<?xml version="1.0" encoding="utf-8"?>
<export-example>
  <doc>
    <id>4483</id>
    <completedYear/>
    <publishedYear>2015</publishedYear>
    <thesisYearAccepted/>
    <language>eng</language>
    <pageFirst>1201</pageFirst>
    <pageLast>1205</pageLast>
    <pageNumber/>
    <edition/>
    <issue/>
    <volume/>
    <type>conferenceobject</type>
    <publisherName>ISCA</publisherName>
    <publisherPlace/>
    <creatingCorporation/>
    <contributingCorporation/>
    <belongsToBibliography>0</belongsToBibliography>
    <completedDate>--</completedDate>
    <publishedDate>--</publishedDate>
    <thesisDateAccepted>--</thesisDateAccepted>
    <title language="eng">HMM training strategy for incremental speech synthesis</title>
    <abstract language="deu">Incremental speech synthesis aims at delivering the synthetic voice while the sentence is still being typed. One of the main challenges is the online estimation of the target prosody from a partial knowledge of the sentence's syntactic structure. In the context of HMM-based speech synthesis, this typically results in missing segmental and suprasegmental features, which describe the linguistic context of each phoneme. This study describes a voice training procedure which integrates explicitly a potential uncertainty on some contextual features. The proposed technique is compared to a baseline approach (previously published), which consists in substituting a missing contextual feature by a default value calculated on the training set. Both techniques were implemented in a HMM-based Text-To-Speech system for French, and compared using objective and perceptual measurements. Experimental results show that the proposed strategy outperforms the baseline technique for this language.</abstract>
    <parentTitle language="eng">Interspeech 2015, Dresden, Germany, 6-10 September 2015</parentTitle>
    <identifier type="doi">10.21437/Interspeech.2015-304</identifier>
    <enrichment key="opus.import.date">2022-06-20T18:36:30+00:00</enrichment>
    <enrichment key="opus.source">sword</enrichment>
    <enrichment key="opus.import.user">importuser</enrichment>
    <enrichment key="BegutachtungStatus">peer-reviewed</enrichment>
    <licence>Keine Lizenz - Es gilt das deutsche Urheberrecht: § 53 UrhG</licence>
    <author>Maël Pouget</author>
    <author>Thomas Hueber</author>
    <author>Gérard Bailly</author>
    <author>Timo Baumann</author>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>HMM-based speech synthesis</value>
    </subject>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>incremental</value>
    </subject>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>TTS</value>
    </subject>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>HTS</value>
    </subject>
    <subject>
      <language>eng</language>
      <type>uncontrolled</type>
      <value>prosody</value>
    </subject>
    <collection role="institutes" number="FakIM">Fakultät Informatik und Mathematik</collection>
    <collection role="othpublikationsherkunft" number="">Externe Publikationen</collection>
    <collection role="othforschungsschwerpunkt" number="16315">Information und Kommunikation</collection>
  </doc>
</export-example>
