Biological activity
Ocean acoustic intelligence
The ocean does not arrive labeled.
Sona helps marine science and monitoring teams move from long-duration underwater recordings to a reviewable account of where attention is needed, what the evidence supports, and what remains unresolved.
Source record
Kept authoritative
Decision state
Supported, uncertain, or unresolved
Review history
Retained with the observation
01 / The work
Built for the review bottleneck.
A hydrophone can record for months. The hard part is deciding what deserves expert attention without losing the evidence needed to revisit or defend that decision.
-
01
Receive
Bring underwater recordings together with their capture context and provenance.
-
02
Review
Work through observations that remain connected to the source record.
-
03
Decide
Mark what is supported, what is uncertain, and what remains unresolved.
-
04
Retain
Carry source and review history into later analysis or re-evaluation.
The acoustic record
Different sounds carry different scientific stakes.
Biology, human activity, environmental processes, and recording faults can coexist in the same material. A useful review process keeps those possibilities open until the evidence supports a narrower conclusion.
Human-made sound
Which observations matter to the monitoring objective?
Environmental context
Is the acoustic setting changing across time or place?
Recording integrity
Is a change in the ocean—or in the recording chain?
02 / Technical standard
Keep the evidence stronger than the claim.
Sona’s public technical position is intentionally simple: results should remain traceable, uncertainty should remain visible, and evaluation should reflect the conditions in which the work will be used.
Source integrity
The original recording remains the authority; derived material does not replace it.
- Capture context retained
- Derived observations traceable
- Integrity problems made visible
Honest uncertainty
An available label is not automatically a supported identification.
- Uncertain states stay explicit
- Abstention remains available
- Expert decisions stay attributable
Evaluation discipline
A favorable average is incomplete without calibration, shift, and failure analysis.
- Development and evaluation separated
- Operating conditions represented
- Limitations reported with results
03 / Research status
Validation before promotion.
Sona is under active validation. Internal results are not presented as public performance claims, and implemented capability is not treated as scientific proof.
Similarity is not identity. Detections are not abundance. A cleaner signal is not automatically preserved evidence.