CLCNet - Audio Samples

Audio samples accompanying the paper CLCNet: Deep learning-based noise reduction for hearing aids using complex linear coding (PDF). All samples are from the test set. The maximum attenuation of both Wiener Filter and Complex Linear Coding is limited to 14dB.

Contents:

• English sample video
• German samples
• English samples

English sample video

German samples

Noisy

Sample 1 (SNR=0): Sample 2 (SNR=-5): Sample 3 (SNR=10):

Wiener Filter (WF)

Sample 1 (SNR=0): Sample 2 (SNR=-5): Sample 3 (SNR=10):

Complex Linear Coding (CLC)

Sample 1 (SNR=0): Sample 2 (SNR=-5): Sample 3 (SNR=10):

English samples

The English samples are also from the EUROM database. Howerver, neither english samples nor the speakers of these samples were used for training. Note that the loudness of the english denoised samples was slightly improved.

Noisy

Sample 1 (SNR=0): Sample 2 (SNR=-5): Sample 3 (SNR=10):

Wiener Filter (WF)

Sample 1 (SNR=0): Sample 2 (SNR=-5): Sample 3 (SNR=10):

Complex Linear Coding (CLC)

Sample 1 (SNR=0): Sample 2 (SNR=-5): Sample 3 (SNR=10):

[1]: Aubreville, Marc, et al. "Deep Denoising for Hearing Aid Applications." 2018 16th International Workshop on Acoustic Signal Enhancement (IWAENC). IEEE, 2018.

[2]: Schroeter, Hendrik, et al. "CLCNet: Deep learning-based noise reduction for hearing aids using complex linear coding" International Conference on Acoustics, Speech, and Signal Processing (ICASSP) 2020.