Start with what fits, not what you'd like to say
Line length is set by the gap, not by how much there is to describe. If a silence between two lines of dialogue lasts three seconds, you have roughly eight to ten spoken words at natural pace — not a full sentence with two qualifiers and a clause. Write to the silence you're given, then cut until it fits.
Describe, don't interpret
Objective over subjective
“She looks sad” is an interpretation. “Her shoulders drop, she looks away” is a description. Leave the emotional read to the viewer — the same way a sighted viewer infers sadness from those same visual cues rather than being told the answer.
Present tense, active voice
“He opens the door,” not “The door is being opened by him.” Present tense and active voice read faster at natural pace and match the immediacy of action happening on screen right now.
Prioritize what changes the story
Not every visual needs describing. Describe what a sighted viewer would need to follow plot, character, or setting — skip what the dialogue or sound design already conveys on its own. A script that describes everything visible runs out of gap space fast and buries the details that actually matter.
Match vocabulary to the register of the piece
A nature documentary and a sitcom don't share a vocabulary. Description should sound like it belongs in the piece it's describing — not read as generic template text dropped into every project the same way.
Read it back against the gap, out loud
The only real test of whether a line fits is reading it at natural pace against the actual timed gap — not counting characters on a page. A line that looks short can still run long once it's spoken with the intended emphasis.
Where Synchrogen fits
Synchrogen shows a running character-count budget against each gap as you write, keeps dialogue locked so a line can't accidentally overlap it, and lets you switch between script and timeline view on the same project. The describer still writes and approves every word — the editor's job is just making the constraints visible while you write. Currently in closed beta.
FAQ
How long should an audio description line be?
As long as the available gap allows and no longer — there's no fixed word count. At natural speaking pace, roughly two to three words fit per second of silence, so a three-second gap holds about eight to ten words.
Should audio description be written in present or past tense?
Present tense, almost always. It matches the immediacy of action happening on screen in real time and generally reads faster than past tense at the same word count.
What shouldn't be included in an audio description script?
Emotional interpretation (“she looks sad” instead of describing the visual cue), anything already conveyed by dialogue or sound, and detail that doesn't affect plot, character, or setting — all three eat gap space that's usually better spent on what a sighted viewer would actually need to follow the story.