Bayesian learning for time-varying linear prediction of speech
2015 (English)In: 2015 23rd European Signal Processing Conference, EUSIPCO 2015, Institute of Electrical and Electronics Engineers (IEEE), 2015, 325-329 p.Conference paper (Refereed)Text
We develop Bayesian learning algorithms for estimation of time-varying linear prediction (TVLP) coefficients of speech. Estimation of TVLP coefficients is a naturally underdeter-mined problem. We consider sparsity and subspace based approaches for dealing with the corresponding underde-termined system. Bayesian learning algorithms are developed to achieve better estimation performance. Expectation-maximization (EM) framework is employed to develop the Bayesian learning algorithms where we use a combined prior to model a driving noise (glottal signal) that has both sparse and dense statistical properties. The efficiency of the Bayesian learning algorithms is shown for synthetic signals using spectral distortion measure and formant tracking of real speech signals. © 2015 EURASIP.
Place, publisher, year, edition, pages
Institute of Electrical and Electronics Engineers (IEEE), 2015. 325-329 p.
Bayesian learning, expectation-maximization, sparsity, Time-varying linear prediction, Algorithms, Forecasting, Maximum principle, Signal processing, Estimation performance, Expectation - maximizations, Expectation Maximization, Linear prediction, Spectral distortions, Statistical properties, Learning algorithms
Electrical Engineering, Electronic Engineering, Information Engineering
IdentifiersURN: urn:nbn:se:kth:diva-186794DOI: 10.1109/EUSIPCO.2015.7362398ISI: 000377943800066ScopusID: 2-s2.0-84963983947ISBN: 9780992862633OAI: oai:DiVA.org:kth-186794DiVA: diva2:928538
23rd European Signal Processing Conference, EUSIPCO 2015, 31 August 2015 through 4 September 2015
QC 201605162016-05-162016-05-132016-07-15Bibliographically approved