Amplitude compression techniques have been used in the hearing aids for the sensory-neural impaired. In this paper, we present a DSP-based amplitude compressor which is based on the conventional filter-bank method, but can realize the relatively smoothed compression characteristics without the FFT method. The system has two kinds of filter-banks. The speech spectrum is estimated using the first bank and a desired compressor is generated at the second one which is similar to FIR filter used in the wide-band compression. Moreover, by estimating the spectrum using the output of the first filter-bank, it is possible to compress adaptively the signal according to the temporal variations of the speech features. We describe the principles of such compression system and its DSP implementation, and show that it is the simplified compression for the DSP realization, in this paper.
Cite as: Veda, Y., Agawa, T., Watanabe, A. (1994) A DSP-based amplitude compressor for digital hearing AIDS. Proc. 3rd International Conference on Spoken Language Processing (ICSLP 1994), 2011-2014, doi: 10.21437/ICSLP.1994-505
@inproceedings{veda94_icslp,
author={Yuichi Veda and Takayuki Agawa and Akira Watanabe},
title={{A DSP-based amplitude compressor for digital hearing AIDS}},
year=1994,
booktitle={Proc. 3rd International Conference on Spoken Language Processing (ICSLP 1994)},
pages={2011--2014},
doi={10.21437/ICSLP.1994-505},
issn={2958-1796}
}