Neural speech enhancement runs on a mobile device so ear-wearable hearing hardware can cut noise and adapt to changing acoustic conditions.
Sound classification selects simplified or complex DNNs to improve speech intelligibility in noise while limiting power and processing demand.
Single chip analog interface unit offloads digital signal processing to smartphone CPU and GPU, reducing device size and cost.
A smartphone-connected ear-level device measures auditory thresholds to generate customized sound processing schemes.
Series-connected lanyard antenna segments compensate for increased distance to maintain reliable near-field data reception despite head movement.
A hearing aid system splits audio signals to process only compensation components on a smartphone, reducing data volume and processing time.
A pivot connection rotates the earphone stem relative to the inner shell.