Hear the different layers
Sound in an interactive storybook usually does several jobs. Narration speaks the written words. Dialogue may add character voices. Effects identify actions or events, while music establishes mood. Treating all of these as one soundtrack makes it difficult for a reader to keep what is useful and remove what is distracting.
A reader may want narration without background music, or environmental sound without a voice reading over a partner. Independent controls make those combinations possible. Labels should describe the result clearly: “Music off” is easier to interpret than an unexplained symbol whose appearance changes after every tap.
Keep narration in the reader's hands
Useful narration can pause, resume and repeat a manageable passage. It does not need to carry the reader automatically to a new page. When a reader stops to ask about a word, the interface should accommodate that conversation rather than continue through the next paragraph.
Text highlighting can connect spoken words with their written forms, but it should remain legible and avoid unnecessary flashing or sweeping effects. Some readers prefer a whole sentence highlighted; others prefer no tracking at all. A narration feature is not a substitute for selectable text, a meaningful reading order or screen-reader compatibility.
Recorded narration and synthetic speech also differ. A recording can offer a deliberate performance; a device voice may allow speed and voice preferences. Neither approach is universally best. What matters is intelligibility, appropriate pacing and whether the reader can choose a comfortable arrangement.
Let silence retain the story
Suppose a distant bell tells a character that the ferry is leaving. If the bell is the only clue, a reader without audio loses an important cause of the next event. The written passage or an appropriate caption can make the same event available: “The departure bell rang across the harbour.” The alternative should convey meaning, not merely report that a sound occurred.
Captions for synchronised audiovisual material and transcripts for audio passages serve related but different purposes. Captions are timed to the media; a transcript can be read separately. Both may need to identify speakers and significant non-speech sounds. A transcript of dialogue alone will not explain a crucial knock at the door.
Borrow lessons from listening traditions
Sergei Prokofiev's Peter and the Wolf connects characters with musical instruments and themes. It is a useful example of sound carrying narrative associations, not evidence that every storybook needs a musical identity. The reader still needs context to understand what an association means.
Margaret Wise Brown's Goodnight Moon, illustrated by Clement Hurd, demonstrates a different kind of sound design: the cadence and repetition of written language. A quiet read-aloud can create rhythm without an added soundtrack. Arnold Lobel's Frog and Toad stories likewise leave room for a reader's own pacing, pauses and character voices.
DAISY, the Digital Accessible Information System, is associated with accessible digital reading and structured navigation. EPUB Media Overlays provide a mechanism for synchronising text and audio in EPUB publications. These references matter because access is not only about a pleasing voice; it is also about finding, navigating and returning to the right passage.
Design for the room around the device
A story may be opened in a shared classroom, a library or a home where someone else is resting. Starting audio only after a deliberate action respects those settings. Headphones are one possible accommodation, not an answer to every need. A reader may be unable or unwilling to use them, and sound should not be required to unlock the story.
During shared reading, ask which sounds help and which can go. Leave time for the listener's response before replaying a passage. Repetition can be welcome, but a child or adult should not have to endure a loud effect each time they revisit a sentence.
For creators, review a complete route with the device muted. Follow every instruction, decision and reveal. Then listen without looking and note where additional description or navigation support is needed. These checks do not replace testing with disabled readers, but they reveal assumptions that are easily missed when a designer experiences all channels at once.