What Is a Video Transcript?
What a transcript contains, the forms it comes in, how it differs from captions, and what people actually use one for.
The Scriptorium team ·
A video transcript is a written record of the speech in a video. At minimum it is the words. Usually it also carries timestamps, so each line can be traced back to the moment it was said, and speaker labels when more than one person talks.
What is in one
- The words. Everything spoken, in order.
- Timestamps. Usually per line or paragraph; some formats carry a start and end time for every individual word.
- Speaker labels. Who said each part, when there is more than one voice.
- Punctuation and capitalisation. Added by the transcription process, since nobody speaks punctuation aloud.
What is not in one: anything that happened on screen. A transcript of a software demonstration will not tell you what was clicked. That is what an audio description is for, and it is a separate document.
Verbatim and clean transcripts
Two conventions, for two purposes.
| Verbatim | Clean | |
|---|---|---|
| Filler words | Kept | Removed |
| False starts and repetition | Kept | Removed |
| Stutters | Kept | Removed |
| Used for | Research, legal, speech analysis | Publishing, journalism, notes |
There is also true verbatim, which marks non-speech sounds and pauses as well. It is used in conversation analysis and legal work and is hard to read for any other purpose.
The rule that applies to both: you may remove disfluencies, and you may not change what someone said. Tidying "I think maybe we should probably look at it" into "we should look at it" is a different statement, not a cleaner one.
How it differs from captions
A transcript is a document you read on its own. Captions are short chunks of timed text shown over the video, and they also describe non-speech sound for viewers who cannot hear. Captions are generated from a transcript, not the other way round.
The full comparison covers subtitles and closed captions as well, which are two more things again.
What people use them for
- Accessibility. Deaf and hard-of-hearing viewers, and anyone using a screen reader.
- Searching. Finding the part of a long recording where something was discussed.
- Quoting. With a timestamp to verify against.
- Repurposing. Articles, show notes, summaries, social posts.
- Making captions. The transcript is the source.
- Reading instead of watching. Many people read faster than they listen, and some simply prefer it.
- Compliance. Many organisations are required to provide text alternatives for recorded material.
The formats it comes in
| Format | Contains |
|---|---|
| Plain text (.txt) | Just the words. |
| Timestamped text | Words with a time against each paragraph. |
| Markdown (.md) | Formatted, with speakers and structure. |
| SRT / WebVTT | Subtitle formats, split into display chunks. |
| JSON | Words, timings and speakers as structured data. |
| CSV | One row per segment, for spreadsheets. |
Getting one
Upload the video to a transcription tool. video to text does it directly, free for a first file up to 15 minutes with no account. If the video is on YouTube, the YouTube transcript page covers the free route.