SPEECH SYNTHESIS METHOD, SPEECH SYNTHESIS DEVICE, PROGRAM, RECORDING MEDIUM, METHOD AND APPARATUS FOR GENERATING CONSTRAINT INFORMATION AND ROBOT APPARATUS

摘要

PROBLEM TO BE SOLVED: To add feeling to synthesized speech while the prosodic feature of a language is maintained.;SOLUTION: In the speech synthesis device 200, a language processing part 201 generates a string of pronunciation marks from a text, and a prosodic data generating part 202 creates prosodic data, expressing parameters such as the time duration of phonemes, pitch and sound volume, on the basis of the string of pronunciation marks. A constraint information generating part 203 is fed with the prosodic data and with the string of pronunciation marks to generate constraint information which limits changes in the parameters, adding the constraint information to the prosodic data so as to maintain the prosodic feature of a speech sentence. An emotion filter 204 is fed with the prosodic data added with the constraint information, changes the parameters of the prosodic data within the constraint in response to given feeling state information, and a waveform generating part 205 synthesizes a speech waveform on the basis of the prosodic data with the changed parameters.;COPYRIGHT: (C)2003,JPO