Understanding ACX Technical Requirements

Before you can achieve optimal vocal balance, you must first know the standards set by ACX (Audiobook Creation Exchange). These requirements are not arbitrary — they are designed to ensure consistent playback quality across all devices and platforms. The core technical specifications for an ACX-compliant audio file include a sample rate of 44.1 kHz or 48 kHz, a bit depth of 16-bit or 24-bit, and peak levels that do not exceed -3 dB. Additionally, the recording must maintain consistent volume levels throughout, with an average loudness (RMS) typically falling between -23 dB and -18 dB. Your final file must be mono (unless specified otherwise), and the background noise floor should be below -60 dB. Any deviation from these specs can result in rejection, so check each recording against these numbers before submission.

Beyond the numbers, ACX also expects a clear, uncolored vocal tone with no audible distortion, clicks, or plosives. Achieving this balance requires a combination of proper recording technique, thoughtful processing, and careful monitoring. The following sections break down every aspect of vocal balance in the context of ACX production.

Core Elements of Vocal Balance for Audiobook Narration

Vocal balance involves managing volume dynamics, tonal consistency, and clarity so that every syllable is audible and comfortable to listen to over long periods. Audiobook listeners often play files at lower volumes or through earbuds, so maintaining a stable perceived loudness is critical. The three pillars of vocal balance are level control, tonal control, and dynamic control.

Level Control: Gain Staging and Microphone Technique

Level control starts at the microphone. Set your preamp gain so that your loudest peaks hit around -6 dB on the meter — this gives you headroom and avoids clipping while keeping the average level strong. Avoid speaking too close to the mic (within 2 inches can cause proximity effect boost and plosives) or too far (over 12 inches yields thin, roomy sound). Aim for a consistent distance of 6 to 8 inches, slightly off-axis to reduce sibilance. Use a pop filter to catch burst consonants. During long sessions, your vocal energy may dip; check your meters periodically and adjust your distance or gain accordingly. A simple test: record a short passage and look at the waveform. If there are huge peaks and valleys, you need to either alter your performance or apply compression later.

Many narrators use a compressor during tracking to catch wild peaks, but for ACX purposes, it’s often better to record without compression and apply it in editing. This way you retain full control over the final dynamics. If you do compress while tracking, use a very gentle ratio (1.5:1 or 2:1) and a slow attack so you don’t kill the natural life of the voice.

Tonal Consistency: Voice Warm-Up and Room Acoustics

A consistent tone and pitch come from good vocal technique and an acoustically neutral environment. Before recording, warm up your voice for at least five minutes with humming, lip trills, and scales. Diaphragmatic breathing (belly breathing) supports a steady delivery and prevents pitch wobble. Record in a quiet, treated room. Even minor room reflections or flutter echo can color the tone and make it harder to balance post-production. Use acoustic panels, blankets, or a portable vocal booth to dampen early reflections. For a low-cost solution, a closet full of clothes or a moving blanket draped over a mic stand works surprisingly well.

Monitor your recordings with closed-back headphones to avoid bleed into the mic. Check for consistent proximity effect — the natural bass boost when you get closer — by listening to the low end across sentences. If your tone shifts as you move, adjust your recording position or use EQ to gently tame the lows when you get close.

Editing Techniques to Perfect Vocal Balance

Once your raw recording is clean and consistent, editing brings it into full ACX compliance. The goal is to even out dynamics, enhance clarity, and reduce background noise without making the voice sound processed. Key tools: compression, equalization, noise reduction, and limiting.

Using Compression for Dynamic Consistency

Compression reduces the gap between loud and soft passages. For ACX narration, a moderate compression setting works best. Start with a ratio of 2:1 or 3:1, a slow attack (10–30 ms) to preserve natural transients, and a medium release (50–100 ms). Set the threshold so that the compressor engages on the louder phrases but leaves softer sections untouched. Aim for around 2–4 dB of gain reduction on peaks. You can also use a second compressor after the first with a lower ratio and faster release to catch any remaining irregularities. Be careful not to over-compress: a flat, lifeless voice fatigues listeners. After compression, makeup gain should bring the average level to around -20 dB RMS. Always A/B test before and after to ensure naturalness.

For those new to compression, Sound on Sound’s compression guide provides a thorough explanation of parameters and common use cases.

Equalization (EQ) for Clarity and Warmth

EQ shapes the tonal balance. Start with a high-pass filter at around 80–100 Hz to remove low-frequency rumble (air conditioning, footsteps, mic handling noise). Then address problem frequencies. A slight boost (1–3 dB) at 3–4 kHz adds presence and clarity — especially important for audiobooks listened to on earbuds. A gentle cut at 200–400 Hz can reduce muddiness. If your voice sounds harsh, dip 1–2 dB around 2–3 kHz or 5–7 kHz. Use narrow Q cuts to target resonances, but avoid cutting too much or the voice will sound thin. Always EQ in context: listen on headphones and speakers, and check the mono compatibility (many podcast players sum to mono).

Remember that EQ can also help reduce sibilance. A de-esser (a specialized dynamic EQ) is more effective than static cuts for sibilance control. Set it to catch frequencies between 5–8 kHz with a threshold that activates only on harsh ‘s’ sounds. iZotope’s EQ guide offers deeper insight into surgical EQ for voiceover.

Noise Reduction and Gating

Background noise must be below -60 dB to pass ACX. If your room is relatively quiet, a noise gate can mute silence between phrases. Set the gate threshold just above the noise floor but below your quietest vocal level. Use a fast attack (1–5 ms) and a medium release (100–200 ms). Avoid too short a release as it can chop off tail ends of words. For persistent low-level noise (hum, fan), use a noise reduction plugin like RX or a spectral editor. Subtract a noise print from a silent portion of your track. Overdoing noise reduction can cause artifacts (warbling, underwater sound), so apply it lightly. If you need to remove clicks or mouth noises, use a dedicated declicker or manually edit them out.

Limiting and Peak Control

After compression and EQ, insert a limiter to catch any stray peaks that exceed -3 dB. Set the ceiling to -3 dB (often written as -3.0 dBFS). Use a brickwall limiter with a short lookahead (1–5 ms) and fast attack. The limiter should only engage on the absolute loudest moments — if it’s activating on every syllable, your compression settings are too weak. A well-compressed recording should require only occasional limiting. Check the final file’s true peak to ensure compliance. Never exceed -3 dB true peak, as ACX will reject the file.

Monitoring and Critical Listening

Your monitoring setup directly affects your ability to judge vocal balance. Use closed-back headphones (e.g., Audio-Technica ATH-M50x or Sony MDR-7506) for recording and mixing. Check your mix on multiple sources: studio monitors, consumer earbuds, and laptop speakers. Each will reveal different aspects of balance. Pay attention to the following:

  • Consistency – Does the volume vary drastically between soft and loud passages? Adjust compression or perform re-records.
  • Clarity – Are consonants smeared? Check high-frequency content and ensure no sibilance spikes.
  • Low-end – Is there a boomy or boxy quality? A high-pass filter at 80 Hz often resolves it.
  • Breath sounds – Excessive breath noise can be distracting. Use a breath gate or manually lower breath levels by 6–10 dB, but don’t entirely remove breaths — natural breathing is part of a good audiobook experience.

Take breaks every 20–30 minutes while listening. Ear fatigue affects judgment. If possible, have a second pair of ears (a friend or a producer) listen to a sample. ACX offers a free “Check Your Audio” tool online; upload a short clip to see if it passes technical specs before submitting the entire book.

Advanced Vocal Balance: Mastering for ACX

Once you’ve processed individual chapters, perform a final master on the entire book to ensure consistent levels across all files. Normalize the RMS of each chapter to the same level (e.g., -20 dB) so that the listener doesn’t have to adjust volume between chapters. Use a loudness meter (many DAWs have one built-in) to match the integrated LUFS to around -20 LUFS, which corresponds to an RMS roughly around -18 to -20 dB for speech. Avoid “loudness wars” — audiobooks should be dynamic but consistent. Export your final files as 16-bit, 44.1 kHz, mono, uncompressed WAV or MP3 at 192 kbps constant bitrate (for distribution).

For a deeper dive into loudness normalization for audiobooks, Adobe’s loudness normalization resource provides industry-standard practices.

Finally, run your files through ACX’s official technical requirements page to confirm you meet every guideline before uploading. Many narrators also use the free app “AudioLab” or “YouLean Loudness Meter” to check LUFS and true peak during the mastering stage.

Common Pitfalls and How to Avoid Them

Even experienced narrators sometimes miss the mark on vocal balance. Here are typical issues and fixes:

  • Inconsistent distance – Use a mic stand with a boom arm and mark your position on the floor. Sit in a fixed chair to prevent shifting.
  • Over-compression – If your recording sounds lifeless and squashed, reduce the ratio or raise the threshold. Aim for natural dynamic flow.
  • Harsh sibilance – Use a de-esser before or after compression. If it persists, try a different microphone that has a gentler high-frequency response (e.g., Shure SM7B, Rode NT1).
  • Room reverb – Add more acoustic treatment or record in a smaller space. A noise gate can help, but it won’t fix reverb on sustained notes.
  • Levels too low – If your average RMS is below -23 dB, you may need to bring up quiet sections with compression or normalizing. But do not just raise overall gain — that lifts noise too.

Record a practice chapter before committing to the whole book. Analyze it with a spectrum analyzer and loudness meter. Adjust your chain until you can consistently produce files that pass ACX’s automated check. Over time, the process becomes second nature.

Final Thoughts: Consistency Is Key

Perfect vocal balance is not a single fixed setting but a discipline that spans performance, acoustics, editing, and mastering. The most reliable way to satisfy ACX’s guidelines is to build a repeatable workflow. Calibrate your gain once, tune your room once, and then use the same processing chain (compressor, EQ, limiter) for every chapter. Only adjust when you change narrating style or microphone. Document your settings — ratios, frequencies, threshold values — so you can quickly recall what worked for previous projects.

Regularly revisit the ACX audio submission requirements because specifications can update. Stay connected with fellow narrators in online communities (e.g., ACX Facebook groups, Audiobook Narrators subreddit) to exchange tips and troubleshooting advice. With meticulous attention to level, tone, and dynamics, you will consistently produce audiobooks that sound professional and pass technical inspection on the first try.