How to Evaluate Caption Text Contrast
A representative-frame audit combining caption text contrast evidence, background treatments, motion, size, placement, and viewer readability.
In short: Test captions over bright, dark, detailed, moving, and mixed-luminance scenes at the real viewing distance. Record text colour, outline, shadow, and background treatment. Use contrast-ratio evidence where the foreground and effective background can be identified, then confirm readability during motion. One paused frame cannot represent an entire title.
Video changes behind the text. A caption that separates clearly from one dark frame may disappear over a bright sky, patterned clothing, or rapid movement seconds later.
Stabilise the presentation
Record device, display, room lighting, viewing distance, caption size, font, text colour, outline or shadow, background colour and opacity, placement, and track.
Keep those settings fixed during the first scene comparison. Otherwise, the audit cannot identify which treatment changed the result.
Build a representative frame set
Collect six authorised screenshots or timestamps:
- bright, low-detail frame;
- dark, low-detail frame;
- bright, detailed frame;
- dark, detailed frame;
- mixed bright and dark regions behind one cue;
- motion crossing the caption region.
Redact unrelated account or profile information.
Understand contrast evidence
WCAG's contrast guidance explains luminance-based ratios for text and backgrounds, including thresholds for normal and large text. Applying a single ratio to video captions can be difficult when pixels behind the text vary continuously.
When an opaque or stable background exists, measure text against that effective background. When the background is transparent, sample the worst relevant regions and document the method. Do not claim exact conformance from an informal screenshot alone.
Original evidence: dynamic frame set
| Frame | Background type | Text treatment | Ratio evidence | Motion result | Viewer result |
|---|---|---|---|---|---|
| Bright | Low detail | Text/outline/box | Measured or not measurable | Replay result | Readable/issue |
| Dark | Low detail | Treatment | Evidence | Result | Result |
| Detailed | Mixed | Treatment | Range/uncertain | Result | Result |
| Moving | Variable | Treatment | Not a single value | Result | Result |
Record the tool and colour samples when calculating a ratio.
Replay at actual speed
Pause supports measurement, but playback reveals whether background motion, short display time, or visual competition makes cues hard to follow. Ask the viewer to read several cues without leaning forward or replaying.
Avoid using the evaluator's vision as the only outcome.
Test background treatments
A solid or translucent box can create a more stable effective background. An outline or shadow can help separate letter edges while preserving more of the image. Results depend on size, thickness, colour, opacity, and scene.
Use the caption background guide for the legibility-versus-occlusion tradeoff.
Keep size separate
Larger text may be easier to recognise but does not automatically produce adequate contrast. It may also cover more variable video. Test size with the readable caption-size method while holding colour treatments steady.
The complete caption accessibility guide connects contrast with timing, content, controls, and placement.
Prioritise failures
Treat unreadable text on common representative frames as a high-impact barrier. Also flag a control that exposes colour choices but no reliably readable combination. Record whether the limitation belongs to the supplied styling, current player controls, or an unknown layer.
Report contrast barriers
Include frame set, settings, device, distance, room context, measurement method, viewer task, expected result, and observed result. Do not attach media or expose private history, credentials, or source addresses.
Common mistakes and limitations
Avoid one-frame conclusions, measuring against the wrong background, changing size during a contrast test, and claiming formal conformance from an informal colour picker.
Current player customisation and supplied styling determine available treatments. Verify official controls instead of promising them.
Recheck during motion
A paused frame can make text look clearer than it is during camera movement or rapidly changing backgrounds. Replay the same cue at normal speed and record whether the outline, shadow, or background treatment remains effective throughout.
Frequently asked questions
Is white text always readable on video?
No. Bright or detailed backgrounds can reduce separation; test representative scenes and supported edge or background treatments.
Is a numeric ratio enough?
No. Use it as evidence where measurable, then test motion, timing, viewing distance, and real reading outcomes.
Does a background box solve every contrast issue?
Not automatically. Its colour and opacity must separate text while avoiding harmful visual occlusion.