Signal processing apparatus, signal processing method, and computer program product
A signal processing apparatus includes one or more processors. The processors acquire a plurality of observed signals acquired from a plurality of microphone groups each including at least one microphone selected from a plurality of microphones. The microphone groups include respective microphone combinations each including at least one microphone, the combinations are different from each other, and at least one of the microphone groups includes a plurality of microphones. The processors estimate a mask indicating occupancy for each of time frequency points of a sound signal of a space corresponding to the observed signal in a plurality of spaces, for each of the observed signals. The processors integrate masks estimated for the observed signals to generate an integrated mask indicating occupancy for each of time frequency points of a sound signal in a space determined based on the spaces.
1. A signal processing apparatus comprising:
one or more processors configured to:
acquire a plurality of observed signals from a plurality of microphone groups, each of the microphone groups comprising a microphone set comprising at least one microphone selected from a plurality of microphones, each microphone set of the microphone groups being different from each other, at least one of the plurality of microphone groups comprising two or more microphones of the plurality of microphones;
for each of the plurality of observed signals, estimate a mask indicating occupancy for each time frequency point of a sound signal of a space corresponding to an observed signal in a plurality of spaces; and
integrate the masks estimated for the plurality of observed signals to generate an integrated mask indicating occupancy for each of the time frequency points of a target sound signal in a target space determined based at least in part on the plurality of spaces.
2. The signal processing apparatus according to claim 1 , wherein each of the plurality of spaces corresponds to one of the plurality of microphone groups, and is determined with at least one of a direction in a plane determined with the plurality of microphones included in a corresponding microphone group and a direction perpendicular to the plane.
3. The signal processing apparatus according to claim 1 , wherein the one or more processors estimate the mask based at least in part on the observed signal using a neural network receiving input information and outputting occupancy for each of the time frequency points of the sound signal of a second space, the input information being one of the observed signal or a feature of the observed signal.
4. The signal processing apparatus according to claim 3 , wherein the neural network is a complex-valued neural network.
5. The signal processing apparatus according to claim 3 , wherein the one or more processors further configured to:
acquire parameters indicating at least part of the spaces and positions of the plurality of microphones included in the plurality of microphone groups, and
correct the input information based at least in part on the acquired parameters.
6. The signal processing apparatus according to claim 3 , wherein the one or more processors further configured to:
acquire parameters indicating at least part of the spaces and positions of the plurality of microphones included in the plurality of microphone groups, and
correct parameters of the neural network based at least in part on the acquired parameters.
7. The signal processing apparatus according to claim 1 , wherein the one or more processors generate the integrated mask with a mean value of the masks, product of the masks, or median of the masks.
8. A signal processing method comprising:
acquiring a plurality of observed signals from a plurality of microphone groups, each of the microphone groups comprising a microphone set comprising at least one microphone selected from a plurality of microphones, each microphone set of the plurality of microphone groups being different from each other, at least one of the plurality of microphone groups comprising two or more microphones of the plurality of microphones;
for each of the plurality of observed signals, estimating a mask indicating occupancy for each time frequency point of a sound signal of a space corresponding to an observed signal in a plurality of spaces; and
integrating masks estimated for the plurality of observed signals to generate an integrated mask indicating occupancy for each of the time frequency points of a target sound signal in a target space determined based at least in part on the plurality of spaces.
9. A computer program product having a non-transitory computer readable medium comprising programmed instructions, wherein the instructions, when executed by a computer, cause the computer to:
acquire a plurality of observed signals from a plurality of microphone groups, each of the microphone groups comprising a microphone set comprising at least one microphone selected from a plurality of microphones, each microphone set of the microphone set of the plurality of microphone groups being different from each other, at least one of the plurality of microphone groups comprising two or more microphones of the plurality of microphones;
for each of the plurality of observed signals, estimate a mask indicating occupancy for each time frequency point of a sound signal of a space corresponding to an observed signal in a plurality of spaces; and
integrate the masks estimated for the plurality of observed signals to generate an integrated mask indicating occupancy for each of the time frequency points of a target sound signal in a target space determined based at least in part on the plurality of spaces.