How OpenNoiseNet explores smartphone sensing, private device networks, and a future public noise map
Seven parts of the architecture that are already in motion while the stabilization gate remains open

“Smartphone node: capture, queue, recover”
The mobile app is being hardened into a long-running field node with local monitoring, recovery, sync diagnostics, and privacy-first uploads.
Mobile runtime and unattended capture

“Derived-first by default”
OpenNoiseNet keeps numeric levels, lifecycle state, and policy-governed evidence as the default path. Raw audio is not the normal cloud artifact.
Privacy-preserving architecture

“event_uuid: detect, queue, upload, ACK”
Stable identifiers, receipts, acknowledgements, and sync diagnostics let reportable events survive reconnects and be traced end to end.
Lifecycle observability and reconciliation

“Event stream → episode”
Server-side logic now groups events into episodes with severity, nuisance score, review state, and export lifecycle instead of leaving operators with raw event noise.
Incident construction and review state

“One private overview”
A private device network is being tested with connected phones. It is not yet generally available.
Private sensor overview

“A shared view of noise”
A free public map is the project vision. It is not yet available.
Community noise reporting

“Testing in real life”
Reliable long-running operation, calibration, and privacy remain important areas of testing.
Validation and transparency
A practical view of what is being built and what remains to be tested
Sound is analysed on the device. Sharing derived information is a separate choice, and the selected server operator can read data sent to it.
Reliable long-running sensing and the quality of smartphone estimates still require real-world testing.
The next steps include testing reliability, calibration, and privacy in everyday use.
Follow the repo, validate the field node, and help make the operator workflow trustworthy enough for early pilots.