Enhancing binaural speech perception in noise via weighted coherence masking for hearables.

A weighted masking method based on the coherent-to-diffuse ratio is presented for robust binaural speech enhancement in real-time hearable devices. The method applies manually tuned weights across custom-defined critical frequency bands to improve th…

Preview only. Read the full abstract at the source