Dushyant Sharma
Dushyant Sharma is a Principal Research Scientist, working as part of the Central Research and Development organization at Nuance on front end signal processing and acoustic data augmentation. His research interests include non-intrusive speech signal analysis, single and multi-channel speech enhancement and acoustic data augmentation and simulation. Dushyant joined Nuance in 2012 after completing his Ph.D. in speech signal processing from the Centre for Law Enforcement Audio Research (CLEAR) at Imperial College in London, UK.
Dushyant Sharma’s stories
Improved Far Field Speech Recognition with Microphone Arrays

The problem of Distant Automatic Speech Recognition (DASR) remains an important topic that has received much attention in the past few years. DASR represents an important component for enabling a truly natural and ambient intelligence system such as Nuance’s Dragon Ambient eXperience (DAX). The current state of the art methods for DASR are based on […]

Delivering non-intrusive signal intelligence with deep learning

When a signal is acquired in a room with reflective surfaces and noise sources, the acquired signal can be modeled by convolution with the room impulse response (RIR) plus an additive noise component. The level of reverberation can be characterized by an RIR, from which the Clarity Index (C50) metric can be obtained which has […]