artefacts that may trigger a quality rejection even if the noise floor number passes. You have two realistic options:

1. Re-record in a quieter environment. This is the clean solution.
2. Aggressive treatment with professional tools (iZotope RX Dialogue Isolate). Results vary. Listen critically. If it sounds unnatural, the human reviewer will catch it.

Why Does Noise Floor Fail After Mastering When the Raw Recording Was Fine?

Boosting gain to meet RMS targets also boosts background noise by exactly the same amount. If your raw recording has a noise floor of -63 dBFS and you apply 6 dB of gain, the noise floor becomes -57 dBFS, above the limit. The fix: apply noise treatment before loudness adjustment in your signal chain.

How Do You Fix a Format Rejection?

ACX requires mono MP3 at 192 kbps CBR and 44,100 Hz sample rate.[1] Every one of these parameters must match exactly.

Common mistakes: Submitting WAV or FLAC files instead of MP3, using variable bitrate (VBR) encoding, leaving the sample rate at 48 kHz from a video workflow, or exporting stereo instead of mono.

Wrong Sample Rate (Not 44,100 Hz)

Fix: Resample to 44,100 Hz using your DAW's resample function. In Audacity: Tracks → Resample → 44100 Hz. Re-master after resampling, since processing at the wrong sample rate may have produced different results.

Stereo Instead of Mono

Fix: Downmix to mono. In Audacity: Tracks → Mix → Mix Stereo Down to Mono. Check for phase cancellation after downmixing and re-verify all specs.

Wrong Bitrate or VBR Instead of CBR

Fix: Re-export with explicitly set encoding: MP3, 192 kbps, Constant Bit Rate. If your encoder doesn't have a CBR option prominently displayed, switch to LAME encoder settings where you can set it explicitly. After export, verify the file properties with a tool like MediaInfo or ffprobe. Don't trust the export dialog alone.

How Do You Fix a Silence Padding Rejection?

ACX requires 0.5 to 1 second of room-tone silence before the first speech and 1 to 5 seconds after the last speech.[1]

Fix if head silence is too short: Add silence at the beginning. In Audacity: click at position 0.0, Generate → Silence, set duration to bring total head silence to 0.75 seconds.

Fix if tail silence is too short: Add silence at the end. Target 3.0 seconds for comfortable margin.

Fix if either is too long: Trim the excess silence.

Important: Use room tone for padding, not digital silence (absolute zero). The transition from absolute silence to room tone creates an audible click. Record several seconds of room tone in your recording space and use that instead. Apply a short 50ms crossfade at each transition to eliminate clicks.

How Do You Fix a Recording Quality Rejection?

Recording quality problems require going back to the raw audio. No mastering tool can fix these. They're flagged by ACX's human reviewers, not the automated check.

Clicks, Pops, and Mouth Noise

Diagnosis: Put on headphones and listen to the flagged chapter(s) carefully. Clicks appear as sharp spikes in the waveform. Mouth noise is the sticky, wet sound of lips and tongue.

Fix: Use a de-click tool (iZotope RX De-click) or manually select and repair in your DAW. For persistent mouth noise: hydrate before recording, use lip balm, position the mic slightly off-axis.

Room Echo and Reverb

Diagnosis: Clap your hands in your recording space. If you hear a noticeable echo or ring, your room needs treatment.

Fix for current recordings: iZotope RX De-reverb can reduce mild reverb, but heavy reverb cannot be fully removed. If severe, re-recording in a treated space is the only reliable fix.

Fix for future recordings: Add soft absorptive materials: bookshelves full of books, heavy curtains, acoustic foam panels. Record in a small space rather than a large one. Position your microphone 6–8 inches from your mouth. A closet full of clothes is a surprisingly effective vocal booth.

Inconsistent Audio Between Chapters

Diagnosis: Listen to the first 10 seconds of several chapters in sequence. Do they sound like they were recorded in the same place, by the same person, with the same setup?

Fix: Process all chapters through the same mastering chain with identical settings. Target the same RMS value for all chapters (e.g., -20 dBFS). If chapters are fundamentally different (different room, different mic, different distance), consider re-recording for consistency.

Narration Issues

Performance or narration quality problems (pacing inconsistency, pronunciation errors, delivery problems) require re-recording the affected sections. No technical fix exists. Listen to professional audiobooks in your genre for pacing reference.

What Order Should You Fix Multiple Issues?

If ACX flagged more than one problem, fix them in this sequence. Some fixes affect measurements that other fixes depend on.

1. Format conversion first. Get to 44,100 Hz mono before anything else. All subsequent measurements depend on being in the correct format.
2. Noise floor second. Apply noise gates and noise reduction on the quiet signal before boosting gain. Normalising first makes noise louder and harder to clean.
3. Loudness (RMS) third. Bring the average level to the -23 to -18 dBFS range. Compression, then gain adjustment.
4. Peak limiting fourth. After loudness adjustment, peaks will be higher. Apply the true peak limiter to bring them below -3 dBFS.
5. Silence padding fifth. Add head and tail silence after all processing. You don't want the limiter or compressor acting on your silence regions.
6. MP3 encoding last. Always encode to MP3 as the final step, then verify the encoded file.

Why this order matters: If you limit peaks before adjusting loudness, the loudness adjustment pushes peaks back above the ceiling. If you pad silence before noise treatment, the noise gate may not handle the transitions correctly. Each step builds on the previous one. Fix things out of order and you end up re-doing earlier steps, or worse, shipping a file that passes one spec but fails another.

How Do You Resubmit After a Rejection?

After fixing the identified issues:

1. Re-master the affected files (or all files if the fix affected your mastering chain)
2. Re-verify every spec on every file, not just the ones that were flagged. Fixing one problem often shifts another measurement.
3. Listen to at least 3 random chapters from start to finish as a quality spot-check
4. Submit through ACX following their re-submission process

When resubmitting, it's tempting to fix only the flagged issue. But always re-verify every spec on the final MP3. The ACX audio requirements guide covers all eight specs in detail.

How Do You Prevent Future Rejections?

The most effective prevention is verification before submission. Check every spec on every file before uploading. See the ACX platform page for a quick-reference summary of all requirements. Automated tools like ChapterPass verify all specifications during processing, so you know whether your files pass before you submit. For answers to common questions about the process, see the FAQ.

The second most effective prevention is a consistent recording environment and workflow. Same room, same mic, same position, same gain, same processing chain for every chapter. Consistency prevents the chapter-to-chapter variation that triggers quality reviews. Understanding the difference between mastering and mixing also helps you focus your processing on what matters. For clarity on what automated tools handle and what they don't, see what ChapterPass does and doesn't do.