Detecting a targeted voice style in an audiobook using voice quality features
2012 (English)In: Acoustics, Speech and Signal Processing (ICASSP), 2012 IEEE International Conference on, 2012, 4593-4596 p.Conference paper (Refereed)Text
Audiobooks are known to contain a variety of expressive speaking styles that occur as a result of the narrator mimicking a character in a story, or expressing affect. An accurate modeling of this variety is essential for the purposes of speech synthesis from an audiobook. Voice quality differences are important features characterizing these different speaking styles, which are realized on a gradient and are often difficult to predict from the text. The present study uses a pa- rameter characterizing breathy to tense voice qualities using features of the wavelet transform, and a measure for identifying creaky seg- ments in an utterance. Based on these features, a combination of supervised and unsupervised classification is used to detect the re- gions in an audiobook, where the speaker changes his regular voice quality to a particular voice style. The target voice style candidates are selected based on the agreement of the supervised classifier en- semble output, and evaluated in a listening test.
Place, publisher, year, edition, pages
2012. 4593-4596 p.
Engineering and Technology
IdentifiersURN: urn:nbn:se:kth:diva-185520DOI: 10.1109/ICASSP.2012.6288941ISI: 000312381404166ScopusID: 2-s2.0-84867584684ISBN: 978-1-4673-0046-9OAI: oai:DiVA.org:kth-185520DiVA: diva2:922794
IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP. MAR 25-30, 2012.
QC 201604262016-04-252016-04-212016-06-02Bibliographically approved