Your heroine is called Siobhan. You've heard her name in your head for three years as shiv-AWN, and every reader who grew up with the name hears it the same way. Someone who has never met a Siobhan reads the letters, and the letters say something else.
That's a familiar audiobook pronunciation problem, and it isn't about who reads the book. Spelling doesn't carry sound: a narrator, human or synthetic, can only read what's on the page until someone who knows the name says how it goes. In an audiobook, that someone is the author. Here is how you tell your production.
The short answer
Open any chapter in the studio, open Martin, and say how the name sounds:
Siobhan is said shiv-AWN, like the Irish name, everywhere in the book.
Martin, Midsummerr's AI audio director, puts a few readings of the name in the chat. Pick one and press Hear to listen before you decide; your book audio stays unchanged while you do. Choose the reading you mean, and Martin saves it to your book's pronunciation guide. Every chapter you produce from then on says Siobhan your way.
Then Martin asks about the audio you already have. Say "yes, this chapter" and he re-records the lines in this chapter where the name is said. Ask for the rest of the book and he shows you an approval card first, with how many finished chapters it covers.
Martin is the AI audio director built into the Midsummerr studio, and Midsummerr is the only audiobook platform with one. Direct your audiobook by talking to it shows the chat in more detail.
Ready to try it on your own book?
Start with 5,000words free →From your note to the finished book
Setting a pronunciation happens in stages, and you decide when each one happens: first you hear it, then you save it, then you place it.
| What you say to Martin | What Martin does | What changes in your audio |
|---|---|---|
| "Siobhan is said shiv-AWN" | Puts a few readings of the name in the chat | Nothing: you're listening |
| You press Hear on each reading | Plays that reading in the chat | Nothing |
| You press Choose on the one you mean | Saves it to the book's pronunciation guide | New chapters say it your way |
| "Put it in this chapter" | Re-records the lines with the name in this chapter | Those lines only |
| "Do the rest of the book too" | Shows an approval card, then re-records your finished chapters | The lines with the name in those chapters |
If none of the readings is the one you hear, say so ("closer, but the stress is on the second syllable") and Martin offers others. You can also give him a sentence or two from the book, and he makes test takes of the name in that context, so you hear it inside a line before you choose.
Your book's pronunciation guide
The reading you choose is saved to your book's pronunciation guide, not to one line. That's why a name you set in chapter three is still right in chapter thirty, and why you only set it once.
Three things stay as they are:
- Your manuscript. The guide tells the production how to say the word. Your text keeps your spelling.
- Everything else in the scene. Only the lines where the name is said are re-recorded. The rest of the cast, the music and the sound effects stay put.
- Your finished chapters, until you approve. Martin asks before he touches audio you already have, and a long book takes a while to work through.
If you change your mind about a reading, tell Martin. He plays you the reading the guide has now and updates it to the one you choose. Removing a word from the guide goes through an approval card too, because it can't be undone.
One line or the whole book
Sometimes what you want is one moment, not the name: a single line said differently, while the name is how you want it everywhere else. Tell Martin which line, and he can record a fresh take of that line alone without touching the guide. How to change one line of your audiobook covers line notes in full.
The guide is for a name the whole book has to say the same way: a character, a town, a family name, an invented word. If the same reading would be right on every page, it belongs in the guide.
What to set before you produce the whole book
The names worth setting are the ones listeners will hear most and the ones a reader is most likely to say another way:
- Character names with spellings that don't match their sound, or more than one common reading.
- Places, real or invented, especially ones the characters say often.
- Invented words: titles, species, spells, brand names inside the story.
- Borrowed words from another language that your characters say naturally.
Authors have had to spell this out for a long time. In E. Nesbit's Five Children and It, the sand-fairy asks whether the children don't know a Psammead when they see one, and one of them answers, "A Sammyadd? That's Greek to me" (Project Gutenberg text). Nesbit let a character say the word the way it sounds. In an audiobook, the guide does that job, and the listener only ever hears it said one way.
When you upload a book, the Pronunciation page may already list names it found, with readings you can hear and approve (how upload suggestions work). Martin is the quick path when you're already listening in the studio and hear a name you want said differently.
Hear a full cast first
Names matter most on a full-cast book, where they pass between many voices. To hear what that sounds like, listen to Five Children and It, produced in Midsummerr with a narrator and a full cast of character voices. How to make a full-cast audiobook covers the whole production, and the features page shows the rest of the studio.
Try it on your own book
Upload a manuscript, generate a chapter, and listen for the names you know best. Open Martin and tell him how they sound. The first 5,000 words are free, and re-recording lines for pronunciation is always free.
FAQ
How do I get a character's name pronounced right in my audiobook?
Open the studio, open Martin and say how the name sounds, for example "Siobhan is shiv-AWN". Martin offers a few readings in the chat; you hear each one and choose. The chosen reading goes into your book's pronunciation guide, and every chapter you produce afterwards uses it.
Will changing a pronunciation regenerate my whole audiobook?
No. Only the lines where the word is said are re-recorded, in the chapters you approve. The rest of the cast, the music and the sound effects stay as they are.
Do I need to know phonetic spelling?
No. Describe the sound the way you'd tell a friend ("the h is silent", "stress on the second syllable"). Martin writes the readings, and you choose by listening.
Does the pronunciation guide change my manuscript?
No. The guide tells the production how to say the word. Your manuscript keeps your spelling.
Does it cost credits to change a pronunciation?
No. Hearing readings and re-recording lines for pronunciation is free.




