Separation of target acoustic signals in a multi-transducer arrangement

• US 7,366,662 B2
• Filed: 08/09/2006
• Issued: 04/29/2008
• Est. Priority Date: 07/22/2004
• Status: Active Grant

• Current Assignee: Qualcomm, Inc., Regents of the University of California (University of California)
• Original Assignee: SoftMax Incorporated (Qualcomm, Inc.)
• Inventors: Visser, Erik, Lee, Te-Won
• Primary Examiner(s): Storm; Donald L.
• Application Number:
  US11/463,376
• Publication Number:
  US 20070038442A1
• Time in Patent Office: 629 Days
• Field of Search: None
• US Class Current: 704/227
• CPC Class Codes: G10L 2021/02161 Number of inputs available containing the signal or the noise to be suppressed
• CPC Class Codes: G10L 2021/02165 Two microphones, one receiving mainly the noise signal and the other one mainly the speech signal
• CPC Class Codes: G10L 21/0208 Noise filtering
• CPC Class Codes: G10L 21/0272 Voice signal separating
• CPC Class Codes: H04R 2430/25 Array processing for suppression of unwanted side-lobes in directivity characteristics, e.g. a blocking matrix
• CPC Class Codes: H04R 3/005 for combining the signals of two or more microphones specially adapted for hearing aids H04R25/407

Abstract

The present invention provides a process for separating a good quality information signal from a noisy acoustic environment. The separation process uses a set of at least two spaced-apart transducers to capture noise and information components. The transducer signals, which have both a noise and information component, are received into a separation process. The separation process generates one channel that is substantially only noise, and another channel that is a combination of noise and information. An identification process is used to identify which channel has the information component. The noise signal is then used to set process characteristics that are applied to the combination signal to efficiently reduce or eliminate the noise component. In this way, the noise is effectively removed from the combination signal to generate a good qualify information signal. The information signal may be, for example, a speech signal, a seismic signal, a sonar signal, or other acoustic signal.

Claims

1. A speech separation process, comprising:

providing a plurality of microphones with respect to a speech source so that each respective microphone generates a signal having a speech component and a noise component in different mixing ratios;

receiving each of the signals generated by the microphones into a separation process;

separating the received signals into a first channel and a second channel using a blind-source separation process, wherein one of the channels provides a noise signal that is substantially noise components and the other channel provides a combination signal that is a combination of noise components and speech components;

identifying which of the first or second channels has the combination signal;

processing the combination signal with the noise signal; and

generating a speech signal indicative of the speech from the speech source.

2. The speech separation process according to claim 1, further including:

measuring a level of speech components in the noise signal;

adjusting, responsive to the measurement, a gain setting for one of the signals generated by the microphones; and

reducing, responsive to the adjustment, the level of speech components in the noise signal.

3. The speech separation process according to claim 1, further including:

providing the plurality of microphones in a first arrangement where a first microphone is closer to the speech source and a second microphone is farther from the speech source;

providing a set of filters within the blind-source separation process;

setting, for the first arrangement, each of the filters with respective filter coefficients to facilitate channel separation; and

providing the plurality of microphones in a second arrangement where the second microphone is closer to the speech source and the first microphone is farther from the speech source; and

rearranging, for the second arrangement, the filter coefficients for the set of filters.

4. The speech separation process according to claim 1, wherein identifying which of the first or second channels has the combination signal further includes:

using a multi-dimensional criterion to determine whether the first channel or the second channel has a greater level of the speech components.

5. The speech separation process according to claim 4, wherein the multi-dimensional criterion includes a measurement of speech recognition accuracy.

6. The speech separation process according to claim 4, wherein the multi-dimensional criterion includes an energy measurement for the first channel or the second channel.

7. The speech separation process according to claim 4, wherein the multi-dimensional criterion includes a measurement of zero crossing rates for the first channel or the second channel.

8. The speech separation process according to claim 4, wherein the multi-dimensional criterion includes a measurement of voice activity detection (VAD) module.

9. The speech separation process according to claim 1, wherein identifying the combination signal further includes:

providing a priori information regarding the speech components; and

using the a priori information to select whether the first channel or the second channel has a greater level of the speech components.

10. The speech separation process according to claim 9, wherein the a-priori information includes an expected distance from one of the microphones to the speech source.

11. The speech separation process according to claim 9, wherein the a-priori information includes an expected frequency profile for the speech source.

12. The speech separation process according to claim 1, wherein the blind-source separation process comprises an independent component analysis process.

13. The speech separation process according to claim 1, wherein the blind-source separation process modulates the mathematical formulation of mutual information directly or indirectly through approximations.

14. The speech separation process according to claim 1, wherein the processing the combination signal with the noise signal comprises filtering the combination signal with a processing filter comprising coefficients at least partially derived from the noise signal.

15. The speech separation process according to claim 1, wherein processing the combination signal with the noise signal comprises processing the combination signal with a filter,

wherein the filter is designed based partly on the first channel and optionally the second channel, and

wherein the filter is designed based partly on a priori information related to the speech from the speech source and expected noise sources.

16. An electronic device for wirelessly transmitting a tranmitted signal, comprising:

a plurality of microphones, each microphone generating a signal having a noise component and a speech component;

a processor operating the steps of;

receiving the generated signals;

separating the received signals into a first and a second channel using a blind-source separation process, wherein one of the channels provides a noise signal comprising substantially only noise components and the other channel provides a combination signal

comprising both noise components and speech components;

identifying which of the first or second channels has the combination signal;

processing the combination signal with the noise signal; and

generating a speech signal indicative of a speaker'"'"'s speech; and

a transmitter for generating the transmitted signal, the transmitted signal comprising speech information indicative of the speech signal.

17. The device according to claim 16, further including:

a housing sized to be portable and to be held in a hand of the speaker; and

wherein said housing comprises the plurality of microphones.

18. The electronic device according to claim 17, wherein each of the microphones has a different direct audio path to an expected location of the speaker'"'"'s mouth.

19. The electronic device according to claim 17, wherein the housing comprises a candy bar shaped housing, a clamshell shaped housing, a mobile phone housing, or a walkie-talkie housing.

20. The electronic device according to claim 16, further including:

a hands-free accessory device; and

wherein at least one of the plurality of microphones is placed on the hands-free accessory device.

21. The electronic device according to claim 16, wherein the blind-source separation process comprises an independent component analysis process.

22. The electronic device according to claim 16, wherein the blind-source separation process modulates the mathematical formulation of mutual information directly or indirectly through approximations.

23. The device of claim 16, wherein said device is a wireless telephone.

24. A headset system, comprising:

a headset device comprising a plurality of microphones, each microphone generating a signal having a noise component and a speech component;

a processor operating the steps of;

receiving the generated signals;

separating the received signals into a first and a second channel using a blind-source separation process, wherein one of the channels provides a noise signal comprising substantially only noise components and the other channel provides a combination signal

comprising both noise components and speech components;

identifying which of the first or second channels has the combination signal;

processing the combination signal with the noise signal; and

generating a speech signal indicative of a speaker'"'"'s speech; and

an output line for transmitting the speech signal to a receiving device.

25. The headset system according to claim 24, wherein the blind-source separation process comprises an independent component analysis process.

26. The headset system according to claim 24, wherein the blind-source separation process modulates the mathematical formulation of mutual information directly or indirectly through approximations.

27. An information separation process, comprising:

receiving a first signal having an information component and a noise component, the first signal being indicative of a signal generated by a first transducer;

receiving a second signal having an information component and a noise component, the second signal being indicative of a signal generated by a second transducer;

receiving the first signal and the second signal into a separation process;

separating the received signals into a plurality of channels using a blind-source separation process, wherein one of the channels provides a noise signal that is substantially noise components and another channel provides a combination signal that is a combination of noise components and information components;

identifying the channel that has the combination signal;

processing the noise signal with the combination signal; and

generating an information signal.

28. The information separation process according to claim 27, further including processing steps for:

determining a level of information components in the noise signal;

adjusting a gain setting for one of the signals generated by the transducers; and

reducing, responsive to the adjustment, the level of information components in the noise signal.

29. The information separation process according to claim 27, wherein the information component includes a sonar signal, a seismic signal, or acoustic information produced by the operation of a mechanical device.

30. The information separation process according to claim 27, wherein the blind-source separation process comprises an independent component analysis process.

31. The information separation process according to claim 27, wherein the blind-source separation process modulates the mathematical formulation of mutual information directly or indirectly through approximations.