King, Simon

Clark, Robert

Yamagishi, Junichi

Watts, Oliver Samuel

Engineering and Physical Sciences Research Council (EPSRC)

2013-10-22T14:32:59Z

2013-10-22T14:32:59Z

2013-07-02

This thesis introduces a general method for incorporating the distributional analysis of textual and linguistic objects into text-to-speech (TTS) conversion systems. Conventional TTS conversion uses intermediate layers of representation to bridge the gap between text and speech. Collecting the annotated data needed to produce these intermediate layers is a far from trivial task, possibly prohibitively so for languages in which no such resources are in existence. Distributional analysis, in contrast, proceeds in an unsupervised manner, and so enables the creation of systems using textual data that are not annotated. The method therefore aids the building of systems for languages in which conventional linguistic resources are scarce, but is not restricted to these languages. The distributional analysis proposed here places the textual objects analysed in a continuous-valued space, rather than specifying a hard categorisation of those objects. This space is then partitioned during the training of acoustic models for synthesis, so that the models generalise over objects' surface forms in a way that is acoustically relevant. The method is applied to three levels of textual analysis: to the characterisation of sub-syllabic units, word units and utterances. Entire systems for three languages (English, Finnish and Romanian) are built with no reliance on manually labelled data or language-specific expertise. Results of a subjective evaluation are presented.

http://hdl.handle.net/1842/7982

en

The University of Edinburgh

O.Watts, J. Yamagishi, and S. King. The role of higher-level linguistic features in HMM-based speech synthesis. In Proc. Interspeech, pages 841-844, Makuhari, Japan, Sept. 2010a.

O. Watts, J. Yamagishi, and S. King. Letter-based speech synthesis. In Proc. Speech Synthesis Workshop 2010, pages 317-322, Nara, Japan, Sept. 2010b.

O. Watts, J. Yamagishi, and S. King. Unsupervised continuous-valued word features for phrase-break prediction without a part-of-speech tagger. In Proc. Interspeech, Florence, Italy, Aug. 2011.

J. Yamagishi and O. Watts. The CSTR/EMIME HTS System for Blizzard Challenge. In Proc. Blizzard Challenge 2010, Sept. 2010.

unsupervised learning

vector space model

speech synthesis

TTS

text-to-speech

Unsupervised learning for text-to-speech synthesis

Thesis or Dissertation

Doctoral

PhD Doctor of Philosophy