SPEECH SYNTHESIS DEVICE

PROBLEM TO BE SOLVED: To provide a speech synthesis device with which it is possible to reduce the cost and time needed for creating and sending out the synthetic sound of a plurality of different programs.SOLUTION: A speech synthesis device 100 comprises: input means 200 for successively inputting...

Full description

Saved in:
Bibliographic Details
Main Authors SEGI HIROYUKI, IIJIMA SHINICHI, KURATA JUN, KOTAKI KUNIHIRO, KURIHARA KIYOSHI, HOSOYA HIROO, SENOO MASUMI, SHIBUYA TOMOHIRO
Format Patent
LanguageEnglish
Japanese
Published 18.01.2016
Subjects
ACOUSTICS
MUSICAL INSTRUMENTS
PHYSICS
SPEECH ANALYSIS OR SYNTHESIS
SPEECH OR AUDIO CODING OR DECODING
SPEECH OR VOICE PROCESSING
SPEECH RECOGNITION
Online AccessGet full text

Cover

More Information
Summary:PROBLEM TO BE SOLVED: To provide a speech synthesis device with which it is possible to reduce the cost and time needed for creating and sending out the synthetic sound of a plurality of different programs.SOLUTION: A speech synthesis device 100 comprises: input means 200 for successively inputting raw data including at least one of a sentence, a word, and a symbol; a plurality of synthetic sound creation means 300 for successively creating a synthetic sound corresponding to the raw data and estimating the length of silence interposed between the synthetic sounds using a predetermined method or generating silence, and creating a speech file in which the synthetic sound and information on the length of silence interposed between synthetic sounds are stored; and synthetic sound send-out means 400 for reproducing a synthetic speech prepared by extracting a necessary synthetic sound from the speech file in a prescribed time frame, changing the length of the synthetic sound or the length of silence interposed between synthetic sounds, and adjusting the length to fit within the prescribed time frame, each of the plurality of synthetic sound creation means 300 creating the speech file in a common data format and sharing the synthetic sound send-out means 400. 【課題】異なる複数の番組の合成音を作成及び送出するために要するコストや手間を低減できる音声合成装置を提供する。【解決手段】音声合成装置１００は、文、単語及び記号のうちの少なくともいずれかを含む原データを順次入力する入力手段２００と、原データに対応した合成音を順次作成すると共に合成音間に挟む無音の長さを予め定められた方法で推定又は無音を生成し、当該合成音と合成音間に挾む無音の長さの情報を格納した音声ファイルを作成する複数の合成音作成手段３００と、所定時間枠において音声ファイルから必要な合成音を取り出して合成音の長さや合成音間に挾む無音の長さを変更して所定時間枠に収まるように調整した合成音声を再生する合成音送出手段４００と、を備え、複数の合成音作成手段３００は、音声ファイルを共通のデータ形式でそれぞれ作成し、合成音送出手段４００を共用する。【選択図】図１
Bibliography:Application Number: JP20140129160