| 摘要 |
The present feature relates to a sound source separating device and method, and a program that allow a sound source to be separated at lower calculation cost. A communication unit receives the spatial frequency spectrum in a sound collection signal obtained when a microphone array collects the planar wave of a voice from a sound source. A spatial frequency mask generation unit generates a spatial frequency mask that masks a component in a prescribed region of a spatial frequency region on the basis of the spatial frequency spectrum. On the basis of the spatial frequency mask, a sound source separating unit extracts a component of the desired sound source from a spatial frequency spectrum and treats the desired component as an estimated sound source spectrum. This feature can be applied to a spatial frequency sound source separator. |
