Captions Quality Guide

Standards-aligned reference for writing and reviewing closed captions (CC) and subtitles. Covers WCAG 1.2, FCC rules, Netflix timed text spec, and BBC subtitle guidelines.

Core Standards

Spec	Requirement
WCAG 1.2.2	Captions required for all prerecorded audio in video (Level A)
WCAG 1.2.4	Captions required for all live audio in video (Level AA)
FCC accuracy	≥99% word accuracy for live captions; 100% target for offline
Reading speed	Max 180–200 WPM on screen; 130 WPM for children's content
Line length	Max 32–42 characters per line (2 lines max on screen at once)
Display duration	Min 1 second, max 7 seconds per caption event
Timing	Sync within 2 frames of speech; never before audio begins

Speaker Identification

When the speaker is not visible on screen, or when multiple speakers are present, identify the speaker.

Good

[NARRATOR] In 1969, humanity reached the moon.
[DR. CHEN] We need to run the test again.

Also acceptable

DR. CHEN: We need to run the test again.

Bad

We need to run the test again.
(No speaker ID when off-screen)

Sound Effects & Music

Non-speech audio that conveys meaning must be captioned using square brackets or italics depending on format.

Good examples

[Glass shattering]
[Tense music]
[Doorbell rings]
[EXPLOSION]
♪ Upbeat pop music playing ♪

Omit if

[Keyboard typing] [Mouse clicking] — ambient sounds that don't affect meaning

Punctuation & Formatting

Key rules

• Use em dash (—) for abrupt speech cutoffs
• Use ellipsis (…) for trailing/uncertain speech
• Italicize off-screen narration (in SRT/VTT use <i>)
• Capitalize the first word of each caption event
• Don't break sentences across lines mid-clause — break at natural pauses