<?xml version="1.0" standalone="yes"?>
<Paper uid="C96-1051">
  <Title>Segmentation and Labelling of Slovenian Diphone Inventories*</Title>
  <Section position="2" start_page="0" end_page="0" type="intro">
    <SectionTitle>
1 Introduction
</SectionTitle>
    <Paragraph position="0"> For the Slovenian language, several attempts were made in the past, where different aspects of a Slovenian text-to-speech synthesis (TI'S) system were covered (Dobnikar95). Nevertheless, none of them succeeded in building a complete system, providing high quality synthetic speech. In the Laboratory of Artificial Pemeption, we started on text-to-speech synthesis one year ago (Gros96). Here we describe the acquisition of an appropriate diphone inventory in a first version of our Slovenian TTS system, which is supposed to serve as a reference system for future improvements.</Paragraph>
    <Paragraph position="1"> We start with a brief overview of the different modules el' the Slovenian TTS system, then we go on to describe how the existing diphone inventory was obtained. The acquisition era labelled diphone inventory or adaptation of a speech synthesis system to synthesise new voices is manually intensive and prone to errors, therefore automatic procedures are required (Taylor91, Sehmidt93, Cosi91, Ottesen93). In section 4 we *This work was partly funded by the Commission of the European Community under COP-94 contract No 01634 (SQEL) explain how we intend to automatically derive additional diphone inventories for building new synthetic voices.</Paragraph>
  </Section>
class="xml-element"></Paper>