摘要 |
PROBLEM TO BE SOLVED: To enable adjusting a speech speed so that contents of an utterance in a synthesized speech can be easily understood.SOLUTION: In speech speed data generation processing, speech text data TD is acquired (S110), the acquired speech text data TD is morphologically analyzed (S120). Familiarity is identified for each morpheme (word) identified by morphological analysis (S130). Furthermore, speech speed data is generated so that a word representing the familiarity is low, of which a ratio of a time required for reading out the word occupying a time required for reading out entire information becomes longer (S170, S200). In familiarity update processing, word familiarity data is updated so that a morpheme mo (word) with high appearance frequency through contents viewed by a user, of which familiarity becomes high. |