Decide what your audience needs
Plan the language, relevant sound information, and delivery format before creating the track. Labels alone do not establish accessibility.
By ToolBrivio editorial · Research approach · Link disclosure
We currently earn no commission from product links on this page.
A client asks for subtitles, but the video also includes an off-screen speaker, a ringing phone, and music that changes the mood. Should the text contain only dialogue, or those details too?
The answer starts with the viewer’s needs. Translation, access to important audio information, and a separate reading copy are related goals, but they call for different deliverables.
This guide explains the distinctions with practical examples. Platform labels are not always consistent, so use the contents of the requested file—and the experience you want to provide—to decide what belongs in it.
Captions describe the audio
Captions include speech and meaningful sounds. A speaker label, a phone ringing, or music starting can help someone follow a scene without hearing it. Closed captions can be switched on or off; open captions remain visible in the video image.
Subtitles focus on dialogue
Subtitles commonly translate speech into another language. Terminology varies between platforms, and “subtitles” is sometimes used for same-language speech text too. Look at what the file contains rather than relying on its label.
A transcript is a separate reading format
A transcript presents the recording as text. It can help with searching, reading, and repurposing an interview, but a plain transcript does not replace timed captions in a video.
Choose for the viewer
- For viewers watching without sound, include the important audio information.
- For viewers who speak another language, provide a checked translation.
- For a readable record, offer a transcript alongside the recording.
Automatic text is a first draft. Check names, technical terms, speaker changes, and timing before publishing.
One scene, three different outputs
Imagine an interview where the guest says, “That was the turning point,” while a phone rings off screen. A speech-only subtitle may contain the sentence. Captions may also need to identify the ringing if it explains why the guest pauses. A reading transcript presents the discussion in a document rather than as timed on-screen blocks.
This is an illustrative scene. Include a sound description when it contributes to understanding; listing every background noise can distract from the conversation. Translation also needs a review in the target language, rather than assuming the source-language wording settles every ambiguity.
Clarify the request before you start
Ask three things: which audience will use the output, where it will be published, and whether the text must be optional in the player. Then agree whether the deliverable includes sound information, translation, or a separate reading copy. Those answers determine the work more reliably than the word “subtitles” alone.