Improving the performance of speech perception in noisy environment based on an FAME strategy
Ying-Hui Lai, Syu‐Siang Wang, Yu-Ting Su, Cheng Han-Che, Fan Kang Fu, Yu Tsao · 2016
Background noise is a significant factor influencing the performance of speech perception. Previous studies showed that the temporal fine structure (TFS) plays an important role in speech perception for normal hearing (NH) and hearing loss individuals. The frequency amplitude modulation encoding (FAME) is a successful approach to enhance the TFS information for cochlear implant (CI) recipients. Following the success of FAME for CI recipients, this study aims to evaluate the speech perception performance of FAME for NH listeners in noisy conditions. Experimental results from the present study confirmed that FAME provides better speech perception performance and lower listening effort for NH listeners than noisy speech. In particular, FAME improved Mandarin disyllabic words recognition by as much as 16.7 percentage points and the ease of listening by 1.6 (MOS scale). This demonstrates that the FAME strategy is promising for improving speech recognition performance for NH listeners in noisy environment.