Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

Direct Input for Japanese Word Accent Registration

IP.com Disclosure Number: IPCOM000116109D
Original Publication Date: 1995-Aug-01
Included in the Prior Art Database: 2005-Mar-30
Document File: 2 page(s) / 74K

Publishing Venue

IBM

Related People

Hashimoto, Y: AUTHOR [+3]

Abstract

This article describes a method of inputting Japanese word accents that can be used for word registration in Japanese text-to-speech.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 52% of the total text.

Direct Input for Japanese Word Accent Registration

      This article describes a method of inputting Japanese word
accents that can be used for word registration in Japanese
text-to-speech.

      <Background> In Japanese, each word has an accent.  This word
accent is located on one of the CV syllables, except on the
Japanese-unique syllables /H/,/Q/, and /N/.  On the other hand, the
written Japanese representation is described by kana character
sequences.  Each CV syllable corresponds to a kana character, except
/SHA/, /SHU/, /SHO/, /MYA/, /MYU/, /MYO/, /NYA/, /NYU/, /NYO/, and so
on.  These exceptional CV syllables are represented by two kana
characters.  When the positions of word accents are assigned by the
conventional method, the accents are shown by the number of CV
syllables from the beginning of the word.  However, the above two
kinds of the exceptional syllables often cause mistakes.

      Fig. 1 shows the relationship between syllables and kana
characters.  The word /ARUFABEQTO/ consists of six syllables,
/A/RU/FA/BE/Q/TO/, and is represented by seven kana characters.  In
this example, /Q/ is a non-accented syllable and /FA/ consists of two
kana characters.  The accent is located on the CV syllable /BE/,
which is the fourth syllable, located on the fifth character from the
beginning of the word.  However, users who are not familiar with
these relations between kana characters and syllables cannot tell the
location of the accent by counting...