Multichannel Speech Enhancement Approaches to DNN-Based Far-Field Speech Recognition
- Marc Delcroix ,
- Takuya Yoshioka ,
- Nobutaka Ito ,
- Atsunori Ogawa ,
- Keisuke Kinoshita ,
- Masakiyo Fujimoto ,
- Takuya Higuchi ,
- Shoko Araki ,
- Tomohiro Nakatani
in New Era for Robust Speech Recognition
Published by Springer | 2017
In this chapter we review some promising speech enhancement front-end techniques for handling noise and reverberation. We focus on signal-processing-based multichannel approaches and describe beamforming-based noise reduction and linear-prediction-based dereverberation. We demonstrate the potential of these approaches by introducing two systems that achieved top performance on the recent REVERB and CHiME-3 benchmarks.