In a Nutshell
The audiobook recording process turns a finished manuscript into retail-ready audio in nine steps: choose a narrator, prepare the script, build a quiet room, pick gear, record a test chapter, record the book, edit to ACX specs, proof the full file, and upload for distribution. U.S. audiobook sales hit $2.43 billion in 2025, up 9% from 2024, according to the Audio Publishers Association. That growth shows readers expect an audio edition, and authors who plan production properly can get there without wasting weeks on failed files.
A finished manuscript is not an audiobook. The file that passes ACX review is a separate product with separate rules and separate failure points.
DIY recording is possible. Clearing ACX technical review without a rejected file is the harder part. The gear, the room, and the learning curve each cost real time. Authors who want to remove the technical risk can hand the files to an audiobook production team that records, edits, and masters to ACX specs before submission. Understanding the full scope of audiobook production for authors helps you decide whether to record yourself or outsource before you commit weeks to a workflow that may not pass review.
The honest range for an 80,000-word book runs from a few hundred dollars for a basic setup to $2,400 or more for a hired narrator at $100 to $300 per finished hour. ACX itself charges no upload fee. The royalty split covers its share.
A quiet room beats an expensive mic. A test chapter beats good intentions. Skip either one and the rejection email arrives anyway. Your book already exists. The audio edition is a second product, and it follows a different process.
How The Collingwood Press Tested This Audiobook Recording Process
The Collingwood Press built this guide from a working home studio, not from a spec sheet. Our setup includes a treated closet, a large-diaphragm condenser on a boom arm, and a laptop with a solid-state drive. We logged raw recording hours across twelve nonfiction chapters and three fiction samples over eight weeks in 2025 and 2026.
Every technical claim in the ACX section traces to the public ACX audio submission requirements page, checked against the April 2026 version. The room measurements came from a handheld sound level meter and Audacity’s ACX Check plugin. One limitation: we couldn’t test ACX’s human quality reviewers by ear. The numbers show whether a file passes the meter. A human still decides whether the narration sounds right.
Key facts box: the five ACX numbers that matter
| Spec | Target |
| RMS loudness | -23 dB to -18 dB |
| Peak ceiling | -3 dB or lower |
| Noise floor | -60 dB RMS or quieter |
| File format | 192 kbps or higher CBR MP3, 44.1 kHz |
| File length | 120 minutes maximum per file |
Industry timeline guidance from ACX, Penguin Books, and Reedsy points to 6 to 10 weeks for an 80,000-word book when you record 2 to 3 hours of studio time per finished hour.
Step 1: Decide Who Narrates
Start with the audio rights question. You need to control the audio rights before you record a single word, and ACX requires the ebook to be live on Amazon before you can claim the title. If a publisher holds those rights, stop and read the contract.
Three paths sit in front of most authors. You can record it yourself as an author-narrated audiobook. Authors who would rather not manage narration, editing, and ACX compliance on their own can choose audiobook publishing solutions that handle audiobook production alongside print and ebook editions. You can hire a narrator through ACX and pay a per-finished-hour rate or split royalties. You can record the raw audio and hand it to a professional editor to fix technical problems you cannot hear yet.
The market shifted toward audio-first publishing in 2025. According to the APA survey, sales revenue from audio-first publications rose 50% to $136 million. Listeners reward production quality, and a poor narrator can quickly lose them. Industry-wide survey data show 64% of listeners consider high-quality narration essential, and 59% say they will abandon a book because of a bad narrator. The self-audition test is simple: record three minutes, play it next to a published audiobook in your genre, and ask one honest listener which one sounds professional.
Audiobook Cost per Finished Hour: A Transparent Formula
Professional narrators typically need 2 to 3 hours of recording time for every finished hour. Editing adds another 3 or more hours per finished hour for a beginner. At roughly 9,000 words per finished hour, a 45,000-word book lands near 5 finished hours.
A hired narrator at $250 to $400 per finished hour puts that book between $1,250 and $2,000 for narration alone. Your own labor changes the math but not the time cost. The global audiobook market reached an estimated $11.0 billion in 2025 and is projected to hit $13.9 billion in 2026, with smartphone listening driving over 44% of device usage. Production values matter because the audience is large and impatient.
ACX Production Deals in 2026
ACX offers three structures. Pay for Production pays the narrator upfront and keeps your royalties clear. Royalty Share splits earnings but requires exclusivity and locks the title for seven years. Royalty Share Plus combines a smaller upfront payment with a royalty split. Under the new royalty model enrolled in 2026, creators earn 50% for exclusive distribution and 30% for non-exclusive distribution. Verify the current rates on ACX the day you publish, because the platform updates terms.
Step 2: Prepare Your Audiobook Script
A manuscript is not a script. Convert it before you open the microphone. Build a pronunciation sheet for character names, place names, brand names, and any technical term you will say more than once. Decide how to handle URLs: say the domain and skip the path unless the listener truly needs it.
Footnotes and tables break audio flow. Rewrite them as spoken sentences or cut them. Replace “see Chapter 3” with a short reminder of the point, because listeners cannot flip pages. Keep the text and the audio identical if you plan to use Whispersync, since sync breaks when the narration skips or adds words.
Mark the opening and closing credits exactly as ACX requires. Use a chapter-file naming scheme that sorts cleanly, such as 01_Chapter_Title.mp3. A downloadable audiobook recording process checklist helps beginners avoid missing these details on day one.
AI narration exists for backlist and academic content, and voice synthesis has cut production costs by up to 80%. The sales data tells a different story. AI-generated audio accounted for just 0.03% of total audiobook sales revenue in 2025, and listener willingness to try AI-narrated books dropped from 70% to 61%. Human narration prep still sets the standard for a book you want readers to finish.
Step 3: Build a Home Audiobook Recording Studio
The room beats the microphone. A $900 condenser in a live, echoey bedroom will fail the noise floor and sound worse than a $120 dynamic mic in a closet full of coats. Find the noises you stopped hearing: the refrigerator hum, the HVAC cycle, the pipe knock, the laptop fan, the phone buzz on the desk.
The Collingwood Press measured the noise floor in an untreated spare room at roughly -42 dB RMS. After moving into a closet with moving blankets on three walls and a rug on the floor, the same gear and voice landed near -62 dB RMS. The treatment didn’t fix the low-frequency rumble from the building’s HVAC. A high-pass filter in editing handled that after recording, but a better room would have removed it at the source.
A home audiobook recording studio does not need foam pyramids on every surface. Thick moving blankets, a carpet, and a corner away from windows do most of the work. If your room cannot get quiet enough, rent studio time by the hour. Audiobook recording process setup costs vary by city, and renting usually beats rebuilding a room you cannot fix.
Step 4: Choose Audiobook Recording Equipment and Software
Microphone: What Actually Works for Narration
The best microphone for audiobook narration matches your room and voice, not the highest price. A USB condenser is the simplest path for a beginner: one cable, no interface, decent results in a treated space. An XLR condenser with an audio interface gives cleaner gain and more headroom. A dynamic microphone with an inline preamp rejects more room noise and suits bright or thin voices that need weight.
Pop Filter, Headphones, Stand, and Interface
A pop filter sits between your mouth and the capsule, catching plosives before they hit the recording. Closed-back headphones let you hear mouth clicks and room rumble that speakers hide. A boom arm keeps the mic off the desk and reduces thumps. For spoken word, a clean preamp interface matters more than a high sample rate.
Software: Pick One and Learn Its Meter
Audacity is free and runs the ACX Check plugin, which measures RMS, peak, and noise floor against the submission numbers. Reaper costs less than most plugins and handles large sessions well. Adobe Audition fits editors who already work in the Creative Cloud. GarageBand works for a test chapter, but its stock tools make precise ACX metering harder than Audacity does. Pick the software that shows you the numbers you need to pass.
Gain staging sets the foundation. Set your input level so loud passages peak around -12 dB to -9 dB. That leaves headroom for the performance and keeps the final peak under the -3 dB ceiling without crushing the dynamics in the edit.
Plan for 2 to 3 hours of studio recording time for every 1 hour of finished audio. That ratio comes from ACX, Penguin Books, and Reedsy guidance, and our own sessions tracked close to it.
Step 5: Record a Test Chapter Before You Commit
Record one full chapter, process it through your chain, and run the ACX Check. Send the file to a listener who will tell you the truth about mouth noise, pacing, and whether your voice sounds tired by the end. Make the pass-or-fix decision here, at chapter one, not at chapter thirty.
The test also validates your timeline. If one chapter takes four hours from setup to exported file, an 80,000-word book with twelve chapters lands near 48 hours of work spread across several weeks. The industry benchmark of 6 to 10 weeks for an 80,000-word book assumes a steady schedule and a room that passes on the first try.
Step 6: Record Each Chapter
Session length matters more than willpower. Three to four hours of recording per day keeps your voice consistent and your attention sharp. Warm up your voice before the first take. Keep water at room temperature, because cold water tightens the throat. Eat something mild before a session so stomach noise stays off the mic.
Leave three seconds of silence between takes. Clap once or say a marker word before a mistake so the edit finds it fast. Record room tone for five seconds at the head and tail of every chapter. Stop when your voice changes, because a tired take sounds different from the take you recorded two hours earlier.
These audiobook narration tips save more time in editing than they cost in recording. Consistent pacing across sessions keeps the book from sounding stitched together. One chapter per file keeps the edit manageable. A session log with date, chapters, raw minutes, and issues catches problems before they compound.
Step 7: Audiobook Editing and Mastering to Meet ACX Audio Requirements
| Requirement | ACX spec | Why it matters |
| RMS loudness | -23 dB to -18 dB | Keeps the book at a consistent, listenable volume |
| Peak | -3 dB or lower | Prevents distortion on playback |
| Noise floor | -60 dB RMS or quieter | Hides room hum and gear hiss under the voice |
| Format | 192 kbps+ CBR MP3, 44.1 kHz | Meets retail encoding standard |
| Channels | All mono or all stereo | Prevents channel mismatch |
| File length | 120 minutes max | Keeps files manageable for QA |
| Room tone | 1 to 5 seconds at start and end | Gives the editor clean edges |
| Narration | Human | ACX requires human narration |
The processing chain runs in order. Edit first: cut mistakes, breaths you do not want, and long gaps. Apply gentle noise reduction only if the floor fails. De-ess and control plosives next. Add a light EQ to remove rumble and tame harshness. Apply light compression, then a limiter, then set loudness to hit the RMS target. Export a WAV master before encoding the MP3.
Boosting the level raises the noise floor with the voice. If your raw file sits at -28 dB RMS with a -54 dB floor, normalizing to -20 dB pushes the floor to roughly -46 dB, which still passes. If the floor starts at -58 dB, the same move fails. Fix the room or the gain staging before you fix the level.
Export a WAV master, then encode a constant bitrate MP3 at 192 kbps or higher and 44.1 kHz. ACX recommends the fre:ac encoder for this step. When the time cost of learning mastering outweighs the money cost of hiring an engineer, hire the engineer. The Collingwood Press keeps audiobook editing and mastering quotes itemized so authors see the line items instead of a single vague number.
Step 8: Proof Your Audiobook
A full proof listen catches what the meters cannot. Follow the text while you listen and mark every mispronunciation, repeated sentence, level jump, edit click, and tone shift. Run pickups in the same room, with the same gain and distance from the mic, or the fix will sound pasted in.
Joanna Penn has described a virtual assistant catching repeated lines she missed in her own proof listen. A fresh set of ears works because the author’s brain fills in the words it expects to hear. If you cannot hire a proof listener, wait a week and listen to the chapter backward, sentence by sentence, against the manuscript.
Step 9: Upload to ACX and Choose Distribution
ACX Audiobook Upload
Log in to ACX, confirm the title is claimed, and upload the files with the opening and closing credits attached. Add the sample, the cover art that meets ACX’s image requirements, and the metadata. If ACX rejects the file, the rejection notice usually points to peak, noise floor, or file length. Fix the exact spec, re-export, and resubmit.
Exclusive vs Non-Exclusive in 2026
Exclusive distribution pays 50% under the new model and keeps the title inside Audible’s ecosystem. Non-exclusive pays 30% and lets you distribute elsewhere. Royalty Share still requires exclusivity and a seven-year term. A book with strong series potential and a listener base in Audible fits exclusive.
A book that sells through multiple channels, or one tied to a course or speaking business, fits non-exclusive. Distribution choice also affects author branding: exclusive titles live only in Audible’s catalog, while wide distribution keeps the author’s name visible across library, Spotify, and retail platforms.
Audiobook Distribution for Authors Beyond ACX
ACX is not the only path. Voices by INaudio, formerly Findaway Voices, reaches libraries and retail channels outside Audible. Spotify for Authors accepts audiobook uploads for direct distribution. Author’s Republic handles wide distribution for authors who want a single dashboard. Once the files go live, marketing audiobook titles requires its own plan, including review requests, sample clips for social media, and cross-promotion with the ebook, because audio discovery works differently from print launches. A broader book marketing guide helps authors connect audiobook promotion to the rest of their launch plan instead of treating audio as a separate campaign.
Findaway Voices vs ACX: A 2026 Comparison
| Factor | ACX | Voices by INaudio (formerly Findaway) |
| Primary reach | Audible, Amazon, iTunes | Libraries, Spotify, wider retail |
| Royalty structure | 50% exclusive, 30% non-exclusive | Varies by channel |
| Exclusivity | Required for royalty share | Non-exclusive by default |
| Fees | No upload fee | Distribution fees may apply |
| Best for | Audible-first launches | Library and wide distribution |
Findaway Voices no longer operates under that name. Authors searching for Findaway Voices vs ACX in 2026 need the current brand: Voices by INaudio.
Audiobook Recording Troubleshooting
| Symptom | Likely cause | Fix |
| Constant hiss | Gain too high or noisy preamp | Lower gain, move mic closer, apply gentle noise reduction. |
| Low hum | HVAC, fridge, or ground loop | Change rooms, use a high-pass filter |
| Echo | Bare walls and hard floor | Add soft surfaces, move into a closet |
| Plosives | Mic directly in breath path | Add a pop filter, angle the mic slightly |
| Mouth clicks | Dry mouth or close mic | Hydrate, record with more distance |
| Clipping | Input level too hot | Lower gain, set peaks near -12 dB |
| Low RMS after export | Weak performance or quiet gain | Raise level in mastering, not by shouting |
| Noise floor fails after normalizing | Floor was too high before gain | Fix the room first |
| ACX rejection | File length, bitrate, or credits | Check the rejection note against the spec table |
Should You Use AI Narration Instead?
ACX requires human narration for titles uploaded through its system. The sales numbers reinforce why that rule exists: AI-narrated audiobooks pulled in 0.03% of total sales revenue in 2025, and the share of listeners willing to try one fell from 70% to 61%. AI narration fits low-cost backlist experiments or academic content where a human read is not practical. For a book you want listeners to finish and review, human narration remains the standard.
Conclusion
Preparation beats talent in the audiobook recording process. A quiet room passes the noise floor. A test chapter catches problems early. A metered edit clears ACX QA. Skip any one of those three and the file fails, no matter how good your voice sounds.
The Collingwood Press handles audiobook production end to end, from script prep to ACX upload, with itemized quotes instead of vague hourly rates. A finished audiobook also becomes an author promotion asset, including sample clips, interview material, and a second format that reaches listeners who never pick up a print copy. Authors who want the technical work off their plate can see the full service list before committing a dollar.
A finished manuscript is not a finished audiobook. The microphone only rewards authors who prepare the room, the script, and the file before they press record.
Frequently Asked Questions
How long does the audiobook recording process take?
An 80,000-word book typically takes 6 to 10 weeks from first session to uploaded files, based on ACX, Penguin, and Reedsy timeline guidance. Plan 2 to 3 hours of studio time for every finished hour.
How many hours does it take to record a 50,000-word book?
At roughly 9,000 words per finished hour, that book runs about 5.5 finished hours. Recording alone takes 11 to 17 hours, and editing adds another 16+ hours for a beginner.
What are the ACX requirements?
Files must measure -23 to -18 dB RMS, peak no higher than -3 dB, and hold a noise floor at -60 dB or quieter. Use 192 kbps or higher CBR MP3 at 44.1 kHz, keep files under 120 minutes, and include 1 to 5 seconds of room tone.
Can I record an audiobook at home?
Yes, if the room meets the noise floor requirements. A treated closet with moving blankets and a decent microphone can meet ACX specs. Test one chapter and run the ACX Check before recording the whole book.
How much does it cost?
DIY costs range from a few hundred dollars for a USB mic and free software to roughly $1,000 for an XLR setup and treatment. A hired narrator at $250 to $400 per finished hour puts a 5-hour book between $1,250 and $2,000.
Is Findaway Voices still available?
Findaway Voices rebranded as Voices by INaudio. Authors searching for the old name should look for the current brand and compare its library reach against ACX before choosing distribution.
