作者: Simon Doclo , Marc Moonen , Ann Spriet , Kim Ngo , Soren Holdt Jensen
DOI: 10.5281/ZENODO.41111
关键词:
摘要: Hearing aids typically use a serial concatenation of Noise Reduction (NR) and Dynamic Range Compression (DRC). However, the DRC in such negatively affects performance NR stage: residual noise after is amplified by DRC, resulting signal-to-noise-ratio (SNR) degradation. In this paper, we present an integrated solution for DRC. The based on estimate amount speech each time segment. case dominant, less active it desirable to have as much possible, whereas dominant segment more idea not compromise operation applying Experimental results confirmed that degrades SNR improvement, proposed offers better improvement compared concatenation.