Dr. Kim Ngo

As a part of the Real-Time Communication (RTC) group the focus is research and development on our Audio-Conferencing solutions such as GoToMeeting or GoToWebinar. Area of audio R&D: - Speech processing (AEC/AES, NR, AGC, VAD, speaker selection, resampling, mixing) - Speech coding (iSAC, iLBC, SILK, OPUS, G.7xx) - VoIP network (jitter buffer, Packet loss concealment, RTP, signalling protocols)

Subband Adaptive filtering approach for Acoustic Echo Cancellation (AEC) for cellphones (speaker-phone or hands-free mode) Compensating for non-linear effects from microphone and loudspeaker in cellphones. Making the AEC algorithm robust against double-talk, noise, and reverberation.

This research project is a joint collaboration between Katholieke Universiteit Leuven and NXP Software. The topic is subband acoustic echo cancellation for cellphones.

Research on adaptive beamformers (MVDR) for small arrays with Prof. Douglas L. Jones. The focus is to estimate the steering vector in a robust way in the the presence of reverberation and noise keeping in mind a low complexity approach.

Research on exploiting pitch estimation in adaptive feedback cancellation for hearing aids with Prof. Søren Holdt Jensen and Prof. Mads Græsbøll Christensen. Pitch estimation is used to model the near-end speech which can be used in the decorrelation process.

Thesis: Digital signal processing algorithms for noise reduction, dynamic range compression, and feedback cancellation in hearing aids.

Thesis: MFCC Based Robust Audio Classification Using The Multivariate Autoregressive Model

Thesis: Noise Robust Automatic Speech Recognition (ASR)