Subtitle accessibility guide

By Subtitle Convert · 8 minute read · Updated for 2025

Subtitles are not just a convenience for people watching foreign-language content. They are a primary access channel for deaf and hard-of-hearing viewers, and an increasingly important tool for neurodivergent audiences, language learners and anyone watching video in a noisy environment or on mute. This guide covers what accessible subtitles need.

Who subtitles serve

Deaf and hard-of-hearing viewers — the primary audience for captions.
Non-native speakers — language learners and immigrants consuming content in a second language.
Viewers in noisy or quiet environments — airports, public transit, shared offices, late-night viewing.
Neurodivergent audiences — viewers with ADHD, auditory processing differences or autism may process text better than audio.
Anyone watching at speed — captions help viewers follow when they speed up playback.

Types of accessible subtitles

Closed captions (CC / SDH)

Subtitles for the deaf and hard-of-hearing (SDH) include not just dialogue but also:

Sound effects in square brackets: [door slams], [phone rings].
Speaker identification: - MARY: Hello, John..
Music cues: [soft piano], ♪ music ♪.
Off-screen dialogue labelled as such.

SDH subtitles are the highest standard for accessibility, because a viewer who cannot hear anything can still follow the whole story.

Open captions vs closed captions

Open captions are burned into the video. They cannot be turned off. Used in cinemas (for special showings) and some broadcast contexts.

Closed captions are delivered as a separate stream or file. Viewers can turn them on or off. Standard for streaming and web video.

Accessibility standards and rules

WCAG 2.2 Level AA requires captions for pre-recorded audio content in synchronized media (1.2.2). Level AAA adds extended audio description and sign language interpretations for live content.
Section 508 in the US requires captions for federal agencies.
ADA is interpreted to require accessible video content for public accommodation.
EAA (European Accessibility Act) applies to many industries in the EU from 2025.
Broadcast TV has mandatory caption requirements in many countries.

Design considerations for accessible subtitles

Reading speed. SDH content should target 17 characters per second or less. This is slower than the general-audience target (20 CPS) because SDH content includes sound effects and speaker labels that add reading load.
Line length. Keep to 42 characters per line, maximum two lines. For SDH with speaker labels, this may require more compressed dialogue.
Contrast. White text with a thin black outline or a semi-transparent background is standard. Pure white on bright scenes is unreadable.
Positioning. Keep subtitles away from faces and important visual content. WebVTT allows per-cue positioning, which is useful for complex shots.
Font. Sans-serif, medium weight, consistent size. Proportional fonts are usually better than monospace for readability.
No flashing. Subtitles that appear/disappear rapidly can trigger seizures in some viewers. This is rare in practice but worth knowing.

Producing SDH subtitles

Start with a transcript of the audio, not just the dialogue.
Identify which sounds matter for comprehension (gunshot, alarm, footsteps) and add them in brackets.
Label speakers when it is not clear from context.
Mark music and lyrics — the convention is ♪ lyrics ♪ or [soft piano music].
Timing: give SDH cues slightly more lead-out time (200–300ms) to accommodate the extra reading.
Compress dialogue where needed to keep total CPS within the SDH target.

Tools for accessibility subtitles

Line Length Checker — flags lines that are too long.
Duration Report — shows cue durations that may be too short for reading.
SRT Validator — catches timing issues.
SRT to VTT — converts SRT to WebVTT for HTML5 video captions.

Common accessibility failures

Fast dialogue not compressed — SDH viewers cannot keep up.
Sound effects missing — a deaf viewer loses important story information.
Speaker confusion — multiple characters not labelled, dialogue becomes ambiguous.
Subtitles on bright backgrounds without outline or background.
Inconsistent positioning — subtitles jump around the screen for no reason.
Auto-generated captions uncorrected — machine ASR output is often wrong and can actively mislead.