Use deep learning to reduce environmental noise and isolate voice.
- 100~300KMemory Usage
- 16~32msLatency
- 25~60dBNoise Reduction Depth
- 50~200MIPSCompute
Our Edge
Flexible sampling rates: 16/48/96 kHz
Configurable latency, model size, and noise reduction depth
Native integration with AudioWeaver and MATLAB
Best-in-class performance per KB footprint
Proven robustness against wind and non-stationary noise