Modern telecommunication technology is present in many areas of everyday life. However for systems carrying out such problems as sound coding, automatic speech recognition, voice identification, etc. is required a high quality signal on an input. This paper describes one-channel neuromorphic motivated speech enhancement system, which explores properties of modulation spectrum of human. It can be used as input signal preprocessor of such systems, with the purpose to increase of its quality. The new adaptive technique of filtering of spectral envelopes is suggested in this paper. This approach allows more accuracy tune modulation filter, and appropriate to speech enhancement systems. Criteria of a choice of controlled modulation filter parameters are submitted. Also work performances of speech enhancement system are resulted.
Cite as: Shadevsky, A., Petrovsky, A.A. (2004) Implementation of time-varying modulation filter in speech enhancement system. Proc. 9th Conference on Speech and Computer (SPECOM 2004), 190-194
@inproceedings{shadevsky04_specom,
author={A. Shadevsky and Alexander A. Petrovsky},
title={{Implementation of time-varying modulation filter in speech enhancement system}},
year=2004,
booktitle={Proc. 9th Conference on Speech and Computer (SPECOM 2004)},
pages={190--194}
}