Sound is not only output.
It is evidence.
Our research connects physical acoustics, human perception, spatial reasoning and deployable machine intelligence.
Research agenda
Auditory scene analysis
Separate events, sources and activities from complex acoustic mixtures.
Source localization and tracking
Estimate position, motion and identity under real spatial constraints.
Spatial scene models
Represent entities, zones, relationships and state as a machine-readable scene.
Context and intent inference
Connect acoustic evidence with temporal, device and application context.
Edge intelligence
Run perception close to the environment for latency, resilience and privacy.
Adaptive interaction
Translate scene understanding into natural, proportionate system responses.
From question to reusable capability
We organize research around problems that can become algorithms, models, interfaces and system behavior—then validate them under real acoustic and operational constraints.
- 01Frame
- 02Model
- 03Prototype
- 04Validate
- 05Productize
Responsible by design
Acoustic intelligence must be designed around purpose limitation, data minimization, edge processing, explicit permissions and observable system behavior.
Read our principles