|
|
|
|
 
  Noise Reduction (NR) single microphone

Speech enhancement systems are widely used for hands-free communication and speech recognition. The most fundamental method is adaptive filtering where noise and voice are assumed to be uncorrelated components within the audio stream. Our NR technology is based on statistical properties of human speech and has shown better results. This method models and estimates the distributions of speech and noise spectra employing statistical estimators. It provides more accurate modeling of speech prior distribution and improves the performance of speech enhancement in terms of both sound quality and speech recognition.

Specifications:

  1. Noise suppression up to -30 dB;
  2. Processing delay: < 30 ms;
  3. Supported sampling rates: 8 KHz, 16 KHz;