Real time Single channel Speech Enhancement with Recurrent Neural Networks

Singlechannel speech enhancement using deep neural networks (DNNs) has shown promising progress in recent years. In this work, we explore several aspects of neural network training that impact the objective quality of enhanced speech in a realtime setting. In particular, we base all studies on a novel recurrent neural network that enhances fullband shorttime speech spectra on a singleframein, singleframeout basis, a framework that is adopted by most classical signal processing methods. We propose tw