| 摘要 |
PROBLEM TO BE SOLVED: To make it possible to generate synthetic speech or sound adaptable to a use environment desired by user without adjusting individual synthetic speech generation variables by user, by inputting environmental sound in a place and status for the use. SOLUTION: Reproduced sound is created in which environmental sound, etc., generated in a environment for using synthetic speech are superimposed on a speech to be evaluated. And it is made possible for a user to hear reproduced sound outputted from acoustic output means 7, 8 and evaluate the reproduced sound heard from a hearing evaluation function 1005. By using the evaluation result, data are totalized, statistically processed, and further, the result of the processing is displayed on a screen. By using the processed result, an utterance velocity, an utterance pressure, an utterance basic frequency, etc., are changed and re-synthesized speech, etc., are outputted according to the changed variables. Namely, user evaluates presented reproduced sound in conformity with the set evaluation items. The evaluation result is inputted to the device, and then, a synthesized speech having obtained an optimal value as the evaluation result is generated and outputted. |
