Listening field note
Can WebVTT captions be used for interview retrieval?
WebVTT cues can provide browser-friendly time ranges and readable text for local passage indexing, provided unsupported metadata, styling, cue settings, and speaker markup are handled conservatively and the original file remains preserved.
Remove display markup carefully
A plain retrieval index may strip cue tags for reading while retaining the original WebVTT as the authoritative caption artifact elsewhere. Write the decision in plain language and retain the original source separately so the index can be audited against something outside the browser project.
Do not infer identities from tags
Voice spans and labels are metadata supplied by the file, not biometric evidence or verified identity. Review the result as a retrieval aid rather than a conclusion. Human-authored fields must remain visibly distinct from exact source-derived text.
Test against actual media
Caption timing can drift or reflect edited media; a human should seek representative passages in the authorized recording. Carry the limitation into the frozen edition and passive handoff so a recipient can understand what was checked, what was not checked, and what remains their responsibility.