You can check most audiobook specs for free using Audacity, MediaInfo, and Youlean Loudness Meter. The gap is true peak. Audacity's ACX Check plugin measures sample peak, not true peak, so a file that passes Audacity can still fail ACX's true peak check by 0.5 to 1 dB. This guide covers every free verification method, what each tool can and cannot check, and when manual checking costs more time than it saves.
If your files have already been rejected and you need specific fixes, the ACX rejection troubleshooting guide walks through each failure type. For the full list of specifications, see the ACX audio requirements guide.
What specs does ACX check on every file?
Before choosing tools, you need to know what to verify. ACX enforces eight technical specifications on every chapter file through automated quality checks:[1]
- RMS loudness: the average signal power across the full file, measured in dBFS (decibels relative to full scale, where 0 dBFS is the maximum level a digital audio system can represent). Must fall between -23 and -18 dBFS.
- True peak: the maximum level of the reconstructed analog waveform between digital samples, measured via 4x oversampling. Must stay below -3 dBFS.
- Noise floor: the background sound level during silent passages (room tone, HVAC, electrical hum). Must stay below -60 dBFS.
- Sample rate: 44,100 Hz (CD-quality audio).
- Channels: mono.
- Format: MP3, 192 kbps CBR (constant bit rate, meaning every second of audio uses the same amount of data).
- Head silence: 0.5 to 1 second of room tone before narration begins.
- Tail silence: 1 to 5 seconds of room tone after narration ends.
You must verify all eight on the final MP3 file, not on the project file, not on an intermediate WAV export, but on the actual MP3 you will submit.[1] MP3 encoding changes peak levels by 0.5 to 1 dB and can shift RMS slightly. Checking the wrong file version is one of the most common reasons audiobooks pass local checks but fail ACX's review.
Most audiobook distributors use the same or similar specifications, so passing ACX's check means your files will work everywhere.
Can Audacity check ACX specs for free?
Audacity is the most widely used free audio editor, and its ACX Check plugin is the standard free verification method for audiobook producers.
What Audacity's ACX Check measures
The ACX Check plugin (built into modern versions of Audacity) measures three specifications:
- RMS loudness: whether it falls between -23 and -18 dBFS
- Peak level: whether it's below -3 dBFS
- Noise floor: whether it's below -60 dBFS
To use it: open your MP3 file in Audacity, select the entire track (Ctrl+A or Cmd+A), then go to Analyze → ACX Check. The plugin reports pass or fail for each metric.
The critical limitation: sample peak vs true peak
The ACX Check plugin measures sample peak, not true peak. Sample peak looks only at the discrete digital sample values in your file. True peak estimates the analog waveform between samples using 4× oversampling, and that reconstructed waveform can exceed the highest sample by 0.5 to 1.5 dB.
ACX checks true peak.[1] A file that passes Audacity's peak check at -3.0 dBFS can have a true peak of -2.5 dBFS or higher, enough to fail ACX's review. This is the single most important gap in free verification tools.
What Audacity doesn't verify
Beyond the true peak limitation, the ACX Check plugin doesn't check:
- File format (MP3, bitrate, CBR vs VBR)
- Sample rate (44,100 Hz)
- Channel configuration (mono)
- Silence duration at head and tail
For format and sample rate, look at the bottom-left of the Audacity window after opening your file. It displays the project sample rate and channel configuration. For detailed format verification (bitrate, CBR vs VBR), you need MediaInfo or a command-line tool.
How to interpret RMS readings correctly
RMS is an average across the entire file, so a chapter with both very quiet and very loud sections can show an acceptable RMS while still sounding inconsistent. If your RMS measures -20 dBFS but your narration jumps between -15 dBFS (loud) and -28 dBFS (quiet), the chapter will sound uneven even though the number passes. Compression reduces this dynamic range before normalisation. ACX's human reviewers listen for this kind of inconsistency, so an RMS reading within range is necessary but not sufficient for acceptance.
Audacity alternatives for RMS checking
Adobe Audition: Go to Window → Amplitude Statistics, then click Scan. The "Total RMS Amplitude" field shows your value directly. Audition also provides true peak measurement in the same scan.
iZotope RX: Open the Loudness tab in the metering panel. It displays RMS alongside LUFS, true peak, and other loudness measurements in real time, useful if you already own the software.
How do you check true peak for free?
True peak measurement is the single most important reason to add a tool beyond Audacity. ACX enforces a true peak ceiling of -3 dBFS.[1] The -3 dBFS ceiling exists because MP3 encoding can raise peak levels, and without headroom, the encoded file may clip and distort on playback devices.
A file can have every sample below -3 dBFS but still have inter-sample peaks that exceed the limit. This happens because digital audio is a series of discrete samples, and the actual analog waveform between samples can be higher than either neighbouring sample.
Youlean Loudness Meter (free, BS.1770-compliant)
Youlean Loudness Meter is a free VST/AU plugin and standalone application that provides detailed loudness analysis including true peak measurement.[2] It complies with the ITU-R BS.1770 standard, the same algorithm broadcast professionals use for loudness measurement.
What it measures:
- Integrated loudness (LUFS, Loudness Units Full Scale, a frequency-weighted loudness measurement that correlates closely with RMS for spoken word)
- True peak (not just sample peak)
- Loudness range (dynamic variation across the file)
- Short-term and momentary loudness
To use it: load your MP3 file in Youlean's standalone mode or as a plugin inside a DAW (Audacity, Reaper, Adobe Audition), play the file through the meter, and read the integrated loudness and true peak values when playback completes. The true peak explained guide covers the measurement in depth.
Adobe Audition (true peak built in)
If you have access to Adobe Audition, Amplitude Statistics includes a True Peak field when you click Scan. This is the most straightforward built-in option for true peak measurement, with no plugins needed.
iZotope RX
The metering panel in iZotope RX displays true peak alongside RMS and other measurements in real time. If you already own RX for noise reduction or other editing work, you don't need additional tools for true peak verification.
FFmpeg loudnorm (command line, free)
For technically inclined producers, FFmpeg's loudnorm filter measures true peak from the command line:
ffmpeg -i input.mp3 -af loudnorm=print_format=json -f null -
This outputs a JSON block including input_tp (input true peak). It uses the ITU-R BS.1770 algorithm and is the most accurate free measurement method available. FFprobe (part of FFmpeg) also reads file metadata:
ffprobe -v error -show_entries format=format_name,bit_rate,duration -show_entries stream=sample_rate,channels,codec_name input.mp3
FFprobe and FFmpeg are ideal when you have many files to check. You can script them to batch-analyse every chapter file and flag any that fall outside specifications. For a 30-chapter audiobook, this takes seconds compared to hours of manual checking in a GUI.
How do you verify file format and metadata?
MediaInfo is a free, open-source tool that reads file metadata, essential for verifying the specifications that Audacity's ACX Check doesn't cover.
What MediaInfo checks
- File format: MP3, WAV, FLAC, etc.
- Bitrate: 192 kbps, CBR vs VBR
- Sample rate: 44,100 Hz
- Channels: mono vs stereo
- Bit depth (for WAV/FLAC files)
How to read MediaInfo output
Download MediaInfo (available for Windows, macOS, and Linux), open your MP3 file, and read the audio stream details. You're looking for:
- Format: MPEG Audio (Layer 3)
- Bit rate mode: Constant
- Bit rate: 192 kbps
- Sampling rate: 44,100 Hz
- Channel(s): 1 (mono)
If any of these don't match, your file will be rejected regardless of how good the audio sounds.[1] The ACX requirements checklist has the complete format specification. VBR (variable bit rate) MP3s are rejected regardless of average bitrate. ACX reads the bit rate mode from the file header.
How do you measure silence padding?
No free tool automatically verifies ACX's silence padding requirements (0.5–1 second head, 1–5 seconds tail).[1] You need to check this manually.
Measuring silence duration in Audacity
Open your MP3 file and zoom into the beginning. Identify where the room tone ends and narration begins. Select the silence region and read the duration from Audacity's selection toolbar at the bottom of the screen. Do the same for the tail of the file.
Measuring silence in Adobe Audition
Switch to the Waveform view and use the selection timer at the bottom to measure the duration of the silent portion at head and tail. The visual waveform display in Audition makes it straightforward to identify the boundary between silence and speech.
Room tone vs digital silence
ACX's silence should be room tone (the natural ambient sound of your recording environment), not digital silence (absolute zero samples). Room tone sounds natural at transitions between chapters. Digital silence creates audible clicks and channel activation artefacts when playback devices switch from zero signal to audio. The silence padding guide covers this requirement in detail. A gentle fade-in over 10 to 20 milliseconds at the start of speech prevents hard transients.
What does the complete free verification workflow look like?
Here's a per-chapter workflow combining all the tools above:
Per-chapter checklist
- Open in MediaInfo → verify MP3, 192 kbps CBR, 44,100 Hz, mono
- Open in Audacity → run ACX Check → verify RMS and noise floor pass
- Run through Youlean Loudness Meter → verify true peak below -3 dBFS[2]
- Check head silence in Audacity → 0.5 to 1 second of room tone
- Check tail silence in Audacity → 1 to 5 seconds of room tone
- Spot-listen → 30 seconds at beginning, middle, and end for artefacts
For a 30-chapter audiobook, this workflow takes approximately 2–4 hours of careful verification work. You cannot skip chapters. ACX checks every file individually, and a single failing chapter means resubmission.
Noise floor measurement tip
For the noise floor check, select at least one second of clean silence (no speech, no clicks, no breaths). The ACX Check plugin finds the quietest section automatically, but if you're measuring manually, make sure the selected section contains only room tone. Selecting a section with low-level speech or breath sounds produces an artificially high reading that doesn't reflect your actual room noise.
Why doesn't checking individual files catch everything?
Individual chapter verification catches spec failures, but it doesn't catch inter-chapter inconsistency. ACX's human quality review listens for:
- Loudness consistency across chapters: listeners shouldn't need to adjust volume
- Noise character consistency: background noise should sound the same across chapters
- Tonal consistency: narration should sound like it was recorded in the same session
Free tools measure individual files but don't automatically compare files to each other. If you recorded chapters across multiple sessions, days, or rooms, inconsistency is a real risk. The best approach is to process all chapters through the same mastering chain with identical settings. The audiobook mastering guide covers how to set up a consistent processing chain.
When should you automate instead of checking manually?
Free manual verification works. It is accurate when done correctly, and it costs nothing except time. But there are scenarios where automation makes more sense:
Large projects. A 50-chapter audiobook requires 50 repetitions of the manual workflow. At 3–5 minutes per chapter, that's 2.5 to 4 hours of pure verification work, before fixing any issues found.
Frequent submissions. If you produce audiobooks regularly, the verification overhead compounds. Professional narrators producing multiple titles per month cannot afford manual checking.
True peak gaps. Audacity's ACX Check doesn't measure true peak. Adding a separate tool (Youlean or FFmpeg) to the workflow adds complexity and opportunities for error.
Batch consistency. Verifying that all 30 chapters are consistent with each other requires comparing measurements across files, which is tedious to do manually.
ChapterPass handles the full verification and mastering chain automatically. Upload your chapter files and the tool checks every specification, including true peak, format, silence padding, and inter-chapter consistency, in a single pass. Files that don't meet specs are processed to meet them. The complete guide to what ChapterPass does explains the scope.
What mistakes do people make with manual verification?
Even with the right tools, these mistakes cause audiobooks to fail after passing manual checks:
Checking the wrong file
You verify your Audacity project or WAV export, not the final MP3. MP3 encoding changes peak levels by 0.5–1 dB and can shift RMS slightly. Always verify the actual file you'll submit.
Ignoring true peak
Audacity's ACX Check reports sample peak, not true peak. A file at -3.0 dBFS sample peak can have a true peak of -2.5 dBFS or higher. This passes Audacity but fails ACX.
Skipping format verification
You assume your export settings are correct because you set them once. But DAW settings can drift. A project setting change, an update, or a wrong click in the export dialog can produce a file at the wrong bitrate, sample rate, or channel configuration.
Measuring noise floor on the wrong section
The noise floor must be measured on a genuinely silent section: no speech, no breaths, no page turns. Selecting a section with low-level speech or breath sounds produces an artificially high noise floor reading that doesn't reflect your actual room noise.
Missing inter-chapter inconsistency
You check each chapter individually but don't compare them to each other. Chapters recorded on different days or in different rooms may have different loudness, different noise characters, or different tonal qualities, all things ACX's human reviewers will flag.
What's the best verification strategy?
For your first audiobook: use the full manual workflow with Audacity, MediaInfo, and Youlean.[2] It teaches you what each specification means and how to interpret the measurements. That knowledge is valuable even if you later switch to automated tools.
For ongoing production: evaluate whether the time spent on manual verification justifies the cost savings over an automated solution. The math depends on your volume, your hourly rate, and how much the resubmission risk costs you in delayed revenue.
Regardless of method, verify every chapter, verify the final MP3, and verify before you upload. A systematic verification process, whether manual or automated, is the difference between first-submission acceptance and the rejection-resubmission cycle. If you want the full list of what ACX checks, including both the automated technical review and the human quality review, start there.
Check every spec automatically. ChapterPass verifies all eight ACX specifications, including true peak, in a single pass. No plugins, no manual checks, no guesswork.