Publish accessible, useful video on a website

Define the role of the video

Write down what it adds: demonstration, answer, place introduction or experience. Place it on a page whose text explains the subject and next step. A generic clip unrelated to the surrounding content does not help.

Plan length, sequence, filmed people, music rights and review date. When a product or process changes often, a short maintainable clip is more useful than a long film that becomes obsolete.

• Purpose and page chosen.
• Rights and people checked.
• Length fits the question.

Make speech and visuals accessible

Prepare synchronized captions covering useful speech and sounds. Fictional example: a demonstration says “fifteen” but automatic captions show “fifty”; check numbers, names and technical terms before publication. Add a structured transcript for locating steps. When a visual action conveys information absent from speech, also plan how to make it understandable without seeing the image.

If an important action is shown only visually, explain it in narration or provide suitable description. W3C caption guidance explains the relationship between captions and transcripts.

For prerecorded video with sound, distinguish the needs covered. Accessibility captions follow speech while identifying speakers and sounds needed for understanding. Translating only dialogue may therefore omit important audio information. Check names, numbers, timing and line placement; the existence of a caption file does not prove its accuracy.

Audio description addresses another need: making important visual information understandable through sound. Under WCAG criterion 1.2.5 at Level AA, a transcript alone does not replace description when it is needed. If the main narration already conveys all important visual information, no additional description is necessary. Plan that choice in the script, then check the player and the version actually published.

• Captions reviewed and timed.
• Essential visual information described.
• Useful text version available.

Scroll the table horizontally to read every column. With a keyboard, focus the table area and use the arrow keys.

• W3C WAI : Captions/Subtitles

  Consulted on 22 October 2026; speech, speaker identification and useful sound information.

• W3C WCAG 2.2 : Audio Description (Prerecorded)

  Consulted on 22 October 2026; scope of criterion 1.2.5 and cases where audio already conveys visual information.

Choose a restrained player and thumbnail

Test play, pause, volume, captions, full screen and keyboard focus. Avoid automatic sound. An original thumbnail should represent the real content and have suitable alternative text if informative.

Load media according to its role rather than slowing every visit. Check file weight and mobile-network behaviour. The performance guide helps measure the effect.

Check the format actually loaded by the player. A WebVTT file is UTF-8 encoded and starts with the WEBVTT signature, using the required line structure. Fictional exercise: a cue from 00:00:01.000 to 00:00:03.500 spans 2 500 milliseconds. The decimal point belongs to timestamp syntax: do not replace it with a comma when translating the file. End time must follow start time, and cue starts follow the format’s ordering rules. Syntax alone proves neither readability nor synchronisation with speech.

Check the selected track, its language and its kind in the final player. srclang describes track language; label helps people choose it. A captions track must convey necessary audio information, not merely translate dialogue. Test pause, track switching and phone display, then review names, numbers and meaningful sounds with the video. The 20 May 2026 WebVTT Candidate Recommendation Draft describes the format; it replaces neither editorial checking nor other accessibility needs.

• Controls keyboard-usable.
• No imposed sound.
• Loading and thumbnail checked.

Scroll the table horizontally to read every column. With a keyboard, focus the table area and use the arrow keys.

Check discovery and updates

Align the title, text, thumbnail and video with the same content. Fictional example: a recorded procedure becomes outdated after an interface change; state its scope and correct passages that could cause a wrong action. Check the player, captions and preview after replacement. Google’s video best practices support discovery checks without guaranteeing indexing.

When content ages, replace the clip or explain its earlier context. Add video structured data only when it accurately describes a media item actually present. Translated versions need their own checks of captions and page copy.

• Page and media consistent.
• Thumbnail and URL checked.
• Languages and captions reviewed.