Sound source separation system, sound source separation method, and acoustic signal acquisition device

Abstract

A sound source separation system separates a target sound and a disturbance sound coming from an arbitrary direction other than the direction the target sound comes from. The system includes different-directional-signal-group generators and a sensitive region formation unit. The generators each generate two or more combinations of spectra of signals each of which has a different directivity, using received sound signals of microphones. The sensitive region formation unit determines, for each frequency band, whether or not a relationship between powers of the spectra in each combination simultaneously satisfies conditions each defined for each combination, using two or more combinations of the spectra of the signals generated by the respective different-directional-signal-group generators, and performs multidimensional band selection of assigning power of a spectrum selected beforehand to a spectrum of the target sound to be separated, for a frequency band where the conditions are simultaneously satisfied.

1 - 111 . (canceled) 
     
     
         112 . A sound source separation system separating a target sound and a disturbance sound coming from an arbitrary direction other than the direction in which the target sound comes from, the sound source separation system comprising:
 a plurality of different-directional-signal-group generators each generating two or more combinations of spectra of a plurality of signals each of which has a different directivity, using received sound signals of a plurality of microphones; and   a sensitive region formation unit determining, for each frequency band, whether or not a relationship between powers of the spectra in each combination simultaneously satisfies a plurality of conditions each defined for each combination, using two or more combinations of the spectra of the plurality of signals generated by the respective different-directional-signal-group generators, and performing multidimensional band selection of assigning power of a spectrum selected beforehand to a spectrum of the target sound to be separated, for a frequency band where the plurality of conditions are simultaneously satisfied,   wherein each of a plurality of the different-directional-signal-group generators comprises at least two of the following three devices:
 a first device acquiring, on a time domain or a frequency domain, a difference between a received sound signal of one of the plurality of microphones and a received sound signal of one of the rest of the plurality of microphones that has undergone a delayed process, and generating a first signal; 
 a second device acquiring, on a time domain or a frequency domain, a difference between a received sound signal of the one of the rest of the plurality of microphones and a received sound signal of the one of the plurality of microphones that has undergone a delayed process, and generating a second signal; and 
 a third device acquiring, on a time domain or a frequency domain, a difference between a received sound signal of the one of the plurality of microphones and a received sound signal of the one of the rest of the plurality of microphones, and generating a third signal. 
   
     
     
         113 . The sound source separation system according to  claim 112 , wherein each of the plurality of the different-directional-signal-group generators further comprises an integration unit comparing each of the spectrum powers for each frequency band and performing a spectrum integration process of assigning inferior power to a spectrum of a fourth signal, using a spectrum of a first target sound superior signal and a spectrum of a second target sound superior signal after defining the first signal generated by the first device as the first target sound superior signal and defining the second signal generated by the second device as the second target sound superior signal. 
     
     
         114 . The sound source separation system according to  claim 112 , wherein the sensitive region formation unit sets a condition for each different-directional-signal-group generator as a condition that power of the spectrum of the fourth signal obtained by each integration unit is larger than power of the spectrum of the third signal obtained by each third device, and determines whether or not those conditions are simultaneously satisfied for each frequency band. 
     
     
         115 . A sound source separation system that separates a target sound and a disturbance sound coming from an arbitrary direction other than a direction from which the target sound comes, comprising
 a plurality of different-directional-signal-group generators each generating two or more combinations of spectra of a plurality of signals each of which has a different directivity, using received sound signals of a plurality of microphones,   wherein each of a plurality of the different-directional-signal-group generators comprises:
 a first device acquiring, on a time domain or a frequency domain, a difference between a received sound signal of one of the plurality of microphones and a received sound signal of one of the rest of the plurality of microphones that has undergone a delayed process, and generating a first signal; 
 a second device acquiring, on a time domain or a frequency domain, a difference between the received sound signal of the one of the rest of the plurality of microphones and the received sound signal of the one of the plurality of microphones that has undergone a delayed process, and generating a second signal; 
 a third device acquiring, on a time domain or a frequency domain, a difference between the received sound signal of the one of the plurality of microphones and the received sound signal of the one of the rest of the plurality of microphones, and generating a third signal; and 
 a fourth device performing spectral subtraction for obtaining a spectrum of a signal including the target sound or spectra of two signals each including the target sound by subtracting, a value produced as result of multiplying, by a coefficient, power of the same frequency band of a spectrum of a target sound inferior signal obtained by one or two of the first, second and third devices, from power of each frequency band of a spectrum of a target sound superior signal obtained by the rest of the first, second and third devices. 
   
     
     
         116 . The sound source separation system according to  claim 115 , further comprising an integration unit comparing powers for each frequency band using, for each of the different-directional-signal-group generators, the spectrum of the signal including a plurality of the target sounds obtained by the fourth device, and performing a spectrum integration process by assigning inferior power to a spectrum of a target sound superior signal.