Frequently Asked Question
Captioning Criteria
I would remove the ahs and umms...as the machine-generated captions generally do that.
If there's more than one speaker, the speaker's name needs to be identified before each person's text.
The captions can include a parenthetical aside that notes there's crosstalk in the background. This is an important component of audio description.
Captions should be dramatically correct, and that can be tricky since we talk differently from how we write. So, I would use best judgment there
Dialogue: speakers must always be captioned. Identify speakers and tones when they cannot be inferred; if they are talking over each other and there is no clear dominant speaker, you put in parenthesis that multiple people are speaking at once. b. Caption verbal/oral bridges (e.g. “um” and “uh”) only if they hold value (this is a rare thing). c. Add proper grammar to run on sentences to help convey meaning.