Show simple item record

dc.contributor.authorGOBL, CHRISTERen
dc.contributor.authorYANUSHEVSKAYA, IRENAen
dc.date.accessioned2014-02-28T15:59:14Z
dc.date.available2014-02-28T15:59:14Z
dc.date.issued2014en
dc.date.submitted2014en
dc.identifier.citationKane, J., Aylett, M., Yanushevskaya, I., Gobl, C., Phonetic feature extraction for context-sensitive glottal source processing, Speech Communication, 59, 2014, 10 - 21en
dc.identifier.otherYen
dc.identifier.urihttp://hdl.handle.net/2262/68180
dc.descriptionPUBLISHEDen
dc.description.abstractThe effectiveness of glottal source analysis is known to be dependent on the phonetic properties of its concomitant supraglottal features. Phonetic classes like nasals and fricatives are particularly problematic. Their acoustic characteristics, including zeros in the vocal tract spectrum and aperiodic noise, can have a negative effect on glottal inverse filtering, a necessary pre-requisite to glottal source analysis. In this paper, we first describe and evaluate a set of binary feature extractors, for phonetic classes with relevance for glottal source analysis. As voice quality classification is typically achieved using feature data derived by glottal source analysis, we then investigate the effect of removing data from certain detected phonetic regions on the classification accuracy. For the phonetic feature extraction, classification algorithms based on Artificial Neural Networks (ANNs), Gaussian Mixture Models (GMMs) and Support Vector Machines (SVMs) are compared. Experiments demonstrate that the discriminative classifiers (i.e. ANNs and SVMs) in general give better results compared with the generative learning algorithm (i.e. GMMs). This accuracy generally decreases according to the sparseness of the feature (e.g., accuracy is lower for nasals compared to syllabic regions). We find best classification of voice quality when just using glottal source parameter data derived within detected syllabic regions.en
dc.format.extent10en
dc.format.extent21en
dc.language.isoenen
dc.relation.ispartofseriesSpeech Communicationen
dc.relation.ispartofseries59en
dc.rightsYen
dc.subjectVoice qualityen
dc.subjectExpressive speechen
dc.subjectSpeech synthesisen
dc.subjectPhonation typeen
dc.subjectGlottal sourceen
dc.titlePhonetic feature extraction for context-sensitive glottal source processingen
dc.typeJournal Articleen
dc.type.supercollectionscholarly_publicationsen
dc.type.supercollectionrefereed_publicationsen
dc.identifier.peoplefinderurlhttp://people.tcd.ie/yanushien
dc.identifier.peoplefinderurlhttp://people.tcd.ie/cegoblen
dc.identifier.rssinternalid91309en
dc.identifier.doihttp://dx.doi.org/10.1016/j.csl.2014.03.002en
dc.rights.ecaccessrightsOpenAccess
dc.subject.TCDTagSpeech processing/technologyen
dc.contributor.sponsorScience Foundation Ireland (SFI)en
dc.contributor.sponsorGrantNumber09/IN.1/I2631en


Files in this item

Thumbnail
Thumbnail

This item appears in the following Collection(s)

Show simple item record