HMM-based speech enhancement using explicit gain modeling
2006 (English)In: 2006 IEEE International Conference on Acoustics, Speech and Signal Processing, 2006, 161-164 p.Conference paper (Refereed)
We propose a hidden Markov model (HMM) based speech enhancement method using explicit modeling of speech and noise gains. The gains are considered to be stochastic variables in an HMM framework. The speech gain models the energy variations of speech phones, typically due to differences in pronunciation and/or different vocalizations of individual speakers. The noise gain helps to improve the tracking of the time-varying energy of non-stationary noise. The time-varying parameters of the gain models are estimated on-line using the recursive expectation maximization (EM) algorithm. The performance of the proposed enhancement system is evaluated through both objective and subjective tests. The experimental results confirm the advantage of explicit gain modeling, particularly for non-stationary noise sources.
Place, publisher, year, edition, pages
2006. 161-164 p.
, International Conference on Acoustics Speech and Signal Processing (ICASSP), ISSN 1520-6149
IdentifiersURN: urn:nbn:se:kth:diva-36125DOI: 10.1109/ICASSP.2006.1659982ISI: 000245559900041ScopusID: 2-s2.0-33947706834OAI: oai:DiVA.org:kth-36125DiVA: diva2:430534
31st IEEE International Conference on Acoustics, Speech and Signal Processing Toulouse, FRANCE, MAY 14-19, 2006
QC 201107112011-07-112011-07-082011-10-07Bibliographically approved