Speech enhancement:
Development and design of algorithms for improving speech intelligibility in noise
Development of perceptually-motivated statistical estimators of the magnitude/power spectrum
Development of noise-estimation algorithms suitable for highly non-stationary noisy conditions
Development of objective measures for predicting speech intelligibility in noisy conditions
Multi-microphone noise reduction algorithms (e.g., beamforming, blind-source separation, etc)
De-reverberation algorithms
Speech Perception:
Understanding how normal-hearing listeners cope with noisy backgrounds - "cocktail party" listening
Understanding the factors influencing perception of speech in noise
Development of objective measures for predicting speech intelligibility in noisy conditions
[Research supported in part by NIH ]