There are numerous single-channel noise reduction strategies to improve speech perception in noise. However, most of them improve speech quality but do not improve speech intelligibility, in circumstances where the noise and speech have similar frequency spectra. Current exceptions that may improve speech intelligibility are those that require a priori knowledge of the speech or noise statistics, which limits practical application.

Hearing impaired (HI) listeners suffer more in speech intelligibility than normal hearing listeners (NH) in the same noisy environment. As a result, developing better single-channel noise reduction algorithms for HI listeners is justified. Researchers from the Institute of Sound and Vibration Research, University of Southampton, UK and the Institute of Acoustics, Chinese Academy of Sciences propose a model-based “sparse coding shrinkage” (SCS) algorithm which extracts key speech information in noisy speech.

It is evaluated by comparison with a state-of-the-art Wiener filtering approach using speech intelligibility tests with NH and HI listeners. The model based SCS algorithm relies only on statistical signal information without prior information. It shows that the SCS algorithm improves speech intelligibility in stationary noise and is comparable to the Wiener filtering algorithm. And both algorithms improve intelligibility for HI listeners but not for NH listeners. Improvement is less in fluctuating (babble) noise than in stationary noise. Moreover, both noise reduction algorithms perform better at higher input signal-to-noise ratios (SNR) where HI listeners can benefit but where NH listeners have already reached ceiling performance. The difference between NH and HI subjects in intelligibility gain depends fundamentally on the input SNR rather than the hearing loss level.

Results show that HI listeners need different signal processing algorithms from NH subjects. Also, the SCS algorithm offers a promising alternative to Wiener filtering. Performance of all noise reduction algorithms is likely to vary according to extent of hearing loss. As well, algorithms that show little benefit for listeners with moderate hearing loss may be more beneficial for listeners with more severe hearing loss.

This research was supported by the European Commission within the ITN AUDIS (grant agreement number PITN-GA-2008-214699).

Journal References:

Jinqiu Sang, Hongmei Hu, Chengshi Zheng, Guoping Li, Mark E. Lutman, Stefan Bleeck. Evaluation of the Sparse Coding Shrinkage Noise Reduction Algorithm in Normal Hearing and Hearing Impaired Listeners. Hearing Research (Vol. 310, pp. 36-47, APR 2014). DOI: 10.1016/j.heares.2014.01.006