Speech Intelligibility Enhancement Based on a Non-causal Wavenet-like Model

• DOI: 10.21437/interspeech.2018-2119
• Publication Date: 2018-08-28T09:55:42Z
• Authors: Muhammed Shifas PV; Vassilis Tsiaras; Yannis Stylianou
• Keywords: Intelligibility (philosophy)

Low speech intelligibility in noisy listening conditions makes more difficult our communication with others.Various strategies have been suggested to modify a signal before it is presented environment the goal increase its intelligibility.A state-of-the art approach, referred as Spectral Shaping and Dynamic Range Compression (SS-DRC), relies on modifying spectral temporal structure of clean has shown considerably improve conditions.In this paper, we present non-causal Wavenet-like model for mapping samples generated by SSDRC.A successful non-linear function potential be used a) improving b) Wavenet-based synthesizers based improvement layer.Objective subjective results show that able reproduce gains SSDRC, while far improves quality modified compared obtained SSDRC.