| 摘要 |
<P>PROBLEM TO BE SOLVED: To provide a sound signal processing device capable of suitably extracting a main sound from a mixed sound of the main sound and an unnecessary sound (for example, a covering sound and a reverberant sound). <P>SOLUTION: A sound signal processing device divides two signals consisting of a mixed sound signal, which is a signal of a time domain of the mixed sound including a first sound and a second sound, and a target sound signal, which is a signal of a time domain of a sound including a sound equivalent to at least the second sound, all or a part of which have a temporal interrelation, into plural frequency bands respectively and extracts the first sound signal included in the mixed sound signal based on a level ratio of the two signals calculated for each frequency. Since the first sound is extracted from the mixed sound with a focus on a frequency characteristics and the level ratio, the first sound can be easily extracted with excellent sound quality without subtracting an artificially generated waveform on a temporal axis or cancelling by spatial composition with a reverse phase wave. <P>COPYRIGHT: (C)2012,JPO&INPIT |
