Method and apparatus for sound processing

1. A sound processing method performed by a sound processing apparatus, the sound processing method comprising:

extracting a desired signal from among signals generated by a plurality of sound sources collected through a plurality of channels using a plurality of microphones,

the extracting the desired signal comprising:

separating a voice signal for each sound source based on a sound signal by using a multi-channel blind source separation method based on an independent vector analysis;

comparing power values of off-diagonal elements of a full frequency band for each channel of an adaptive filter in a sound source separator; and

extracting the desired signal based on a magnitude of a power value of a voice signal generated by a sound source and the sum of off-diagonal values of the adaptive filter.

2. The sound processing method according to claim 1 , further comprising performing a time to frequency transformation.

3. The sound processing method according to claim 1 , further comprising at least one of:

removing an operation noise of a tone component generated by an operation of a device and inputted to one of the plurality of microphones; and

removing white noise generated by the operation of the device.

4. The sound processing method according to claim 3 , wherein the removing the operation noise of the tone component comprises removing noise by using a noise source sensor and an adaptive line enhancer (ALE).

5. The sound processing method according to claim 4 , wherein at least one of a microphone, an acceleration sensor, and a vibration sensor is used as the noise source sensor to remove the noise.

6. The sound processing method according to claim 3 , wherein the removing the white noise comprises removing the white noise by using multi-channel stationary noise reduction.

7. The sound processing method according to claim 1 , further comprising removing residual variation noise from the desired signal.

8. A sound processing apparatus configured to process a sound signal, comprising:

a memory storing one or more instructions; and

one or more processors configured to execute the one or more instructions to:

separate a desired signal from among signals generated by a plurality of sound sources collected through a plurality of channels using a plurality of microphones,

wherein the separating the desired signal comprises performing multi-channel blind source separation so as to separate sound sources based on an independent vector analysis, and extracting the desired signal based on a magnitude of a power value of a voice signal generated by the sound source and the sum of off-diagonal values of an adaptive filter,

wherein the sound source separator the extracting the desired signal comprises:

separating a voice signal for each sound source based on a sound signal; and

extracting a desired voice signal from the separated voice signals,

wherein the extracting the desired voice signal further comprises extracting the desired voice signal by comparing power values of off-diagonal elements of a full frequency band of the adaptive filter.

9. The sound processing apparatus according to claim 8 , wherein the one or more processors are further configured to execute the one or more instructions to:

perform a time to frequency transformation.

10. The sound processing apparatus according to claim 8 ,

wherein the one or more processors are further configured to execute the one or more instructions to:

remove an operation noise of a tone component generated by an operation of a device and inputted to one of the plurality of microphones; or

remove white noise generated by the operation of the device.

11. The sound processing apparatus according to claim 10 , wherein the removing the operation noise comprises removing the operation noise by using a noise source sensor and an adaptive line enhancer.

12. The sound processing apparatus according to claim 10 , wherein the removing the operation noise comprises removing the operation noise by using at least one of a microphone, an acceleration sensor, and a vibration sensor, as a noise source sensor.

13. The sound processing apparatus according to claim 10 , wherein the removing the white noise comprises removing the white noise by using multi-channel stationary noise reduction.

14. The sound processing method according to claim 8 , wherein the one or more processors are further configured to execute the one or more instructions to:

remove residual variation noise from the desired voice signal.

15. A sound processing apparatus for extracting a desired voice signal from a sound signal in which an operation noise of a device and other voices are mixed, the sound processing apparatus comprising:

a plurality of microphones configured to allow a sound signal to be collected through a plurality of channels;

a memory storing one or more instructions; and

one or more processors configured to execute the one or more instructions to:

remove noise from the sound signal; and

perform multi-channel blind source separation so as to separate sound sources in the noise-removed sound signal based on an independent vector analysis,

wherein performing the multi-channel blind source separation comprises extracting a desired voice signal from the channel having a smallest sum of off-diagonal values of an adaptive filter when an intensity of a desired voice is larger than an intensity of the other voices.