How to write a style guide for captions in educational videos

Captions are part of the learning experience, not an afterthought added when a video is ready to publish. A well-designed captioning system gives Deaf and hard-of-hearing learners access to spoken content, while also helping students who speak English as an additional language, study in noisy spaces or review complex material at their own pace.

A caption style guide turns accessibility goals into consistent editorial decisions. It explains how captions should represent speech, identify speakers, describe meaningful sound and handle terminology. Without shared rules, one lesson may use full stops and speaker labels while the next relies on fragments, unexplained abbreviations and poorly timed text.

For Australian schools, universities, training providers and education publishers, the guide should reflect local language and culture. That includes Australian spelling, Auslan awareness, Indigenous names and terminology, and the expectations of learners using platforms such as Canvas, Moodle, YouTube or institution-specific learning management systems.

The strongest guidance is practical enough for a captioner, teacher, video editor and quality reviewer to use in the same way. It should also be created with Deaf people and accessibility specialists, whose experience can reveal problems that a purely technical checklist may miss. Christopher Tester’s accessibility practice provides useful context for organisations considering communication access across education and public-facing media.

Define the purpose and audience

Begin by stating who the captions are for and what they need to achieve. Captions should provide equivalent access to dialogue, important non-speech information and the structure of the lesson. They are different from subtitles that translate one spoken language into another, although a single video may require both captions and translated subtitles.

Identify the educational settings covered by the guide. A university lecture, a primary science animation and a vocational safety demonstration may need different levels of detail. A lesson watched independently may benefit from fuller descriptions, while a live-recorded tutorial may require careful treatment of questions, interruptions and audience responses.

Set an accessibility baseline using recognised standards, such as WCAG guidance on prerecorded media. In Australia, organisations should also consider obligations under the Disability Discrimination Act 1992 and the practical expectations of students, families and education providers. A style guide cannot replace legal advice, but it can establish a clear, repeatable process for meeting access responsibilities.

Include the intended delivery formats. Captions embedded in a video behave differently from a selectable WebVTT track, and mobile screens may display fewer characters than a desktop player. The guide should tell production teams whether captions must be provided as an editable file, a platform-ready file, an open-captioned video or all three.

Establish rules for readable caption text

Choose a readable presentation style before production begins. Use Australian English spelling and punctuation unless the subject requires another convention. Write captions in concise phrases that preserve the speaker’s meaning, rather than editing speech so heavily that important nuance disappears. Avoid placing too many words on screen at once, and keep line breaks at natural grammatical boundaries.

A style guide should specify maximum lines, preferred characters per line and minimum display duration. Many teams use two lines as a practical maximum and aim for a comfortable reading speed, but the exact settings should be tested with real learners. Young students, readers with language-processing differences and people viewing on a phone may need more time than a fast adult reader.

Decide how to handle hesitations, repetitions and false starts. Include them when they affect meaning, emotion or the lesson itself; remove distracting verbal clutter when it adds no value. For example, a repeated word may show uncertainty in a role-play, while repeated “um” sounds in a long lecture may make captions needlessly difficult to follow.

Preserve essential formatting cues. Italics can indicate a voice heard through a recording, while bold text may be reserved for defined educational purposes rather than used decoratively. Avoid capitalising whole sentences for emphasis, since all-capital text is harder to read. Define how mathematical notation, formulas, URLs, book titles and unfamiliar terms should appear, and ensure the rules work in the institution’s chosen captioning software.

Represent speakers, sound and visual information

Speaker identification is essential when more than one person appears or speaks. The guide might require a name label, such as “Mia:”, when the speaker is visible, and a descriptive label such as “[Lecturer]” when the person is off screen. Use labels consistently and avoid relying only on caption colour, because colour may be unavailable, unclear or inaccessible to some viewers.

Describe sounds that carry educational, emotional or contextual meaning. Suitable examples include “[door closes]”, “[students laugh]”, “[alarm sounds]” and “[music fades in]”. Do not caption every minor background noise. The test is whether removing the sound description would make the lesson, scene or speaker’s intention harder to understand.

Visual information should not be duplicated unnecessarily. If a diagram visibly displays the words “water cycle”, repeating that text in the captions may be useful when the presenter refers to it, but a caption track should not become a complete description of every image. Important visual content that is not spoken should be supplied through the video’s narration, audio description or other accessible learning materials.

Take particular care with classroom discussion and recorded webinars. Identify students when they speak, distinguish overlapping voices where possible and mark unintelligible audio honestly rather than guessing. A note such as “[unclear]” is preferable to an invented statement that could misrepresent a learner. If several people speak at once, preserve the main educational contribution and explain the overlap according to the guide.

Include language, culture and specialist terminology

Educational captioning must account for language diversity. Create a process for checking names, place names, technical vocabulary, acronyms and words from languages other than English. In Australia, this may include Aboriginal and Torres Strait Islander language names, local place names and discipline-specific terms used in health, law or environmental studies.

Do not treat Auslan as a visual version of spoken English. Auslan has its own grammar, vocabulary and Deaf cultural context. Captions provide access to spoken or recorded content, but they do not replace a qualified Auslan interpreter when a learner needs signed communication. Where an interpreter appears on screen, the guide should explain whether the interpreter’s signing is captioned, identified or supported by a separate transcript.

Cultural sensitivity belongs in the editorial rules as well as the training process. Teams should understand how to refer respectfully to Deaf people, hard-of-hearing people and disability-related topics. Deaf culture guidance can help hearing colleagues recognise why communication choices, eye contact, turn-taking and access to information matter.

Build an approved terminology list for each course or organisation. Include the preferred spelling, pronunciation notes for voice talent, abbreviation rules and definitions for specialist words. In a Sydney nursing course, for example, a medication name must be checked against the teaching materials; in a Melbourne engineering course, a technical abbreviation should not be expanded inconsistently across modules.

Create a quality assurance workflow

Assign responsibility at every stage. The scriptwriter or educator checks subject accuracy, the captioner prepares the timed text, the editor reviews synchronisation and presentation, and an accessibility reviewer assesses the learner experience. A Deaf reviewer should be involved where possible, especially for content about Deaf communities, signing, disability or communication access.

Review synchronisation as well as spelling. Captions should appear close to the relevant speech, remain long enough to read and clear before the next idea begins. They should not cover a presenter’s face, key diagrams, demonstrations or on-screen text. Test the final video on a laptop, mobile phone and the actual learning platform, since player controls and line wrapping can change the result.

Keep a reusable checklist with measurable checks. It can cover caption file format, speaker identification, sound descriptions, terminology, reading speed, line breaks, punctuation, visual placement and accuracy. Sample several minutes from the beginning, middle and end of every video, and review the whole item when the content is high-risk, assessment-related or technically dense.

Finally, version the style guide and record exceptions. Educational teams often work with external captioning vendors, freelance editors and lecturers who produce their own recordings. A shared document should state which rules are mandatory, who approves deviations and where the latest terminology list is stored. Regular feedback from students can then lead to specific updates, rather than informal changes that make future videos inconsistent.

A caption style guide becomes effective when it is treated as part of teaching quality. Clear standards, local cultural awareness, reliable technology and meaningful Deaf participation can produce captions that support comprehension while respecting the people who depend on them.