Automatic glottal inverse filtering with non-negative matrix factorization

Research output: Chapter in Book/Report/Conference proceedingConference contributionScientificpeer-review

1 Citation (Scopus)

Abstract

This study presents an automatic glottal inverse filtering (GIF) technique based on separating the effect of the glottal main excitation from the impulse response of the vocal tract. The proposed method is based on a non-negative matrix factorization (NMF) based decomposition of an ultra short-term spectrogram of the analyzed signal. Unlike other state-of-theart GIF techniques, the proposed method does not require estimation of glottal closure instants. The proposed method was objectively evaluated with two test sets of continuous synthetic speech created with a glottal vocoding analysis/synthesis procedure. When compared to a set of reference GIF methods, the proposed NMF technique shows improved estimation accuracy especially for male voices.

Original languageEnglish
Title of host publicationProceedings of the Annual Conference of the International Speech Communication Association
Subtitle of host publicationInterspeech'16, San Francisco, USA, Sept. 8-12, 2016
PublisherInternational Speech Communication Association
Pages1039-1043
Number of pages5
Volume08-12-September-2016
ISBN (Electronic)978-1-5108-3313-5
DOIs
Publication statusPublished - 2016
MoE publication typeA4 Article in a conference publication
EventInterspeech - San Francisco, United States
Duration: 8 Sept 201612 Sept 2016
Conference number: 17

Publication series

NameProceedings of the Annual Conference of the International Speech Communication Association
PublisherInternational Speech Communication Association
ISSN (Print)1990-9770
ISSN (Electronic)2308-457X

Conference

ConferenceInterspeech
Country/TerritoryUnited States
CitySan Francisco
Period08/09/201612/09/2016

Keywords

  • Glottal inverse filtering
  • Nonnegative matrix factorization
  • Speech analysis

Fingerprint

Dive into the research topics of 'Automatic glottal inverse filtering with non-negative matrix factorization'. Together they form a unique fingerprint.

Cite this