Writing an audio story: what the format really changes

When we started preparing the scripts for Voctale, we assumed the adaptation would be light work: take stories written to be read, hand them to a voice, adjust a few turns of phrase along the way. The first recording session settled the matter in ten minutes. The sentences stood up on the page and collapsed the moment they were spoken. Too many subordinate clauses, pronouns whose thread got lost, descriptions that assumed a way back the ear simply does not offer.

Writing an audio story is not the same as voicing a text. It is a different writing regime with its own constraints, and it is better to know them before you start than to discover them in the edit. Here is what the format actually shifts, in the order we noticed it.

The listener cannot go back

A reader who drifts off scrolls back three lines without even registering it. A listener moves forward. If they lose a name, a date, a causal link, they do not go looking for it: they carry on with a gap, and the gap widens. Two minutes later they have stopped following the story and started trying to reconstruct it.

The practical consequence is easy to state and hard to hold: one piece of information per sentence. Constructions like “Marta, whom her aunt had taken in the winter the mill burned down, refused to open the door” work perfectly on the page and dissolve out loud. They need unrolling: Marta refuses to open the door. Her aunt took her in years earlier, the winter the mill burned down.

We also learned to repeat first names far more often than we would in print. A “she” is enough on the page, where the eye finds the antecedent instantly; in audio, two female characters in one scene make every pronoun ambiguous. That is not clumsiness, it is signposting.

Writing for a voice that breathes

A written sentence has no duration. A spoken one does, and it is paid for in breath. An actor who reaches the end of a line out of air produces a slack landing, and that slackness is audible even when nobody could say where it comes from. Since then we read every paragraph aloud, watching for the point where the breath falls: if there isn’t one, the sentence is too long.

The text also starts producing sonic accidents you cannot see when rereading in silence. Runs of sibilants, two words whose ending and beginning glue together, a proper noun that becomes unrecognisable once it is chained to the word before it. It is pure sound work, close to what we described in our piece on finding your writing voice: rereading stops being a check and becomes an act of listening.

Silence is part of the text

In print, the white space between two paragraphs costs nothing: the reader crosses it at their own pace. In audio that space has a real duration, and the duration carries meaning. A second and a half after a revelation is a breath. Four seconds is a scene changing direction. Zero seconds, and the revelation goes unnoticed.

We now mark silences in the script, like stage directions, with a rough length. It looks excessive until you try it; in practice it stops the edit from having to rescue intentions nobody ever wrote down. Silence is the one piece of punctuation the format adds over the page, and probably the most powerful one.

Episode length replaces chapters

A chapter ends when the author decides it ends. An audio episode ends when the listener is about to get off the bus. We work with eight to twelve minutes for adult stories, and much shorter running times for children’s material – the rhymes and bedtime pieces of Filastrocche della Luna Storta rarely go past four minutes, because beyond that attention scatters and the text is doing a different job.

Each episode has to open something and close part of it. No abrupt ending, and no gratuitous cliffhanger either: real movement, plus a question. The mechanics are close to those of interactive stories, where every segment must stand on its own while calling for the next.

Description moves into sound

This is the change that demanded the most unlearning. In print, describing an abandoned house runs through adjectives. In audio the house exists through a floor that shifts, a window that never closed properly, the way a voice rings in an empty room. The adjective becomes almost always redundant.

The opposite trap follows immediately: doubling the sound with the text. If the rain is audible, nobody needs a voice announcing that it is raining. Text and sound have to split the work rather than do it twice. We now mark in the script whatever the ambience will carry, so it does not get rewritten in full sentences later.

Read aloud, very early

The only method that has saved us time is recording a rough read of the first draft, with no performance intent, just to hear it. A decent USB microphone is plenty at that stage – the point is not the quality of the take, but spotting the sentences that resist. Three quarters of our corrections come out of that listen, before the text is considered finished at all.

On the theory side, the literature on writing for radio remains the most directly applicable resource: it is fifty years ahead of narrative podcasting and wrestles with the same problems. For practitioner accounts we also follow Allison Lister on story construction, and the SubwayPress columns whenever they touch on audio formats.

What audio changed in the rest of our writing

The side effect arrived quickly. The texts we now write for the page have shorter sentences, more explicit connections, and far less dependence on the reader’s comfort in going back. The audio format works as a developer bath: it makes audible everything a written sentence manages to keep vague.

If you are hesitating, take a text you know well – a story already written, a scene you believe is finished – and read it aloud while recording yourself. Five minutes will tell you what the format changes. It is the same kind of displacement a change of pen name or a change of medium produces: a different frame, and the writing starts moving again.