Speech recognition systems that operate in real world environments have to be robust against additive noises. In this work, a technique that uses a voicing-dependent exponent in the computation of the filter-bank parameters to improve their robustness to additive noise is presented. Speech recognition experiments with the Aurora 1.0 database and recognition setup are reported to show the potential of this new technique.
Cite as: Macho, D., Nadeu, C. (2000) Use of voicing information to improve the robustness of the spectral parameter set. Proc. 6th International Conference on Spoken Language Processing (ICSLP 2000), vol. 1, 766-769, doi: 10.21437/ICSLP.2000-190
@inproceedings{macho00_icslp,
author={Dusan Macho and Climent Nadeu},
title={{Use of voicing information to improve the robustness of the spectral parameter set}},
year=2000,
booktitle={Proc. 6th International Conference on Spoken Language Processing (ICSLP 2000)},
pages={vol. 1, 766-769},
doi={10.21437/ICSLP.2000-190}
}