Listening field note
Can SRT captions become an interview passage index?
SRT cues can supply bounded time ranges and text segments for passage capture, but caption blocks, speaker labels, line breaks, and timing may reflect display needs rather than the project's preferred transcript or archival structure.
Parse cue timing strictly
Each cue needs a valid start, end, and nonempty text. Reject malformed ranges instead of guessing. Write the decision in plain language and retain the original source separately so the index can be audited against something outside the browser project.
Treat speaker prefixes carefully
A leading Name: pattern can seed a speaker label, but it cannot identify a voice or prove that the captioner used names consistently. Review the result as a retrieval aid rather than a conclusion. Human-authored fields must remain visibly distinct from exact source-derived text.
Preserve the source role
Record that the imported source was SRT and retain the original separately; passage exports should not masquerade as the canonical transcript. 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.