| 摘要 |
PROBLEM TO BE SOLVED: To separate a mixed voice which is generated by a various kinds of sound sources, and to mix again the separated voice by a desired ratio, by using microphones having different characteristics. SOLUTION: A voice processing device 10 includes: an input correction section 104 for correcting difference of characteristics of a first input voice which is input from a first input device, and a characteristic of a second input voice which is input from a second input device, the characteristic being different from the first input voice; a voice separation section 112 for separating the first input voice corrected by an input correction section and the second input voice into a plurality voices; a sound voice classification estimation section 122 for estimating a voice classification of the plurality voices which are separated by the voice separation section; a mixing ratio calculation section 120 for calculating a mixing ratio of each voice according to the voice classification which is estimated by the voice classification estimation section; and a voice mixing section 124 for mixing the plurality voices which are separated by the voice separation section with the mixing ratio calculated by the mixing ratio calculation section. COPYRIGHT: (C)2010,JPO&INPIT |
