| 摘要 |
PURPOSE: An assistive robot with an emotional speech synthesis function and an emotional speech synthesis method for the assistive robot are provided to change a sentence final syllable according to a sentence type, and to add expressions on the basis of a rhythm module. CONSTITUTION: An emotional speech synthesis method for the assistive robot(800) comprises following steps. A private voice is analyzed, and the character of a personal emotional prosody structure is extracted. The emotional prosody structure is saved in a voice database(890). An input text is determined according to situation information about the surrounding environment. The sentence type of the input text is identified. The personal emotional prosody structure is searched from the database. The input text is converted into an emotionless speech, and the emotionless speech is modified on the basis of the personal emotional prosody structure corresponding to a desirable emotion. The emotional voice corresponding to the speaker is generated, and the sentence final syllable is modified according to the sentence type. [Reference numerals] (810) Unit for extracting the characteristic of an emotional prosody structure; (820) Unit for modifying the sentence final syllable; (830) TTS system; (860) Unit for modifying the syllable level; (870) Unit for modifying the international phrase level; (880) Unit for modifying the sentence level; (890) Voice database; (AA) Data set; (BB) Emotional prosody structure for individuals; (CC) Input text; (DD) Voice without emotions; (EE) Emotional voice; (FF) Target emotion; (GG) Text type |
