How an audiobook actually gets made
You have made a book, so an audiobook looks like one more export from it. It is not. It is a separate production with its own stages, and each stage can hold up every stage after it.
We make no audio tools, so there is nothing here we are steering you toward. This is the sequence, so you can see what you are agreeing to before you agree to it.
A script is not a manuscript
A narrator cannot read your manuscript as it stands. Someone — you, before any recording starts — has to go through it and decide what happens to everything on the page that is not plain speech. The text has to be final first: a line changed after recording is a line somebody records again, which is one of the reasons Knowing when a draft is finished treats the words going final as a production boundary rather than a mood.
- Chapter and part headings. Usually spoken, in a form you fix once and keep. Inconsistency between chapter one and chapter twenty is audible.
- Footnotes and endnotes. Read where they fall, gathered at the end of the chapter, or dropped. All three are defensible; pick one for the whole book rather than deciding note by note. Footnotes, endnotes and the reader's patience covers what each kind of note is doing.
- Tables, charts and figures. None of these can be read aloud as themselves, and "as you can see in figure three" is meaningless in audio. Either the sentence states the finding or the thing moves to a supplementary document — and distributors differ on whether they accept one and how it reaches the listener, so confirm that before you depend on it.
- Front and back matter. Copyright page, dedication, acknowledgments, about the author. Most audio editions read short versions of these and drop the index, because an index is a page-number object and audio has no pages. An epigraph needs a decision of its own — read, attributed, or cut — and Epigraphs and dedications is the shape you are converting.
- Anything visual in the prose. A letter set in a different face, a text message, italics doing work the ear cannot hear. Each needs a spoken convention agreed before the microphone is switched on, because everything Letters, texts and documents inside a novel does with type has to be done with the voice instead.
Do this in a separate copy, so the manuscript your print and ebook editions are built from stays clean — the same discipline EPUB and print PDF are two different books argues for across the other two formats. While you are in there, build a pronunciation list: character and place names, invented words, foreign phrases, anything you have only ever seen written down.
Casting
Then you decide who reads it — you, or someone you hire. It sets the cost, the schedule and the ceiling on quality, so it gets its own page: Narrating it yourself, or hiring someone. If you are hiring, the work is casting rather than shopping, and Finding and auditioning a narrator covers where narrators are found and what to agree before anyone records.
Recording
Recording is long, repetitive and unglamorous, and the room has to be quiet in a way ordinary rooms are not: a fridge compressor two rooms away, traffic, a plane, your own chair. That is the whole subject of Where to record, and what you need, which is about the space before it is about the microphone. Sessions also have to match each other — a book recorded across three weeks can arrive with a different voice in it at chapter nine, because the narrator had a cold, moved the microphone, or simply got better; the habits that stop that are in The recording session.
Editing, then mastering to a spec
Editing is where the takes become a performance: bad takes cut, breaths managed, pauses evened out, the mouth noise nobody notices until it is in headphones. Editing, mastering and hitting the spec is the stage in detail, including why most rejections happen here.
Mastering is a separate, measurable job. Every distributor publishes a technical specification covering file format, sample rate, loudness, noise floor, the silence at the head and tail of each file, and how the book is split across files. Those requirements differ by distributor and are revised, so read the current spec sheet from the one you intend to use rather than a number from a blog. The failure is binary: a file that misses the spec is rejected until it is fixed, however good the performance is.
In the tools. We do not make audio, and nothing here needs us. What the audio edition still needs is the rest of the book around it — a print interior and an EPUB from one manuscript, and cover artwork the audio edition can be built from. Start your book → · about the tools
The proof-listen
Someone then listens to the whole finished audiobook with the script in front of them. Not a spot check — all of it. This is the stage authors try to skip, and it is what catches a dropped word, a sentence recorded twice, a name pronounced two ways, and the level shift where two sessions were joined. If you hired a narrator, ask who does it and whether it is included — Proof-listening before you publish sets out what you are listening for and how to write notes an engineer can act on.
Where the time goes
The industry's unit is the finished hour: one hour of the completed audiobook as a listener hears it. Cost and schedule are both quoted in it, per finished hour rather than per word or per chapter.
Two facts follow, and they surprise people. Recording takes longer than reading, because of retakes, restarts and interruptions. And finished audio takes several times longer again: editing, mastering and the proof-listen all happen at a multiple of the running time. That multiple varies with the book, the narrator's experience and how clean the raw recording was, so nobody can quote you a fixed one honestly.
Plan a full-length book in weeks. Anyone who tells you a weekend has not done one.
What makes a book easy or hard
Easier: straightforward nonfiction, single voice, no apparatus. The narrator reads sentences and the sentences work.
Harder: fiction with a large cast. Every character needs a voice that is distinguishable, consistent across twelve hours, and sustainable for the narrator. Dialect raises the difficulty again.
Hardest to plan: anything with tables, diagrams or footnotes — not because it cannot be done, but because each is a decision that must be made before recording, and making it afterwards means recording again.
Packaging, and the fact underneath all of this
Last comes the upload: files named and ordered as the distributor requires, the square cover art, a description, and metadata that has to match the artwork exactly. Which route you upload to — direct, an aggregator, or one ecosystem exclusively — is decided in Getting an audiobook into stores, and a royalty-share narrator agreement can settle it for you before you ever look.
Which points at the fact worth holding on to. Your audiobook is not a version of your book. It is a third edition alongside print and ebook, with its own ISBN, its own listing, its own metadata and a cover that is a square rather than a tall rectangle — see Audiobook cover art is a square, and that changes everything.
Treat it as a separate product with its own schedule and it behaves. Treat it as an export, and it is late.
Make your book
The tools are free to use: name your book, bring your manuscript or your artwork, and work as long as you like. You only pay when you want the files without the preview watermark.
Next: Narrating it yourself, or hiring someone · More: How an audiobook earns its money back · Should you make an audiobook? · Audiobook cover art is a square, and that changes everything · Do you need your own ISBN?