Style Guide References
Rev’s transcribers, captioners and subtitlers follow shared style guidelines so every file is accurate, consistent and easy to read. This article explains what those guidelines mean for the files you receive: how speech is written, how speakers and sounds are labeled, how captions and subtitles are formatted, and how you can help us get names and terms right.
In this article
- Transcripts
- Captions
- Premium Captions
- Subtitles
- Glossaries, Speaker Names and Scripts
- How We Check Quality
Transcripts
Accuracy
- We type what’s said, in the speaker’s own words. We don’t paraphrase, add content, or guess what someone meant to say.
- Profanity is typed as spoken. Words that are bleeped or muted in the recording are marked (beep) or (censored).
- A clear mispronunciation may be corrected when it would get in the way of understanding. For example, “warshar” is typed as “washer.”
- Names, companies and industry terms are researched. If a spelling can’t be confirmed, the transcriber uses their best guess and keeps it consistent. Terms in your glossary are always used.
- Transcripts use standard US spelling and American English grammar and punctuation.
Standard and Verbatim Transcripts
Standard (the default) is lightly edited for readability without changing meaning. It leaves out:
- Filler words such as um, uh, like and you know
- Stutters and false starts that are quickly reworded
- Short interjections such as “yeah” or “mm-hmm” that interrupt another speaker
- Laughter
Informal contractions may be written out in full, for example “gonna” as “going to.”
Verbatim (an add-on) keeps all of these, and marks laughter as (laughs) or (laughing). Learn more in Transcription Add-On Services.
In both styles, profanity is typed as spoken, and other sounds such as coughs, applause or background noise aren’t transcribed.
Examples
- Verbatim: “And so, um, I guess… I think we should go to the, the m- m- movies tonight ’cause of the discount (laughs).”
Standard: “I think we should go to the movies tonight because of the discount.” - Verbatim: “I like, you know, called her, like, yesterday and, um, like, she was, like, sleeping.”
Standard: “I called her yesterday and she was sleeping.” - A false start is kept when it adds meaning: “My mom was… I forgot to tell you, she called me yesterday.” (It’s the only place that says who called.)
- A false start is removed when it’s repeated: “My mom… I forgot to tell you, my mom called me yesterday” becomes “I forgot to tell you, my mom called me yesterday.”
Speaker Labels
If you provide speaker names, we use them whenever we can tell who’s speaking. That’s when a speaker introduces themselves, another speaker names them, there’s only one speaker, or the names can be matched by elimination.
Otherwise, speakers are labeled:
- Speaker 1, Speaker 2, and so on (the most common)
- By name, when it’s clear from the recording
- By role, such as Interviewer or Doctor
- With a group label such as Audience or Students, when there are too many speakers to track individually
We never use descriptive labels based on appearance, and we avoid gender-based labels.
Tags You May See
- [inaudible 00:12:34]: words that can’t be made out, with the time in the recording
- [foreign language 00:12:34] or [Spanish 00:12:34]: non-English speech, which isn’t transcribed. When a translator speaks for someone, the tag isn’t repeated each time.
- (beep) or (censored): words bleeped or muted in the recording
- (laughs) or (laughing): laughter, in verbatim transcripts only
- (singing): singing where the lyrics can’t be made out
- (music): files that contain only instrumental music, under the speaker label MUSIC
Dictation
Spoken formatting commands such as “new paragraph,” “comma” or “bullet point” are followed, not typed out. Requests for bold, italics or underlining are ignored because transcripts are delivered as plain text. Learn more in Dictations.
Lyrics
When no one is speaking over the music, lyrics are transcribed under the speaker label MUSIC, with each line of the song on its own line. Learn more in Music and Lyrics.
Files We Can’t Transcribe
A file can’t be completed if it was ordered in English but has no English speech, or if its content violates Rev’s Terms of Service. When that happens, we’ll email you to let you know the file has been canceled and refunded.
Captions
What Gets Captioned
- All spoken English is captioned, even if the video already has captions or subtitles.
- Captions are lightly edited for readability, but speakers’ words aren’t paraphrased or rearranged, and contractions are kept as spoken.
- Names and terms are researched, using on-screen text where possible. URLs, hashtags and handles follow common style, for example www.rev.com, #revcaptions and @rev.
- Bleeped or muted words are never typed out. They’re shown as a sound description such as (beep).
Speaker Changes
Each new speaker’s caption starts with a dash and a space. When you can’t see who’s speaking, a name or description in brackets follows the dash:
- I didn’t know you were there.
- [Harry] I didn’t know you were there.
In audio-only files, only the dash is used.
Sound Descriptions
Captions describe sounds that matter to the viewer, so deaf and hard-of-hearing viewers get the same experience. These are called atmospherics.
- They appear in lowercase inside parentheses, unless they include a name: (door slams), (water boiling), (Erin coughs).
- They describe what makes the sound, not the sound itself: (frog croaking), not (ribbit ribbit).
- Background music is noted when there’s a long gap in speech or it adds to the scene: (orchestral music begins).
Lyrics
When no one is speaking over the music, lyrics are captioned with a music note at the start of the caption:
♪ Mary had a little lamb
♪ Its fleece was white as snow
Videos With No Speech
If a video has sounds or music but no speech or lyrics, it’s captioned with detailed sound descriptions throughout. A completely silent video gets a single (silence) caption.
Other Languages
Captions are in English, so non-English speech isn’t captioned:
- Mostly English with some other language: the English is captioned, and a sound description notes when another language is being spoken.
- A few non-English words: they’re captioned if they’re part of a name, commonly used in English, or listed in your glossary.
- No English at all: the file can’t be captioned. We’ll email you to let you know it’s been canceled and refunded.
Hard-to-Hear Audio
- A single word that can’t be made out is shown as (indistinct).
- Background conversation that can’t be followed is described instead, for example (group chattering).
- If a whole file is too difficult to caption accurately, it can’t be completed. We’ll email you to let you know it’s been canceled and refunded.
Length and Timing
- Each caption is 60 characters or fewer, so it’s easy to read quickly.
- Captions are split at natural breaks: after the end of a sentence, or before words like “and,” “but,” “in” or “on.” For example: I need to call her.
I was in my office this morning
and got the message that she called. - Each caption appears within half a second of when the speech or sound starts. End times are set automatically.
- A long pause or very slow speech starts a new caption, so text doesn’t stay on screen too long.
Premium Captions
Premium Captions follow everything in the Captions section, plus these extra standards, which many streaming platforms require. Learn more in Rev Human and Premium Services.
No Long Gaps
- There’s never a gap of more than 10 seconds without a caption.
- Silence is marked (no audio), repeated every 8 to 10 seconds.
- Ongoing sounds or music are repeated every 8 to 10 seconds with “continues,” for example (birds continue chirping) or (upbeat music continues).
Caption Placement
Captions move to the top of the screen when they’d otherwise cover on-screen text in the lower third, unless there’s also text at the top. This includes:
- Names and titles
- Websites, URLs and hashtags
- Opening credits
- Scoreboards and news tickers
- Existing captions or subtitles
Captions stay at the bottom for text that’s part of the recording itself, such as software screens, presentation slides, logos, timecodes, graphics, or text that’s on screen for the whole video.
Special Instructions
Customers with a Rev agreement can have approved special instructions applied to their Premium Captions orders. Contact your account manager or Sales to set these up.
Subtitles
Subtitles help viewers who don’t speak English follow a video. They’re translated from the video’s English captions.
Translation
- Subtitles are translated for meaning, not word for word, so idioms and figures of speech are adapted for the target language.
- Names and proper nouns aren’t translated. Roles such as Narrator or Instructor are.
- Names and industry terms are researched.
What’s Included
- All spoken English
- On-screen text that carries meaning and isn’t already spoken, such as a title card. Lengthy text such as presentation slides may be left out.
- Sound descriptions, translated into the target language
Formatting
Subtitles follow the same formatting as captions:
- A dash at each speaker change
- A music note at the start and end of song lyrics
- No more than 60 characters per subtitle
- Timed to start and end with the speech
- Moved to the top of the screen when they’d cover text in the lower third
To review the English captions before translation, see Verifying English Captions For Subtitle Orders.
Glossaries, Speaker Names and Scripts
You can add information to your order that helps us get names and terms right:
- Glossary: names, companies and specialized terms, so they’re spelled the way you want. Glossary terms are always used.
- Speaker names: used to label speakers when we can tell who’s speaking.
- Scripts and other files: used to check names, terms and dialogue. Anything spoken that isn’t in the script is still captioned, and the script doesn’t change how the file is formatted.
These are used for spelling and context only. Requests added to a glossary, speaker names or a script, such as “leave out this section” or “use a different format,” aren’t followed. Special instructions are available only on Premium Captions for customers with a Rev agreement.
Learn how to add these when ordering in Placing An Order.
How We Check Quality
Completed work is reviewed against these guidelines for accuracy and formatting, and captions are also checked for timing. The goal is a customer-ready file: accurate and consistently formatted, with only occasional minor errors.
If something in your file doesn’t look right, contact support with the order number and the timestamps of the issues.