Step 1 — find the gaps
Before anything gets written, you need every silent gap between lines of dialogue mapped out — that's where description has to live. Doing this by ear, scrubbing the timeline manually, is the slowest part of a first attempt; automatic scene and dialogue detection collapses it into minutes.
Step 2 — write to the gap, not around it
Each gap sets a hard word budget — roughly two to three words per second at natural pace. Write the description to fit that specific gap rather than writing a paragraph and cutting it down afterward; it's faster and the result reads more naturally.
Step 3 — voice it
Record narration yourself, hire a voice talent, or generate it with text-to-speech. Whichever route, keep the same voice consistent across the whole project — a narrator that changes tone or pacing between clips is one of the fastest ways to break immersion for a viewer relying on the track.
Step 4 — mix or keep separate
Description can be burned into the final mix, or delivered as a standalone AD-only audio track the viewer switches on. Increasingly, platforms and broadcasters want both — confirm the delivery spec before you commit to only producing one.
Step 5 — export to the format the platform actually requires
Different platforms and broadcasters specify different container formats, track layouts, and loudness standards for the AD track. Check the actual delivery spec before export — a technically correct track in the wrong format still gets rejected.
The shortcut that isn't
Skipping step 1 and writing description “by feel” while watching is where most first attempts go wrong — lines end up overlapping dialogue, or missing gaps that were there the whole time. The gap-finding step is tedious, but it's the one step that can't be shortcut without the rest of the process getting harder.
Where Synchrogen fits
Synchrogen automates steps 1 and 3 — scene/dialogue detection and narration — and keeps step 2, the actual writing, and the step 4/5 export decisions in the describer's hands, with both burned-in and AD-only exports built in. Currently in closed beta.
FAQ
Can I add audio description to a video myself without professional software?
Yes, with a general audio editor and enough manual time to find gaps, write, and voice each line by hand. Purpose-built software mainly saves time on the gap-finding and mixing steps, not on the writing itself.
Do I need a separate audio track for audio description or can I mix it in?
It depends on the delivery spec. Many broadcasters and platforms want a standalone AD-only track the viewer can toggle, in addition to or instead of a burned-in mix — check the specific requirement before producing only one.
What file format does audio description need to be delivered in?
It varies by platform and broadcaster — there's no single universal standard. Always confirm the target platform's specific container format, track layout, and loudness requirements before final export.