Music and Lyrics
How music and lyrics appear in your file depends on the service you order. Human transcriptionists and captioners follow set rules for when to include lyrics and how to mark music. AI transcription and captions write down whatever speech the system detects. This article explains what to expect from each service.
In this article
Human Transcription
Rev transcribes English lyrics when they’re the focus of the recording or when a main speaker sings. Background music isn’t transcribed.
Background Music
Lyrics from background music aren’t transcribed. Short music breaks in a file with speech, such as a podcast intro or transition music, aren’t marked either.
Example: A song plays on a TV in the background while people talk in the foreground. Only the conversation is transcribed.
Lyrics
Lyrics are transcribed when the song is the focus of the recording. For example:
- The file is a music video, a performance, or a clip built around a song.
- A main speaker sings.
- The speakers sing along with a song or talk about it while it plays.
Lyrics are transcribed the same way as speech.
Instrumental Music and Silence
If a file has only instrumental music, with no singing or speech, the transcript contains a single (music) tag with “Music” as the speaker label. A completely silent file comes back with a single (silence) tag.
Lyrics in Other Languages
Human transcription covers English lyrics.
- A few foreign-language words in an English song are transcribed as sung.
- Several lines of non-English singing are marked with a foreign language tag and the timestamp, such as [foreign language 00:01:15].
- A file with only non-English lyrics can’t be transcribed.
Human Captions and Subtitles
Human captioners follow the same rules for when lyrics are included: lyrics are captioned when the song is the focus or a main speaker sings, and not when the music is in the background. Music is handled differently in captions in two ways:
- Music becomes a sound description. Instrumental music, including background music, is shown as a sound description such as (music). Your Atmospherics setting in File preferences controls whether these use parentheses, brackets, or are left out.
- Lyrics you can’t make out: If a song’s lyrics can’t be understood, the captions may show only a sound description for that part, or for the whole file if it’s all music.
Subtitles are translated from the English captions, so lyrics that appear in the captions are translated too.
AI Transcription and AI Captions
AI transcription and AI captions don’t apply these rules. The system writes down whatever it recognizes as speech, so:
- Lyrics may be transcribed whether they’re the focus or in the background, and may be less accurate than speech.
- Instrumental music isn’t marked with a (music) tag or a sound description.
- Loud music behind speech can make nearby words harder for the AI to recognize.
If your file has a lot of music or singing, human transcription or captions will give you more accurate, consistently formatted results.