A speech reproduction system using flexible time-axis and speech database for researchers

研究了柔性时间轴再现系统的话率估计技术和话率转换的信号处理技术。在研究期间，拓展了软件研究环境。在这个扩展的环境中。对各个处理单元的信号处理进行了分解和描述，将各个处理单元连接起来，形成了一种新的信号处理方法。针对采用FPGA的系统，在Xilinx公司的评估板上实现了该系统的再现。此外，本研究还建立了专门的语音数据库，供全国各地的研究人员免费使用。在本学术研究期结束时，该数据库有2216个句子，并已向公众开放。

Speaking rate estimation techniques for a reproduction system using flexible time-axis and signal processing techniques for a speaking rate conversion are studied. In the study period, the software research environment is expanded. In this expanded environment. signal processing is decomposed and described for every processing element, and a new signal processing method can be made by connecting elements. About the system with FPGA, the proposed method of a reproduction system was implemented on the evaluation board of Xilinx. Moreover, in this research, construction of a special speech database is also performed so that researchers all over the country can use free. At the time of the end of this academic study period, the database has 2216 sentences, and has already opened to the public.