Describe the visual information a listener needs to understand the film, and place that description where it does not obscure essential dialogue or sound. If the approved edit has no suitable gaps, prepare a separately described version or an accompanying descriptive transcript. Speaking faster over the vows is not a useful accessibility solution.
W3C's audio-description guidance explains the purpose and delivery choices. Start from a relative's information needs, not a desire to narrate every visible detail.
Find the information missing from sound
Listen to the film without watching it. Note moments where the story changes but the soundtrack does not explain the change: a box opens, someone reveals a ring, a date appears, or the couple moves into a new setting.
Separate necessary information from decorative detail. A brief description of a ring being offered may matter; an inventory of every flower may not. Identify people consistently when that is necessary to understand who acts or responds.
For a fictional AI sequence, describe what the scene shows. Do not present an imagined journey as a real event from the couple's history simply because the image looks photographic.
Mark usable gaps against the actual edit
Make a cue list with the visual event, the shortest useful description, the available gap and any important sound already occupying it. Listen to the entire gap, including the end of a sentence and its emotional pause.
Silence can carry a reaction. Filling every pause changes the scene even when no words overlap. Preserve a meaningful laugh, an intake of breath or the original proposal question when that sound helps explain the event.
Captions address another access need: W3C's caption guidance covers meaningful spoken and non-spoken audio in text. A described soundtrack does not replace captions for a relative who cannot hear it.
Keep the wording observable
Prefer actions, relationships and essential on-screen text. “She opens the small box and holds out a ring” describes visible evidence. “She realizes he is her destiny” interprets an internal state the picture cannot establish.
Use a distinguishable narration voice with the speakers' permission when real recordings are involved. Avoid imitating a family member in a way that makes new description sound like an original ceremony recording.
Check names and pronouns in context. A listener should not need to see a costume color to know which partner is speaking or acting.
A fictional ring-reveal description plan
In an invented proposal film, a quiet shot shows the lead taking a closed box from a coat, opening it and looking toward the partner. The planned description is “He opens a ring box.” It belongs after the action becomes clear and before the spoken question, provided the actual cut has room.
If the question begins immediately, the proposed described version adds an appropriate pause through external editing or supplies that information in a separate transcript. This is a planning example, not a tested generated clip. The description must follow the delivered action if the shot changes.
Approve access to the story
Ask a listener who has not watched the picture to describe who offered the ring, what the response was and which parts were imagined. Their answers reveal information gaps that a technically audible narration may still leave.
Review the full mix for overlaps, rushed wording and unnecessary interpretation. Label the described version clearly and keep the clean master. Confirm that the chosen player can deliver the intended version.
Develop the visual story in Imagild Film Studio. Description writing, authorized recording and mixing require an external finishing workflow unless the currently available controls explicitly provide them.