User Customizable Real-Time Single and Dual Microphone Speech Enhancement and Blind Speech Separation for Smartphone Hearing Aid Application

Speech Enhancement (SE) is a vital algorithmic component in the Hearing Aid pipeline. Over the years, several algorithms have been developed to work in real-time and to improve the quality and intelligibility of speech. However, noise suppression with minimal distortion to speech is still a prime challenge that needs to be addressed. In this work, a new single microphone SE is introduced that is implemented on a smartphone to work as an assistive device to Hearing Aids via wireless connectivity. The uniqueness of the developed method is in the introduction of varying parameters that allow the smartphone user to control the amount of noise suppression and speech distortion in real-time, which allows the user to customize the perceptual audio to their preference. A super-Gaussian extension of this approach is explored and analyzed. With the recent accessibility of the two microphones on the smartphones, doors were opened to use beamformer as a pre-filtering stage to the proposed single microphone SE. Real-time blind speech separation technique is also proposed to yield superior quality for speech. Objective and subjective results show that the developed methods outperform traditional SE techniques.