Improved Speech Intelligibility With Cochlear Implants Using State-of-the-Art Noise Reduction Algorithms

Speech intelligibility in noisy environments remains a difficult challenge for cochlear implant (CI) users. Signal processing techniques such as beamforming and singlemicrophone noise reduction can partially alleviate this problem. We tested the effect of one adaptive monaural and one binaural beamformer by hearing-aid manufacturer Phonak on speech intelligibility (SI) in largely diffuse noise in a group of 12 unilateral users of the Advanced Bionics (AB) Harmony speech processor. Both beamformers were further combined with AB’s noise reduction algorithm ClearVoice. Two test sessions were performed per subject. In session A, we found significant benefits for both the adaptive beamformer and ClearVoice, yielding a combined SRT improvement of 6.1dB. Session B revealed a further benefit for the binaural beamformer. Together with ClearVoice, the resultant SRT improvement over the unprocessed signal was 7.86dB.