Evaluation of Frequency-Lowering Algorithms for Intelligibility of Chinese Speech in Hearing-Aid Users
Xianbo Xiao,Hui Zhang,Guangshu Hu,Chunhong Liu,Jia Liu
DOI: https://doi.org/10.1016/j.pnsc.2008.08.007
2009-01-01
Abstract:Despite the attention being given to and the knowledge of the benefits of evaluating frequency-lowering algorithms for hearing-impaired people, the causality between these algorithms and their benefits is still not clear. This is aggravated by the fact that comparative research on the methodologies and skills required for fitting an appropriate algorithm to individual patients is lacking. Against this back-drop, the current study has attempted to make progress in this area. In this experiment, six experienced traditional hearing-aid users with severe impairments in the high frequencies were fitted with two different frequency-lowering methods, and weekly hearing tests were conducted to track the benefits of such methods. After the experiment, five of the listeners accepted the frequency-lowering algorithms. Both methods showed superior results when compared with the listeners’ own hearing aids in most of the tests, and the segmented compression algorithm was indicated to have better "anti-noise" quality and speech intelligibility improvement capabilities. A preference for the application of proportional compression to unvoiced speech was also found in this algorithm. Unlike in previous studies, all speech materials here are recorded in Chinese. Therefore, the results could also be used to evaluate the benefits of frequency-lowering to the intelligibility of Chinese speech for patients. As a result, an improvement was found, especially in the recognition of consonants. Moreover, no adverse effect was found in intonation recognition.