Transcription Style Guide & Handbook
Official formatting and quality standards for transcribers on TranscribeHub.
📖 1. Clean Verbatim vs. Strict Verbatim
Unless specified otherwise in the client's guidelines, our default format across the platform is Clean Verbatim.
✓ Clean Verbatim (Default)
- Remove filler words: "um", "uh", "ah", "er"
- Remove false starts and stuttering (e.g., "I th- I think" → "I think")
- Preserve slang like "gonna", "wanna" only if essential to tone; prefer formal spelling
- Maintain speaker meaning without editorial rewriting
★ Strict Verbatim (When Requested)
- Capture every vocalization, hesitation, and repetition
- Keep all stutters: "I, um, I was going to, ah, mention..."
- Include throat clearing and verbal tics
- Required for legal depositions and linguistics studies
🗣️ 2. Speaker Labels & Tags
Clearly separate dialogue whenever speakers change.
Interviewer: Welcome to the conference. Can you explain your architecture?
Dr. Sterling: Absolutely. We transitioned to a distributed ledger model last quarter.
Interviewer: How did that impact transaction latency?
⏱️ 3. Timestamps & Non-Speech Events
Use standard bracket tags for sound events and speech interruptions:
[inaudible 02:15]
When a word or phrase cannot be deciphered after 3 listens, add timestamp inside tag.
[crosstalk]
When two or more individuals talk simultaneously over each other.
[laughter] / [applause]
Significant crowd reactions or prolonged laughter that pauses the speaker.
[music]
Intro/outro theme tracks or background audio interludes.