Correct The Caption, Not The Speaker
A fluent caption can reverse an instruction or attach it to the wrong person. Review the wording and the speaker label separately. When the reference does not resolve a detail, keep the uncertainty visible.
This text-only exercise uses fictional, timestamped excerpts and a reference transcript supplied with the input. There is no audio file or captured model run. It checks whether a reviewer preserves evidence, not how well a tool recognizes speech or synchronizes captions.
Get The Pack
Send only the input file. Keep the review key and worksheet separate, and save the first response before requesting a correction.
Download comparison worksheet (.txt)
Review Three Caption Errors
| Case | Draft caption | What to check |
|---|---|---|
| CAP-01 | “Do publish the sample today.” | The reference says “Do not publish the sample today.” Restore the negation. |
| CAP-02 | “Use test ticket Q-510.” | The reference says Q-501. Copy the identifier exactly. |
| CAP-03 | “Mara: I will check the export.” | The words are verified, but the speaker is not. Use an unknown-speaker label; do not assign the work to Mara. |
For CAP-03, an acceptable caption is “Unknown speaker: I will check the export.” Leave the attribution unresolved until there is better evidence. Removing the whole caption would also lose verified words.
Keep the supplied time ranges in all three cases. They identify where to inspect the recording; without audio, there is no basis to shift them.
Return With The Same Input
Keep the exact input and first output with your worksheet. Record any later corrections separately. After changing tools or instructions, repeat the same pack and compare each error type. Without a comparable earlier answer, record a baseline instead of claiming improvement.
Return to the caption practice shelf entry and choose your next check based on the evidence you have.
Review The Recording Before Sharing
For a real caption file, replay the affected segment with the approved recording and check timing, speaker changes and important sounds. Verify the export in the intended player before sharing. The text exercise cannot establish speech-recognition accuracy or the readability of the actual file.
Use only approved recordings and destinations. See the meeting tool review guide before uploading or sharing a transcript.