# Reference-based review

The model detected one speaker, matching the excerpt reference. Some wording still needs correction.

Reference-based comparison against the AMI manual word annotations v1.6.2, including speaker IDs and original source times. Not a listening audit.

- The opening reference “And the DIVA” is transcribed as “In the diva”; the wording and technical-name capitalization differ.
- The final segment adds “a product”, which is absent from the AMI word reference. This has not been adjudicated by listening.
- DirectShow and the phrase “two cards, one for audio and one for video” are present in the model output.
- One model speaker label is used throughout, matching the one annotated source speaker B.

## Limits

- These are two short, selected excerpts from real non-scenario multi-person meetings, using the official mixed headset recording. They are not privately recorded one-person/two-person sessions.
- The reference contains false starts, fillers and overlapping speech. Word-count differences are not an error rate, and no WER/DER or general accuracy claim is made.
- Timestamps are model output, not independently listened-to alignment measurements. All normalized timing statuses remain unknown, even where numeric times exist.
- No output text, speaker label or timestamp was manually corrected. Original provider output and normalized original exports are preserved.
- This direct-provider test uses the unchanged production input adapter, pinned model, normalizer and exporters. It does not test application login, credits, job database, webhook handling or browser processing.