• The Mert audio algorithm library supports AI for removing stationary noise and wind noise, AI speech separation, and AI echo cancellation. Inference supports speech flow mode, and a single network model can remove stationary noise, wind noise and echo cancellation. It supports dynamic range control, equalization, and multi microphone voice enhancement.
  • Specifications

    8k and 16k sample rate are supported.


    1. Flow pattern inference with zero latency;

    2. A single model can simultaneously reduce noise and eliminate echo;

    3. Seamless switching between listening mode and intercom mode;

    4. Echo elimination during intercom, no cutting sound phenomenon.