Audiobook Pronunciation Guide for Narrators: What to Include and How to Write It

A name read wrong in chapter one is read wrong for the whole book. The narrator says it the way it looks on the page. By chapter 20, that reading is baked into the master.

An audiobook pronunciation guide for narrators stops that before the first take. This piece is for both sides of the handoff: authors sending a book to a narrator, human or AI, and narrators receiving the guide.

More books are making that handoff. The Audio Publishers Association survey puts US audiobook sales revenue at $2.43 billion in 2025, 9% over the previous year. US publishers reported over 750,000 active titles in 2025, a 43% increase from 2024.

From the producer's chair, the pronunciation guide is the cheapest document in the production. It costs nothing to write, and every fix it prevents costs booth time. Below: what goes in the guide, how to write fast respellings, how narrators use them, and how to check the audio afterward.

What Is an Audiobook Pronunciation Guide, and Why Does It Matter?

An audiobook pronunciation guide is a short, separate list of every word a narrator could stumble on, each with a plain respelling and a note. You send it before recording, so each name is said one way for the whole book.

Some authors call it an audiobook pronunciation list or a style guide. The name matters less than the timing. It has to land before the first session.

Timing matters because fixes after recording are expensive. Someone has to find every instance of the wrong word, then re-record each passage so it matches the takes around it. As our cost breakdown notes, fixes on a human production mean rebooking the narrator.

Here is the math. Human narration runs $200 to $400 per finished hour, and a finished hour is roughly 9,300 words. Each finished hour takes 2 to 4 booth hours to record.

So a 75,000-word novel is about 8 finished hours and $1,600 to $3,200 of narration. That is 16 to 32 hours in the booth. A lead character's name can turn up in every chapter, and every misread line goes back on the narrator's calendar.

Our audiobook production cost breakdown runs the full numbers. The short version: retakes are booked in booth hours, and a pronunciation guide is the one document that keeps them off the invoice.

The listener pays too. The Journal of Speech, Language, and Hearing Research published a lab study on listening effort that measured pupil dilation in listeners with cochlear implants. It tested single mispronounced words in sentences. Its finding: "When a word was mispronounced, there was a brief but significant elevation in effort."

The study did not test audiobooks. Still, it is reasonable to expect a recurring wrong name to add that cost each time it comes up in a ten-hour read.

What Should You Include in a Pronunciation Guide?

A pronunciation guide holds two kinds of entries. Word entries tell the narrator how to say one specific term. Rule entries set how a whole class of text is read, such as years or clock times.

Start with the words. Read your manuscript for every term in these categories:

• Character names: every named person, including walk-ons who appear once.
• Place names: real and invented, with extra care for any that aren't English.
• Invented terms: spells, species, titles, and anything else you coined for the book.
• Foreign words: any phrase a character says in another language, with the language named.
• Acronyms: whether each one is spelled out letter by letter or said as a word.
• Numbers, dates, and times: one rule for each, so every year and every clock time is read the same way.
• Words with more than one valid pronunciation: mark the reading you want wherever context alone won't settle it.
• Possessives and invented plurals: how a coined name sounds once it takes an s.
• Near-identical names: two names that look alike on the page and must sound different in the ear.
• Deliberate variations: names that different characters say differently on purpose, with a note on who says which.

Words with two valid readings are easy to miss, because they look ordinary on the page. Google Play Books narration tips, written for Google's auto-narration tool, flag it: "Some words like 'read,' 'minute,' and 'present' have more than one pronunciation." The guide covers the cases where the sentence alone won't tell the narrator which one you meant.

Here is what a finished guide looks like. This is the sample from our own style guide.

Term	Say it	Note
Kehlani	keh-LAH-nee	character
Caerwyn	CARE-win	place
2024	twenty twenty-four	year rule
3:15 p.m.	three fifteen p m	time rule

Look at the last two rows. They are rules, and one entry covers every year and every time in the book. That is how a short guide covers a long manuscript.

A page or two is plenty. Send it as a separate file. Our guide to how to format a manuscript for narration covers the rest of the handoff.

How Do You Write Phonetic Respellings a Narrator Can Read at Speed?

Write every respelling for speed. The narrator reads it once in prep, then has to produce the word mid-sentence with the script moving.

A respelling that works at speed follows a few habits:

• Hyphenate the syllables and capitalize the stressed one, as in keh-LAH-nee.
• Build from common English sound chunks the narrator already knows, like CARE in CARE-win.
• Add a rhymes-with note when a chunk could be read two ways.
• Record a short audio clip for any sound a respelling can't carry.

We tell authors the same thing every time: you don't need IPA. A simple phonetic hint in a separate list is enough. For invented names and tough proper nouns, add a phonetic line in your notes, such as Kehlani (keh-LAH-nee).

IPA has its place. The International Phonetic Association publishes the official IPA chart, and its symbols are precise. They are also slow to read mid-sentence, so use IPA only when the narrator asks for it.

Spoken-audio studios treat name pronunciation as reference work worth writing down. The Library of Congress National Library Service (NLS), for one, publishes the NLS Say How? guide, "A Pronunciation Guide to Names of Public Figures."

How Do You Handle Foreign Words, Accents, and Dialects?

Foreign words need a decision before they need a respelling. When several pronunciations are defensible, pick one and say which in the note. A narrator who has to choose in the booth will choose differently on different days.

Our rule on accents is narrow. Note accents only when the read should change, as when a character code-switches or a place name isn't English. Everywhere else, the narrator's natural read is the right one.

Regional variation gets the same treatment. When a word is said differently in different places, write the region in the note. The narrator then knows which version to aim for, and you can check it against that target.

Then give the narrator something to listen to. The International Dialects of English Archive holds roughly 1,800 samples from 135 countries and territories, and more than 170 hours of recordings. A clip of a real speaker carries what spelling can't.

How Do Narrators Use a Pronunciation Guide Before Recording?

We recommend starting with a full read of the book before recording. The narrator should mark tricky words phonetically in the manuscript and query whatever the guide misses. Your guide is the cue sheet for that read-through.

If you narrate, ask for the guide before prep starts. If you write, expect questions, and treat each one as a fix you didn't have to pay for. The prep we recommend runs in four moves:

• Annotate the manuscript: mark each tricky word where it appears, so the respelling is in view at the moment of the read.
• Check the guide against the text: confirm every name in the book is on the list and every spelling on the list matches the book.
• Ask the author about gaps: query anything missing before the first session, while the answer is still cheap.
• Decide the edge cases up front: settle numbers, acronyms, front matter, and footnotes before the first take.

Character voice notes are the companion document. Give each speaking character an age, an origin, an accent if there is one, and one defining trait. Voices should be distinct enough to tell apart in a fast exchange, and they don't need cartoon accents.

We direct AI narration the way a studio directs a human narrator, and the guide feeds both. A producer marks affect and pace on every line and checks names and pronunciation against the manuscript before you hear a word. Our walkthrough on how to cast and direct a narrator shows the process line by line.

Distinct voices are easier to hear than to describe. Listen to our multi-voice audiobook samples to hear a cast that stays apart in dialogue.

What Ambiguities Should You Settle Before the First Session?

Some ambiguities aren't words at all. They are formatting choices the narrator will otherwise make on the fly, and possibly make differently in chapter 3 and chapter 30. Settle each one in writing, as a rule entry in the guide.

Item	Question to settle	Example rule
Numbers and years	Digits read as one number or in pairs?	2024 is read twenty twenty-four
Times	How is a clock time voiced?	3:15 p.m. is read three fifteen p m
Acronyms	Letters or a word?	IPA is read as letters, I-P-A
Abbreviations	Read literally or expanded?	Dr. is read Doctor
Non-English words	Which pronunciation, and how much accent?	Pick one and name the language
Front and back matter	What gets read aloud?	List each page that is read, in order
Footnotes	Read, folded in, or cut?	One decision for the whole book

For abbreviations, spell out the ones that sound odd when read literally (Dr., St., etc.) or flag them in your style guide. St. is the classic trap, since it can be Street or Saint. One line in the guide removes the guess.

How Do You Check Pronunciation After Recording?

Most advice stops at the handoff. The guide earns its keep a second time after recording, when you proof the audio against it.

Proof with the manuscript and the guide open side by side. Listen for every entry on the list, and mark each miss as you hear it. Names pronounced three ways and a scene read in the wrong register are the common failures.

Flag one clear note per issue, with the location and the fix: "Chapter 4, paragraph 12: should be breath not breathe." Retake the line, not the whole book.

The proof is only as good as the text behind it. Send the final manuscript, not a draft with track changes or comments still open. A narrator who reads an open comment can't know which version you meant.

Test early. Hear a real chapter before you commit to the whole book, because one chapter shows whether the guide holds up. Send your manuscript for a free first chapter and it comes back in about a day.

On our productions, that check is built in. We run a pronunciation pass in production, with pronunciation review in every pack. When you flag a line in the delivery room, we retake that line.

If you are cloning your own voice, the sample becomes part of the guide. Say names and invented words the way you want them forever. Our guide on how to record a voice sample covers what to read.

Summary

An audiobook pronunciation guide for narrators buys one thing: a book that sounds the same in its last chapter as in its first. That consistency is what you keep, in the master and in every sale after it.

• Keep one separate list, a page or two, sent as its own file.
• Respell in plain syllables with the stressed syllable in caps.
• Write rules for numbers and acronyms alongside the word entries.
• Note accents only where the read should change.
• Send the final manuscript with every tracked change resolved.
• Check the audio line by line and retake only the line that missed.

Your next step is small. Open the manuscript and list every proper noun now. That list is the first draft of your guide.

FAQ

Do I Need IPA in My Pronunciation Guide?

No. Plain respelling is faster to read mid-sentence, so add IPA only if your narrator asks for it.

How Do Narrators Handle Words That Aren't in the Guide?

The best fix is to mark unfamiliar words during the read-through and query the author before recording. Answer those queries quickly to keep them out of the booth.

Should I Send My Narrator the Final Manuscript or a Draft?

Send the final manuscript only, with every track change resolved and every comment closed. A narrator records what is on the page.

Do Footnotes Get Read Aloud in an Audiobook?

Only if you decide they should. Choose in advance whether each footnote is read, folded into the text, or cut, and write that rule into the guide.