Write audio description for a mostly visual scene

Gemini: logs the visuals and drafts the script

Gemini makes the visual log and a first draft. You fit lines to pauses; a blind or low-vision reviewer listens.

You
Get ready
Gemini
Log and draft
You
Check against pauses
The reviewer
Reviewer listens

Have ready: the 3-minute scene video
Use: You
The 3-minute scene video.
Check
Upload only if the owner allows the scene in an online tool.
Log the visuals and draft the descriptions
Use: Gemini
Upload the scene for a time-stamped log of visible actions, the natural pauses with times, and one draft description per pause.
Details
What to paste into the tool
Task: Visual log, pause list and draft script. Watch this scene and list what is visible on screen by time; describe actions, people, objects and text only; do not state feelings, motives or story meaning. Mark the natural pauses with start and end times. Write one description per pause, short enough to be spoken inside it, and add a short style note.
Read each draft line aloud against its pause
Use: You
Read each draft description aloud against its pause and fix lines that overrun or add interpretation.
Check
Check each line fits its pause read aloud and describes only what is shown on screen.
Listen and mark unclear lines
Use: The reviewer
A blind or low-vision reviewer listens and marks any unclear or interpretive description.
Check
Describe only what is shown on screen; the reviewer must confirm clarity.

ChatGPT: turns your scene note into the script

You write the scene note and pauses. ChatGPT drafts the lines; you and a reviewer check them.

You
Watch and note the scene
ChatGPT
Draft the descriptions
You
Fix against the scene
The reviewer
Reviewer listens

Watch the scene and mark each pause
Use: You
Watch the scene, write a short plain note of what happens with times, and mark each natural pause with start and end times.
Draft a description for each pause
Use: ChatGPT
Paste the scene note and pause list; ask for one description per pause, short enough to fit the pause and free of interpretation.
Fix each line against the scene
Use: You
Play the scene with the draft and fix any line that is wrong, too long for its pause, or adds opinion.
Check
Read each line aloud and time it against its pause; describe only what is shown, no interpretation.
Listen and mark unclear lines
Use: The reviewer
A blind or low-vision reviewer listens and marks any unclear or interpretive description.
Check
Describe only what is shown on screen; the reviewer must confirm clarity.

Compare the two approaches

Option A: Gemini: logs the visuals and drafts the script

Best when: You want a first draft to correct, not a blank page

What it does: Gemini drafts; you fit lines, the reviewer listens

Strong at: 3-minute scene fits the 5-minute base total video length on the page

Limit: Fitting lines to pauses unconfirmed; check each line aloud

Free plan: Yes

Option B: ChatGPT: turns your scene note into the script

Best when: You can write a short note of what happens

What it does: ChatGPT drafts lines; you and the reviewer check

Strong at: No video leaves your computer; only text is pasted

Limit: ChatGPT never sees the scene; check every line

Free plan: Yes

Write audio description for a mostly visual scene

An audio description script with insertion times in natural pauses and a style note.

Gemini: logs the visuals and drafts the script

Have ready: the 3-minute scene video

Log the visuals and draft the descriptions

Read each draft line aloud against its pause

Listen and mark unclear lines

ChatGPT: turns your scene note into the script

Watch the scene and mark each pause

Draft a description for each pause

Fix each line against the scene

Listen and mark unclear lines

Compare the two approaches

Option A: Gemini: logs the visuals and drafts the script

Option B: ChatGPT: turns your scene note into the script

Tools in this workflow

Final check

Write audio description for a mostly visual scene

An audio description script with insertion times in natural pauses and a style note.

Gemini: logs the visuals and drafts the script

Have ready: the 3-minute scene video

Log the visuals and draft the descriptions

Read each draft line aloud against its pause

Listen and mark unclear lines

ChatGPT: turns your scene note into the script

Watch the scene and mark each pause

Draft a description for each pause

Fix each line against the scene

Listen and mark unclear lines

Compare the two approaches

Option A: Gemini: logs the visuals and drafts the script

Option B: ChatGPT: turns your scene note into the script

Tools in this workflow

Final check

Share this page