What happens when two people talk at once?

Most style guides ask you to label each speaker separately and mark the overlap, rather than merging the turns. Some use a specific tag for simultaneous speech; check the brief.

Should labels be corrected if diarization got the count wrong?

Yes. Diarization decides how many voices it heard, which is frequently wrong when speakers sound similar or the audio is poor. Merging or splitting speakers is a normal review step.

Do speaker labels belong in subtitle files?

Usually only in SDH captions, where the viewer cannot hear who is speaking. Standard subtitle tracks rely on the picture to identify the speaker.

Can speaker labels be changed after delivery?

In a text transcript, yes, though renaming consistently across a long file is easy to get wrong by hand. If the labels came from diarization, changing them at the source and re-exporting is safer than find-and-replace.

Definition

What Are Speaker Labels?

VideoText editorial team. Updated September 13, 2026.

Direct definition

Speaker labels are the names or codes written next to each turn in a transcript — “Alex:”, “Speaker 2:”, or “Interviewer:”. They come from diarization, a human, or both.

The sections below explain how speaker labels is used in transcription and subtitle work, what people often mix up, and which VideoText workflow applies when the next step is a file or review pass.

Key takeaways

Pick one format and keep it consistent (name, role, or Speaker N).
Style guides specify punctuation and whether labels sit on their own line.
Wrong labels are a QA failure even if the words are right.

Example

Speaker 1: Did the export include VTT? Speaker 2: Yes, and the SRT.

Choosing a labelling scheme

There are three common schemes and they are not interchangeable. Names (Alex:) read best and require you to know who is who. Roles (Interviewer:) suit research and legal work where the function matters more than the person. Numbers (Speaker 2:) are the fallback when identity is unknown.

Pick one before starting and apply it throughout. A transcript that opens with Speaker 1 and switches to names halfway through is a rejection on most marketplaces, even when every word is correct.

Formatting rules that vary by client

Style guides differ on whether the label sits on its own line or inline with the speech, whether it is bold, whether it repeats on every turn or only on change, and whether the colon is followed by one space or two. None of these affect meaning and all of them are checked in QA.

Where a speaker is unidentified for part of the recording, most guides ask for a consistent generic label rather than a guess, and a note at the top of the transcript explaining the convention used.

Frequently asked questions

What happens when two people talk at once?: Most style guides ask you to label each speaker separately and mark the overlap, rather than merging the turns. Some use a specific tag for simultaneous speech; check the brief.
Should labels be corrected if diarization got the count wrong?: Yes. Diarization decides how many voices it heard, which is frequently wrong when speakers sound similar or the audio is poor. Merging or splitting speakers is a normal review step.
Do speaker labels belong in subtitle files?: Usually only in SDH captions, where the viewer cannot hear who is speaking. Standard subtitle tracks rely on the picture to identify the speaker.
Can speaker labels be changed after delivery?: In a text transcript, yes, though renaming consistently across a long file is easy to get wrong by hand. If the labels came from diarization, changing them at the source and re-exporting is safer than find-and-replace.

Related terms

Sources

Rev: Rev Transcription Style Guide (speaker labels)