What Does an Audiobook Production Service Actually Do for an Author?

An audiobook production service turns a finished manuscript into retail-ready audio. For a fee, usually charged per finished hour of completed audio, it casts and directs a narrator. It then edits the recording, checks it word-for-word against your book, and delivers files that meet store specifications. You keep the rights to the finished audiobook; getting it into stores is a separate step.

What Does an Audiobook Production Service Actually Handle?

A production company manages every stage between a finished manuscript and a compliant set of audio files. That covers casting, script preparation, performance direction, recording, editing, proof-listening, and mastering. Hiring one is only one route into audio. Our earlier primer on how self-publishing an audiobook works covers the format as a whole, including DIY approaches, while this article stays on the production-service route. The recording itself passes through three distinct technical stages: editing, proof-listening, and mastering. Each one is explained in its own section further down.

You are not passive in that pipeline. Most services build in the same author approval moments, in the same order. You choose the narrator, agree the pronunciation list, sign off a short opening sample, and approve the completed book. ACX formalises all four in its contract, and other studios arrange them by agreement. The worked example below therefore uses ACX’s published terms.

Casting, Script Preparation and the Pronunciation List

Casting comes first, and how it works depends on the kind of service. A marketplace platform such as ACX posts your project to its pool of freelance narrators, who record auditions for it. A boutique studio instead proposes voices from its own roster of narrators. Either way, you review audition recordings of a short passage and pick the narrator. ACX (Audiobook Creation Exchange) revised its standard production terms in May 2026. Under those terms, you must supply the final manuscript within three business days of accepting a narrator’s offer. The text therefore needs to be genuinely finished before casting starts.

Script preparation follows. You supply the pronunciation list: character names, place names, invented terms, and anything else a narrator could reasonably say two different ways. If you have been through copyediting, this list will feel familiar. It does the same job as a copy editor’s style sheet, recording one agreed decision per term so every chapter matches.

The 15-Minute Checkpoint and Final Approval

The first checkpoint is a fully edited and mastered sample of the opening 15 minutes. This is a contractual gate, not a courtesy. Under ACX’s terms you may request up to two rounds of changes at this stage, and you have four business days to respond. It is also the last realistic point to redirect pacing, tone, or character voices before the narrator records the remaining hours.

Full recording and chapter-by-chapter proofing then typically takes two to six weeks. Once the narrator finishes, you review the completed audiobook. The same ACX terms allow up to two further revision rounds here, with ten business days to respond.

It is worth knowing what those revision rounds are for. The terms grant exactly three approval rights (narrator, opening sample, completed audiobook) and expressly no others. Your feedback must be specific and detailed, both sides must act in good faith, and the producer makes requested revisions “to the extent practicable”. Nowhere do the terms define an “error” versus a “change”. The working boundary is the final manuscript you supplied at the start: the producer’s job is to deliver that text as audio, so correcting a misread line sits inside the revision rounds. Rewriting a passage after hearing it read aloud changes the underlying text instead. The terms make no provision for that, so treat it as a separate agreement with your producer, not a revision request. Two quiet defaults raise the stakes here. No response within the review window counts as approval, and approval is final once given. Reject the producer’s second revision of the completed book and the remaining exit is termination with a cancellation fee; an impasse at the 15-minute stage, by contrast, lets either side walk away without one. That asymmetry is one more reason to take the checkpoint seriously.

How Much Does Audiobook Production Cost Per Finished Hour?

Professional audiobook production in the US typically costs between $150 and $400 per finished hour. Union-scale narrators and premium studios charge $600 or more. A “finished hour” is one hour of completed, mastered audio, not one hour of studio time. Author’s Republic, one of the larger audiobook distributors, defines it as a polished, edited 60 minutes of audio. Check whether the production company applies a rounding rule, though: some bill any partial hour of finished audio as a full hour.

To estimate your own book, divide the word count by 9,300. That is ACX’s planning figure for how much text most narrators perform in one finished hour (roughly 155 words a minute). Let’s work it through for an 80,000-word novel. First, divide 80,000 by 9,300, which gives about 8.6 finished hours. Next, multiply 8.6 by the quoted rate. At $250 per finished hour the production costs about $2,150; at $400 it costs about $3,440. Those figures rise to $2,250 and $3,600 if the company rounds the 8.6 hours up to 9.

Each of those finished hours hides around six hours of real labour, sometimes more: performance, retakes, editing, and mastering. That is why the rate looks high next to an hourly wage. ACX’s own cost guidance splits a retail-ready hour into roughly $200 of narration and $200 of post-production. That split is a useful sanity check when a quote bundles everything into one number.

Royalty Share: The No-Upfront-Fee Alternative

There is one common alternative to paying cash up front. Under a royalty share arrangement, the narrator produces the book for no fee and instead takes half of your royalty earnings. On ACX this split runs for a seven-year exclusive commitment. It removes the upfront cost, but experienced narrators tend to accept royalty share selectively. On a book that sells well, half the royalties can also end up costing more than the flat fee would have.

Editing, Mastering and Proof-Listening: What’s the Difference?

Editing removes mistakes from a recording. Proof-listening checks the performance word-for-word against the manuscript, and mastering sets the technical character of the finished files. These are three separate jobs, and different people usually do them. A quote that says “editing and mastering included” has not necessarily promised the one in the middle.

Editing is the assembly stage. An editor keeps the chosen takes and removes repeated lines, coughs, chair creaks, and overlong pauses. Each chapter comes out as a clean, continuous performance. Proof-listening (some studios call it QC, for quality control) is different. A person follows the manuscript while listening. This pass catches skipped sentences, misread words, and duplicate takes the editor left in. It is the audio equivalent of proofreading a typeset book against the approved text.

Mastering is the final technical pass. It evens out loudness between chapters, cleans up the background noise level, and encodes the files to the retailer’s published specification. The exact numbers ACX requires appear in the FAQ below. Automated upload checks are no substitute for the human pass. ACX notes that its own Audio Analysis tool measures loudness, peaks, and file format. It does not detect noise-floor problems (constant background hiss sitting under the voice) or editing errors. A file can pass the robot and still say “chapter seventeen” twice. So when you compare quotes, check that proof-listening appears as a named line item.

What Does a Production Contract Include, and What’s Left Out?

A typical production agreement covers casting, direction, recording, editing, proof-listening, mastering, and a defined number of revision rounds. It does not usually include distribution, cover artwork, store metadata, or the raw studio files. Each of those exclusions is worth a question before you sign.

Revision Rounds and File Ownership

Revision structures vary more than authors expect. ACX allows the two-plus-two pattern described earlier: two rounds at the 15-minute checkpoint and two on the completed book. Other services instead define a set number of “opportunities to cure” before extra charges apply. A cure is simply a chance for the studio to fix a fault you have flagged. Neither structure is wrong, but you should know which one you are agreeing to and what the response deadlines are.

Ownership and possession are separate questions. Under ACX’s standard terms the finished audiobook is a “work made for hire”. That legal phrase means the copyright belongs to the person who commissioned the work, not the person who performed it. In practice, the finished sound recording is yours. The raw session files (the individual takes and the engineer’s project files) are a different matter. They rarely form part of the standard deliverable, so if you want an archive copy, name it in the contract. The same pattern holds with cover designers and formatters. The finished files are the deliverable, and working files change hands only when the agreement says so.

AI Clauses and What Stays Your Job

Finally, ask how the contract handles artificial intelligence, because the fine print here has already caught authors out. In early 2023, authors and narrators discovered that a major audiobook distributor’s standard agreement quietly licensed their finished audio files for machine-learning training. The companies involved halted the practice after a public outcry. So read for it: check whether your agreement grants anyone the right to train AI on your recordings, and narrow or strike that wording if it does.

The narrator’s side carries the same exposure. SAG-AFTRA, the performers’ union, bars its members from working with companies seeking digital replicas (AI copies of a narrator’s voice) unless the company has signed the union’s agreement. Its stated guardrails are consent, compensation, and control. A clear clause protects both sides. Nobody can clone the narrator’s voice, and nobody can reuse your recordings in ways you never agreed to.

Retailer policy is the third reason to care, because it decides where an AI-narrated master can be sold at all. ACX’s submission requirements call for human narration unless Audible has specifically authorised otherwise. An AI-narrated audiobook therefore cannot take the main route into Audible. Spotify sits at the other end: it accepts AI-narrated audiobooks, labels them for listeners, and now offers authors its own AI narration tool. If AI narration is anywhere in your plan, check each store’s current policy before production starts. Store metadata also stays your job, including whether the audiobook needs its own ISBN.

Is Distribution Included? Production Companies, Publishers and Distributors

Making the master and putting it in stores are two different transactions, covered by different contracts. Three kinds of company operate here, and they take very different shares of your book.

A production company works for a fee. You pay per finished hour, you keep all rights, and when the master is delivered the relationship ends.

An audiobook publisher pays for production itself. In exchange it acquires the audio rights, usually for a term of years, and pays you a royalty on sales. No money leaves your pocket, and no rights stay in it either.

A distributor takes your finished master and licenses it to retail and library platforms. Some companies sit in more than one category. ACX, the source of the production terms quoted earlier, is both a production marketplace and a distributor. Its contract therefore needs reading with both hats on.

Where the Distribution Options Differ

The distribution options each work differently, so compare them on their own terms. ACX distributes to Audible, Amazon, and Apple Books. Under Audible’s new royalty model, effective 26 May 2026, exclusive distribution (those three stores only) pays a 50% royalty. Non-exclusive distribution (you may also sell elsewhere) pays 30%. The distribution term is seven years, and ACX does not place titles in libraries.

Spotify for Authors is Spotify’s own upload portal. It has accepted free direct uploads from self-published authors since August 2025, non-exclusively, with the uploader keeping ownership.

Apple Books takes no direct uploads from indie authors. Reaching it independently of ACX means using an approved aggregator, meaning a distributor that feeds many stores and library apps from one upload. Author’s Republic, for example, passes on roughly 80% of what the stores pay out. Library apps sit behind aggregators too. Our post on how libraries buy self-published books covers how those purchases actually work.

Treat the engagement as four decisions rather than one purchase. Estimate the finished hours from your word count before asking for quotes, so you can compare rates on equal terms. Confirm that proof-listening is a named line item and count the revision rounds. Ask exactly which files you will receive, and add the archive files if you want them. Then make the distribution choice separately, on its own merits. The company that made your master does not have to be the company that sells it.

Frequently Asked Questions

How many finished hours will my audiobook be?

To estimate an audiobook’s running time, divide the manuscript word count by 9,300, the figure ACX uses for average narration speed. The result is the length in finished hours. A 60,000-word novel comes to about 6.5 finished hours, and a 100,000-word epic to just under 11. Quotes priced per finished hour scale directly from this estimate, so also confirm how the company bills a partial final hour: some charge it proportionally, while others round it up to a full hour.

Can I narrate my own audiobook and still use a production service?

An author who narrates their own audiobook can still hire a production service for everything else. Many studios coach and direct author-narrators. Most also sell post-production on its own, meaning editing, proof-listening, and mastering of audio you record yourself. Rates for that package are commonly quoted at around $75 to $150 per finished hour. Your recording space and equipment still have to be good enough to meet retailer specifications.

How long does professional audiobook production take?

Once an author approves the narrator’s edited 15-minute opening sample, full recording and proofing typically takes two to six weeks. The exact schedule depends on the book’s length and the narrator’s availability. Retailer review adds more time after submission; ACX’s audio review takes up to ten business days before the audiobook goes on sale.

What audio specs do the finished audiobook files have to meet?

ACX publishes exact technical requirements that a mastering engineer works to. The list includes MP3 files at 192 kbps constant bit rate or higher and a 44.1 kHz sample rate. Average loudness must sit between -23 dB and -18 dB RMS, with peaks no higher than -3 dB and background noise no higher than -60 dB RMS. Each file holds one chapter, and no file may exceed 120 minutes. Other platforms publish similar but not identical profiles, so the contract should name the platform the master must pass.

What should you check before signing with an audiobook production company?

Before signing with an audiobook production company, check five things. Get the per-finished-hour rate and the finished-hour estimate in writing. Confirm proof-listening is a named line item. Count the revision rounds and their response deadlines. Agree exactly which files are delivered, including any archive files you want. Finally, look for a clause covering AI training and voice cloning of the narrator’s recordings.