Model
DeepFilterNet
A deep learning audio noise suppression network employing a two-stage architecture that separately processes speech envelope features and fine detail features. It demonstrates effective suppression of non-stationary noise compared to earlier approaches.
Timeline (last 90 days)
Sep 17
DeepFilterNet employs a two-stage architecture to process the envelope features and detail features of speech separately, and also performs well on non-stationary noise
Unverified50%