Binaural Heterophasic Superdirective Beamforming

Area of Science:

• Acoustic Signal Processing
• Psychoacoustics
• Auditory Perception

Background:

• Superdirective beamformers are effective for broadband acoustic signals but amplify white noise.
• Conventional beamformers require extremely low self-noise sensors, which are difficult to achieve.
• Existing methods struggle with effective noise reduction in complex acoustic environments.

Purpose of the Study:

• To propose a novel binaural superdirective beamformer design.
• To address the challenge of white noise amplification in acoustic signal processing.
• To improve the separation of target acoustic signals from noise for enhanced auditory perception.

Main Methods:

• Developed a binaural beamformer composed of two orthogonal sub-beamformers.
• Applied psychoacoustic principles to design filters that decorrelate white noise components.
• Maximized interaural coherence for diffuse noise to aid sound source localization.

Main Results:

• The proposed binaural beamformer produces in-phase signals of interest and random-phase white noise.
• Experimental results demonstrate superior performance compared to conventional monaural beamformers.
• The design facilitates better separation of acoustic signals from white noise by the human auditory system.

Conclusions:

• The novel binaural superdirective beamformer effectively mitigates white noise amplification.
• This approach leverages psychoacoustic principles for improved noise reduction and source localization.
• The proposed method offers a significant advancement over traditional monaural beamforming techniques.