Audiobook Mastering Glossary
This glossary defines the core technical terms narrators encounter when submitting audiobooks to ACX, Findaway Voices, Authors Republic, and Google Play Books. These definitions are written for people who make audiobooks, not audio engineers. Each term is explained in plain language with the specific numbers that distribution platforms require.
Audiobook Mastering
Audiobook mastering is the final technical processing step that adjusts a recorded audio file to meet the requirements of distribution platforms like ACX, Findaway Voices, Authors Republic, and Google Play Books. For ACX and Audible, mastering means setting RMS loudness to between -23 and -18 dBFS, limiting true peak to below -3 dBFS, ensuring the noise floor is below -60 dBFS, converting to mono MP3 at 44.1 kHz with 192 kbps constant bit rate, and adding the correct silence padding at the head and tail of each chapter file.[1]
Mastering is not the same as editing. Editing fixes recording problems: clicks, pops, mouth noise, long pauses, and mispronunciations. Mastering adjusts the technical measurements of an already-edited file so it passes automated quality checks. ChapterPass handles all measurable ACX requirements automatically, but the recording quality and editing must be done before uploading. Platforms like Findaway Voices share the same core specs, so one mastering pass works across multiple distributors.
ACX Compliance
ACX compliance means an audiobook chapter file passes all 8 of ACX's automated technical quality checks. These checks are: RMS loudness between -23 and -18 dBFS, true peak below -3 dBFS, noise floor below -60 dBFS, sample rate of 44100 Hz, mono channel configuration, MP3 constant bit rate format at 192 kbps or higher, head silence of 0.5 to 1 second, and tail silence of 1 to 5 seconds.[1]
ACX's automated system checks each file independently. One failing chapter can hold up an entire audiobook submission. After the automated checks pass, ACX also performs a human quality review that evaluates recording clarity, consistency, and accuracy against the manuscript. ChapterPass handles all 8 automated technical checks, but cannot address the human quality review. See the full breakdown of what ACX checks and what ChapterPass handles.
RMS Loudness
RMS loudness (Root Mean Square) measures the average power of an audio signal over time, expressed in dBFS (decibels relative to full scale). RMS represents how loud a recording sounds overall, not just the peaks, but the sustained volume a listener perceives. ACX requires each chapter file to have an RMS loudness between -23 and -18 dBFS.[1]
Files quieter than -23 dBFS sound noticeably softer than other audiobooks on Audible, forcing listeners to turn up their volume. Files louder than -18 dBFS trigger an automatic rejection from ACX's quality check system. The range is intentionally narrow to ensure consistent loudness across all titles on the Audible platform. ChapterPass measures each chapter's RMS loudness and adjusts gain to land within this range automatically.
True Peak
True peak measures the highest amplitude of an audio signal, including peaks that occur between digital sample points, called inter-sample peaks. ACX requires true peak to stay below -3 dBFS.[1] This is stricter than standard peak measurement, which only looks at the recorded sample values and can miss peaks that happen during digital-to-analog conversion.
A file that looks fine by standard peak measurement can still fail ACX's true peak check and get rejected. Inter-sample peaks cause audible distortion (crackling and clipping), especially on Bluetooth speakers and earbuds. ChapterPass detects and limits both sample peaks and inter-sample peaks to ensure every output file stays below -3 dBFS true peak without audible artefacts.
Noise Floor
The noise floor is the level of background noise in an audio file during silent sections, the pauses between sentences and paragraphs when no one is speaking. Noise floor is measured in dBFS, and ACX requires it to be below -60 dBFS.[1] A higher noise floor (a number closer to 0 dBFS, such as -50 dBFS) means audible hiss, hum, or room noise that ACX's automated checks will flag as a rejection.
Lower numbers mean quieter backgrounds. A -70 dBFS noise floor is quieter than -60 dBFS. Room treatment, microphone choice, and recording environment all affect the noise floor before mastering begins. ChapterPass manages noise in silent sections, producing a natural-sounding floor that meets the ACX specification without the metallic artefacts that aggressive noise reduction plugins can introduce.
dBFS
dBFS (decibels relative to full scale) is the standard unit for measuring audio levels in digital systems. 0 dBFS is the absolute maximum a digital signal can reach before clipping. All values are negative: -20 dBFS is quieter than -10 dBFS, and -60 dBFS is nearly silent.
Every ACX specification is expressed in dBFS. RMS loudness must fall between -23 and -18 dBFS, true peak must stay below -3 dBFS, and noise floor must remain below -60 dBFS.[1] Understanding dBFS is essential for reading any audio measurement tool, from Audacity to ChapterPass compliance reports.
Sample Rate
Sample rate is how many times per second a digital audio system captures the sound wave, measured in hertz (Hz). Higher sample rates capture more detail. ACX requires 44100 Hz (44.1 kHz),[1] the same standard used for CDs.
Many audio interfaces and DAWs default to 48000 Hz because that is the standard for video production. Recording at 48 kHz is fine, but the file must be resampled to 44100 Hz before ACX submission. Submitting at the wrong sample rate fails the automated check. ChapterPass converts any sample rate to ACX's required 44100 Hz.
MP3 CBR
MP3 CBR (constant bit rate) is an encoding method that uses the same data rate throughout the entire file. ACX requires MP3 at 192 kbps CBR or higher.[1] The constant rate means every second of audio uses the same amount of data.
Variable bit rate (VBR) encoding adjusts quality per segment and produces smaller files, but ACX does not accept it. Audible's streaming platform needs predictable file sizes and consistent streaming behavior, which CBR provides. ChapterPass always encodes output as MP3 192 kbps CBR at 44100 Hz.
Mono vs Stereo
Mono audio uses a single channel. Stereo uses two channels (left and right). ACX requires mono[1] because spoken word does not benefit from stereo separation, and mono files are half the size of stereo files with identical content.
Many DAWs export stereo by default, even when you recorded with a single microphone. Submitting a stereo file to ACX will get it rejected. Findaway Voices and Authors Republic accept both mono and stereo,[2] making ACX the strictest platform for channel configuration. ChapterPass converts all input to mono automatically.
Dynamic Range
Dynamic range is the difference in volume between the loudest and quietest parts of an audio recording, measured in decibels. A recording with wide dynamic range has very loud peaks and very quiet passages. A recording with narrow dynamic range sounds more consistent.
Audiobooks need controlled dynamic range so listeners do not constantly adjust their volume. To keep average loudness (RMS) within ACX's -23 to -18 dBFS window[1] while staying below the -3 dBFS peak ceiling, mastering tools like compression and limiting bring the loudest and quietest parts closer together. This gives the final file a tighter, more consistent sound.
Headroom
Headroom is the space between the loudest peak in an audio file and 0 dBFS (the digital maximum). More headroom means the signal stays further from clipping. ACX requires at least 3 dB of headroom by setting the true peak ceiling at -3 dBFS.[1]
This buffer exists because MP3 encoding can raise peak levels by 0.5 to 1 dB. A WAV file at -3.0 dBFS true peak might become -2.5 dBFS after encoding, which would exceed the limit if the ceiling were set at 0 dBFS. ChapterPass sets the limiter below -3 dBFS to guarantee the final MP3 output stays within spec.
Limiter
A limiter is an audio processor that prevents the signal from exceeding a set ceiling. In audiobook mastering, a brickwall limiter catches peaks that would otherwise exceed ACX's -3 dBFS true peak requirement.[1] The term “brickwall” means nothing passes above the ceiling, no matter how loud the input.
For ACX compliance, the limiter must be true-peak-aware, meaning it detects inter-sample peaks between digital sample points. Standard limiters that only check sample values can miss these peaks, causing files to fail ACX's automated check. ChapterPass uses a true-peak-aware limiter that catches both types. See how to fix true peak rejection issues.
Noise Gate
A noise gate silences audio that falls below a set volume threshold. When the signal drops below the threshold (during pauses between sentences), the gate closes and mutes the output. When speech resumes above the threshold, the gate opens and lets audio through.
In audiobook mastering, noise gates help keep the noise floor below ACX's -60 dBFS requirement[1] by muting background noise (HVAC, fans, hiss) during silent sections while leaving speech untouched. The gate must be set carefully to avoid cutting off quiet syllables or creating unnatural silence. ChapterPass manages noise in silent sections automatically. See how noise floor management works in the mastering process.
Ready to stop worrying about technical specs?
ChapterPass handles all of these automatically.