Last Confirmed Usable Sample
mast_center_aquarium_20260827T143843.ts. About 24 seconds, video plus audio. The aquarium scene is still visible.
This page records why the MaST Center video detector batch was stopped before publication: the camera continued streaming, but water on or inside the camera blocked the aquarium view before the Friday batch window.
5,015captured batch segments, Aug 28-31
0published detection runs
20260827T143843last confirmed usable sample
20260828T123220first confirmed blocked sample
Interpretation: the correct public result is a source-quality gap, not a no-detection result. Missing or water-blocked visibility must not be counted as biological absence.
Evidence
First Friday Sample Checked
mast_center_aquarium_20260828T123220.ts. About 60 seconds, video plus audio. The view is already water-blocked.
Later Blocked Batch Sample
mast_center_aquarium_20260828T183638.ts. About 60 seconds, video plus audio. This confirms the blocked-camera state persisted later on Aug. 28.
| Partition | Segments | First item | Last item | Quality note |
|---|---|---|---|---|
| 2026-08-27 spot-check context | 811 | mast_center_aquarium_20260827T000023 | mast_center_aquarium_20260827T143843 | Last sampled clip was visually usable; late frames showed aquarium scene and fish. |
| 2026-08-28 batch start day | 689 | mast_center_aquarium_20260828T123220 | mast_center_aquarium_20260828T235948 | First sampled clip at 12:32 PDT was already water-blocked. |
| 2026-08-29 | 1,442 | day partition | day partition | Blob sizes match the blocked-camera regime; treat as captured but poor source quality. |
| 2026-08-30 | 1,442 | day partition | day partition | Blob sizes match the blocked-camera regime; treat as captured but poor source quality. |
| 2026-08-31 | 1,442 | day partition | day partition | Blob sizes match the blocked-camera regime; treat as captured but poor source quality. |
Transition Bound
The last visually usable sampled clip was mast_center_aquarium_20260827T143843, labelled 2026-08-27 14:38 PDT in the video overlay. The first captured Friday clip checked was mast_center_aquarium_20260828T123220, labelled 2026-08-28 12:32 PDT, and was already water-blocked.
No local laptop clips were found between those two timestamps, so the current bound for the transition is the capture gap between those samples. The blocked condition may have begun any time inside that interval.
Run Decision
The YOLO-Fish / marine-review full-fidelity batch was started, then intentionally stopped after the source-quality spot check. Continuing would primarily spend compute on blurred water/overlay artifacts. The saved private handoff keeps the contact sheets and partial logs for audit.
Limits
This report does not claim that no fish or marine mammals were present during the batch window. It says the camera view was not reliable enough for detection reporting from the captured Friday-to-cutoff video.