# Transcripts, captions and accessibility: the needs to cover

A transcript gives someone a way to read the recording without listening. Captions provide synchronized text while watching. A descriptive transcript also communicates essential visual information. These representations overlap, but none proves that the whole page or media experience is accessible.

## One scene, information in two channels

Consider this original teaching scene for a generic file-export tutorial. It is a written storyboard, not a captured VidSEO interface or a claim about a tested recording.

<div class="hub-diagram">
<section><h3>Spoken</h3><p>“Select this option to keep the original, then export.”</p><p>The speaker refers to an option without naming it.</p></section>
<section><h3>Visible</h3><p>The pointer selects <strong>Keep original file</strong>. A notice reads <strong>The destination folder must be writable</strong>.</p><p>Neither label is spoken.</p></section>
<section><h3>Needed to act</h3><p>The reader needs the option's name and the folder requirement, as well as the spoken sequence.</p><p>Transcribing speech alone leaves the instruction incomplete.</p></section>
</div>

## Compare the representations

**Speech-only transcript:** “Select this option to keep the original, then export.” This is faithful to the words, but a person who cannot see the pointer still does not know which option to select.

**Captions:** the same words appear at the time they are spoken, along with relevant sounds and speaker identification when needed. Captions help a viewer follow the audio while watching; the presence of a timed file does not prove that the player exposes it correctly.

**Descriptive transcript:** “Select this option to keep the original, then export. [Visual information: the pointer selects ‘Keep original file’. A notice says the destination folder must be writable.]” The brackets make the added visual description distinct from the speech. They do not imply the speaker said those words.

In a real recording, verify the added description against the image. Do not fill a visual gap with a plausible instruction. If a correction or broader explanation is needed, place it in an editorial note outside the faithful transcript.

## Review the text as a content owner

Listen and watch with the transcript open. Check names, numbers, negations, warnings and changes of speaker. A missed “not” can reverse an instruction even when the paragraph reads fluently. Identify meaningful sounds, such as an error alert, when they contribute information.

Then review the recording without relying on its audio: which essential labels, diagrams, gestures or status changes remain unrepresented in the text? Add concise descriptions at the relevant points. Decorative motion does not need the same treatment as a warning or a demonstrated action.

Finally, read only the text. Can someone understand the sequence and find the result? Use paragraphs and headings, but preserve the distinction between the recording and your additional guidance. Our [page-editing guide](/blog/how-to-turn-a-video-transcript-into-a-publishable-page/) shows that editorial separation.

## Check the player and the page

Test the actual captions control, language selection, timing and readability. Check keyboard access to playback and chapter actions, visible focus, and text wrapping on a narrow screen. A downloadable SRT file is not the same thing as a working captions track. A visible transcript hidden behind an inaccessible control can also fail its purpose.

In VidSEO, [TXT, SRT and VTT imports](/guides/txt-srt-vtt-html-transcripts/) prepare transcript content. Caption-track and speaker-profile editing has the edition boundary described in [Free versus Pro](/free-vs-pro/). Supplying a track does not change the captions inside YouTube's or Vimeo's own player. Verify the actual publishing arrangement instead of assuming that one upload covers every player.

## State the result precisely

The [WordPress comparison](/blog/before-and-after-a-video-page-with-and-without-an-exposed-transcript/) demonstrates public HTML text. It is not a WCAG conformance audit. Depending on the media and audience, other needs can include audio description, accessible player controls and further page-level checks. Record what was reviewed and what still needs a specialist assessment.

**Next step:** choose a short passage from your own video, identify one spoken and one visual piece of information, then check whether each reaches someone using the transcript or captions.

## References

- [W3C WAI: transcripts, including descriptive transcripts](https://www.w3.org/WAI/media/av/transcripts/).
- [VidSEO documentation](/docs/).

[Browse all resources](/resources/).
