| 摘要 |
PROBLEM TO BE SOLVED: To solve the problem, in which sufficient sound source separation used to be unachievable in a conventional device.SOLUTION: An audio signal processor includes: a frequency area conversion part 11 for generating a plurality of frequency area information from a plurality of audio input signals acquired at different positions; a relative value calculation part 12 for calculating a relative value between time frequency components included in own frequency area information and time frequency components included in the other frequency area information, for each frequency area information; a mask generation part 16 for comparing an enhanced range set on the basis of a relative value threshold stored in advance with the relative value to generate a time frequency mask for reducing the value of the frequency area information corresponding to the relative value outside the enhanced range; a mask multiplication part 17 for multiplying the frequency area information by the time frequency mask to generate enhanced frequency area information; and a time area conversion part 18 for converting the enhanced frequency area information into an audio output signal expressed as the information of time area. |
