Submitting high-quality audio is essential when applying for audiobook narration on ACX. Poor audio quality can lead to rejection or delays in approval. This article provides practical tips to identify and fix common audio artifacts before submitting your work, helping you achieve a polished recording that meets ACX technical standards.

Understanding Common Audio Artifacts

Audio artifacts are unwanted sounds or distortions that can occur during recording or editing. Recognizing these issues early helps ensure your audiobook meets ACX standards. Failing to address them can result in rejection notices, wasted time, and additional recording sessions. Below is a detailed breakdown of the most frequent artifacts and what causes them.

Clicks and Pops

Clicks and pops are sudden, short-duration noises. They often arise from digital glitches, static electricity hitting the microphone diaphragm, or handling noise when you bump the mic stand. Mouth sounds—such as lip smacks, tongue clicks, or dry mouth noises—also fall into this category. These artifacts appear as sharp vertical spikes in your waveform and can be distracting even at low volume.

Hiss and Background Noise

Background hiss is a constant ambient noise generated by your microphone's self-noise, your room’s HVAC system, computer fans, or electrical hum. Even if low in level, persistent hiss adds a layer of interference that fatigues listeners. ACX requires noise floor to be at -60dB or lower, so controlling this is critical.

Distortion and Clipping

Distortion occurs when the input signal exceeds the maximum level your recording device can handle, resulting in a harsh, buzzy, or muffled sound. Clipping is the visible flattening of waveform peaks. Even slight distortion can ruin a narration; severe cases often require re-recording.

Dropouts and Gaps

Dropouts are brief silences or missing segments of audio. They can stem from buffer underruns in your audio interface, dropped USB connections, or simply forgetting to press record. Inconsistent gaps between sentences or within words also count as artifacts if they break the natural flow.

Sibilance and Plosives

Sibilance refers to harsh “s”, “sh”, “ch”, and “z” sounds that cause a piercing or lisping effect. Plosives occur when air from consonants like “p”, “b”, “t”, “k” hits the microphone diaphragm, creating a low-frequency thump. Both are common and manageable with proper mic technique and post-processing.

Room Echo and Reverb

If your recording space has hard surfaces (bare walls, windows, tile floors), your voice will bounce and create a hollow, swimming-pool-like sound called room echo or flutter echo. ACX requires a clean, close, “telephone-like” sound free of noticeable room acoustics.

Microphone Handling Noise and Rumble

Vibrations from desk bumps, footsteps, or mic cable movement transmit low-frequency rumble. These subsonic noises may not be audible on small speakers but become obvious on headphones or through a subwoofer. They also waste headroom and can cause clipping in the low end.

How to Identify Artifacts

Systematic detection is the foundation of effective cleanup. Relying on your ears alone can miss subtle problems. Combine listening with visual analysis for a thorough check.

Use a Controlled Listening Environment

Listen in a quiet room using closed-back studio headphones that isolate you from outside noise. Open-back headphones may bleed sound and mask low-level artifacts. Check your recording at a moderate volume—listening too loud masks hiss and low-level clicks, while too quiet hides sibilance and mouth noises.

Visual Inspection with Waveforms and Spectrograms

Zoom in on your waveform to see sharp spikes (clicks/pops) or flattened tops (clipping). Use a spectrogram view (available in Audacity, Adobe Audition, Reaper, etc.) to identify tonal issues. Hiss shows as a uniform haze across all frequencies, electrical hum appears as vertical lines at 60Hz (or 50Hz depending on region), and room echo creates repeating patterns in the midrange. Sibilance glows brightly around 6–10 kHz.

Listen in Multiple Playback Systems

What sounds clean on studio headphones may reveal problems on laptop speakers or earbuds. Test your submission on at least two different playback devices: one high-fidelity (headphones) and one consumer-grade (smartphone speaker). This catches artifacts that are masked by your monitoring chain.

Check ACX Requirements First

ACX publishes detailed audio specifications. Your submitted file must be: sample rate 44.1 kHz, bit depth 16-bit, mono channel, constant bitrate 256 kbps, RMS levels between -23dB and -18dB (with peak not exceeding -3dB), and noise floor below -60dB. Any deviation can trigger automatic rejection. Familiarize yourself with these specs on the ACX help page before you start editing.

Steps to Fix Common Audio Artifacts

Each artifact requires a targeted approach. Apply fixes in a logical order: start with noise reduction and de-essing, then handle clicks/pops, then adjust dynamics. This sequence prevents one tool from re-creating problems another tool just fixed.

Removing Clicks and Pops

Begin with manual editing for the most precise results. Zoom into the waveform until you see the spike clearly. Select a tiny region around the spike (just a few samples) and delete it, then crossfade (3–5ms) to blend the gap. For repetitive mouth clicks, use a de-click plugin: Audacity’s built-in Click Removal (Effects → Click Removal) works well, but reduce its strength to avoid creating digital artifacts. In Adobe Audition, the Declicker (Effects → Diagnostics → DeClicker) offers presets for mild, moderate, or strong removal. Apply de-click only to problem areas, not the whole file, to preserve natural transients.

If you have many clicks throughout the recording, adjust your recording habits. Drink room-temperature water between takes, avoid dairy or sugary drinks that thicken saliva, and position the microphone slightly off-axis to reduce breath noise.

Reducing Background Noise and Hiss

Noise reduction is a two-step process. First, capture a noise profile: select a few seconds of pure silence (between breaths or at the end of a sentence) and apply the “Noise Print” or “Capture Noise Reduction Profile” tool. Then, select the entire audio and apply the noise reduction effect with settings around 20–30dB reduction and sensitivity to 6–12dB. Over-reduction creates a metallic “watery” sound; if you hear that, lower the reduction amount or try a multi-band approach using spectral editing.

For more precise control, use a noise gate. Set the gate threshold just above the noise floor (around -55dB) so it closes during silent passages. Be careful: too aggressive a gate can cut off room tone and cause unnatural gaps. A gentle gate combined with light noise reduction yields the best results.

If constant hiss persists, consider a broadband de-hisser plugin (like iZotope RX De-hum) or EQ: roll off frequencies below 80Hz (rumble filter) and above 12–15kHz (air noise), but do not cut too much or you’ll dull the voice.

Eliminating Electrical Hum

Electrical hum (50/60Hz and harmonics) can be removed with a notch filter or spectral editing. In Audacity, use Effects → Notch Filter, set frequency to 60Hz (or 50Hz) and Q to 30. This removes the fundamental but may not affect harmonics at 120Hz, 180Hz, etc. Repeat for each harmonic. In Audition, use the De-hum tool (Effects → Noise Reduction → DeHummer) and select the power line frequency. For stubborn hum, a hand-drawn spectrogram selection and silencing works best.

Fixing Distortion and Clipping

Clipped audio cannot be truly restored—the waveform peaks are irreversibly flattened. However, you can reduce the harshness. Use a declipper tool: Audacity has an optional plug-in (Clap & Clip Removal) but Audition’s DeClipper (Effects → Diagnostics → DeClipper) is more capable. Apply it with moderate strength; if the distortion is severe, the best fix is re-recording the affected sections.

To prevent future clipping, monitor your recording levels. Keep peaks around -6dB to -3dB to leave headroom. Never exceed 0dB. Use a compressor during recording (or as a post-effect) to control dynamic range gently—set a ratio of 2:1 or 3:1 with a threshold that catches only loud peaks.

Addressing Dropouts and Gaps

If you notice a dropout (complete silence), first determine if it’s a recording error or an intentional pause that went too long. For an error, re-record the missing sentence or phrase. Align the new take to the surrounding audio using crossfades (10–20ms). If the dropout is only a few milliseconds (a tiny gap), you can fill it with a small amount of room tone captured from a silent moment. Paste that tone into the gap and crossfade to disguise the cut.

For gaps between words that are too long or too short, use time-stretching or simply rebuild the sentence by re-recording. ACX expects natural pacing, so avoid editing large sections with aggressive time compression—it introduces warbling artifacts.

Mitigating Sibilance and Plosives

Start on the recording side: use a pop filter (at least 4 inches from the mic), position the mic slightly above or to the side of your mouth (not directly in front), and practice controlled articulation. For existing recordings, use a de-esser plugin. In Audacity, you can use a multiband compressor (set the sidechain to 5–8kHz, reduce by 3–6dB). In Audition, the DeEsser effect (Effects → Filter and EQ → DeEsser) offers two presets. Apply sparingly; over-de-essing creates a lisp.

Plosives can be reduced with a low-cut filter (high-pass filter) set to 80–100Hz. Manually select the plosive burst (a few milliseconds) and apply a fade-in or reduce the gain by 6–12dB. If the plosive is within the word (e.g., “pop”), you may need to re-record that word.

Cleaning Up Room Echo and Reverb

Heavy reverb is extremely difficult to remove without degrading the voice. Light reverb can be reduced using a de-reverb plugin (Audition’s DeReverb, iZotope RX De-reverb). These tools analyze the mix of direct sound and reflections, then subtract the reverb tail. If your room echoes noticeably, prevention is far better than cure. Record in a small, carpeted closet or use acoustic panels to dampen reflections. Hanging heavy blankets around your recording area helps dramatically.

Handling Rumble and Low-Frequency Noise

Use a high-pass filter at 60–80Hz with a slope of 12dB/octave to remove subsonic rumble, mic handling noise, and HVAC rumble. Do not set the cutoff too high or you’ll thin out your voice, especially for male narrators. A gentle filter preserves natural warmth while cleaning the low end.

Preventing Artifacts During Recording

Microphone Placement and Technique

Maintain a consistent distance of 6–12 inches from the microphone, slightly off-axis to avoid plosives. Use a boom arm or shock mount to isolate from vibrations. Never hold the microphone. Place a pop filter between you and the mic. For mouth noise reduction, keep hydrated and avoid talking on an empty stomach (dry lips).

Room Treatment

Even a budget setup can achieve ACX-quality acoustics. Record in a space with soft furnishings—carpet, curtains, blankets. Build a portable vocal booth using PVC pipes and moving blankets. Avoid corner positions where bass builds up. Aim for a “dead” sound with minimal reflections; you can always add subtle ambience later if desired.

Gain Staging and Recording Levels

Set your preamp gain so that normal speech peaks between -12dB and -6dB (in your DAW). This leaves headroom for loud exclamations. Record in 24-bit or 32-bit float (then convert to 16-bit for submission) to capture more dynamic range. Always monitor with headphones, not speakers, to avoid feedback and to hear room issues.

Final QC Checklist Before Submission

  1. Check RMS levels: Use the ACX RMS check tool or a loudness meter plugin. Target -20dB RMS for audiobooks. Ensure peak never exceeds -3dB.
  2. Verify noise floor: In a silent section, measure the noise level. It must be below -60dB. If higher, re-apply noise reduction or re-record.
  3. Listen end-to-end: Play the entire file without skipping. Note any remaining clicks, breaths that sound too loud, or unnatural edits.
  4. Export in correct format: Mono, 44.1 kHz, 16-bit, 256kbps constant bitrate, MP3 (or WAV if required by the rights holder).
  5. Use ACX’s submission checker: ACX provides a web-based tool to test your file. Submit a sample before the full book to catch issues early.
  • Audacity (free, open source) – solid for noise reduction, click removal, and basic EQ. Download here.
  • Adobe Audition (paid) – professional-level spectral editing, declicking, and de-clipping.
  • Reaper (inexpensive) – highly customizable with many free plugin packs.
  • iZotope RX Elements (paid) – industry-standard repair suite, but pricey. Elements version covers most needs.
  • Youlean Loudness Meter (free) – check RMS and LUFS compliance.

By systematically identifying and fixing these common artifacts, you can produce audiobook narration that meets ACX’s stringent standards. A clean recording not only passes quality checks faster but also pleases listeners and helps you build a reputation as a reliable narrator. Take the time to master these techniques—your voice-over career will benefit from every artifact you remove.