SYSTEMS, METHODS, APPARATUS, AND COMPUTER PROGRAM PRODUCTS FOR SPECTRAL CONTRAST ENHANCEMENT

• US 20090299742A1
• Filed: 05/28/2009
• Published: 12/03/2009
• Est. Priority Date: 05/29/2008
• Status: Active Grant

• Current Assignee: Glaxo Group Limited (Glaxosmithkline Holdings Ltd.), Qualcomm, Inc.
• Original Assignee: Qualcomm, Inc.
• Inventors: Lin, Hung-Chun, Visser, Erik, Toman, Jeremy
• Granted Patent: US 8,831,936 B2
• Time in Patent Office: Days
• US Class Current: 704/233
• CPC Class Codes: G10L 21/0208 Noise filtering
• CPC Class Codes: G10L 21/0272 Voice signal separating

Abstract

Systems, methods, and apparatus for spectral contrast enhancement of speech signals, based on information from a noise reference that is derived by a spatially selective processing filter from a multichannel sensed audio signal, are disclosed.

Claims

1. A method of processing a speech signal, said method comprising performing each of the following acts within a device that is configured to process audio signals:

performing a spatially selective processing operation on a multichannel sensed audio signal to produce a source signal and a noise reference; and

performing a spectral contrast enhancement operation on the speech signal to produce a processed speech signal,

wherein said performing a spectral contrast enhancement operation includes;

calculating a plurality of noise subband power estimates based on information from the noise reference;

generating an enhancement vector based on information from the speech signal; and

producing the processed speech signal based on the plurality of noise subband power estimates, information from the speech signal, and information from the enhancement vector, and

wherein each of a plurality of frequency subbands of the processed speech signal is based on a corresponding frequency subband of the speech signal.

2. The method of processing a speech signal according to claim 1, wherein said performing a spatially selective processing operation includes concentrating energy of a directional component of the multichannel sensed audio signal into the source signal.

3. The method of processing a speech signal according to claim 1, wherein said method comprises decoding a signal that is received wirelessly by the device to obtain a decoded speech signal, and

wherein the speech signal is based on information from the decoded speech signal.

4. The method of processing a speech signal according to claim 1, wherein the speech signal is based on the multichannel sensed audio signal.

5. The method of processing a speech signal according to claim 1, wherein said performing a spatially selective processing operation includes determining a relation between phase angles of channels of the multichannel sensed audio signal at each of a plurality of different frequencies.

6. The method of processing a speech signal according to claim 1, wherein said generating an enhancement vector comprises smoothing a spectrum of the speech signal to obtain a first smoothed signal and smoothing the first smoothed signal to obtain a second smoothed signal, and

wherein the enhancement vector is based on a ratio of the first and second smoothed signals.

7. The method of processing a speech signal according to claim 1, wherein said generating an enhancement vector comprises reducing a difference between magnitudes of spectral peaks of the speech signal, and

wherein the enhancement vector is based on a result of said reducing.

8. The method of processing a speech signal according to claim 1, wherein said producing a processed speech signal comprises:

calculating a plurality of gain factor values such that each of the plurality of gain factor values is based on information from a corresponding frequency subband of the enhancement vector;

applying a first of the plurality of gain factor values to a first frequency subband of the speech signal to obtain a first subband of the processed speech signal; and

applying a second of the plurality of gain factor values to a second frequency subband of the speech signal to obtain a second subband of the processed speech signal,

wherein the first of the plurality of gain factor values differs from the second of the plurality of gain factor values.

9. The method of processing a speech signal according to claim 8, wherein each of the plurality of gain factor values is based on a corresponding one of the plurality of noise subband power estimates.

10. The method of processing a speech signal according to claim 8, wherein said producing a processed speech signal includes filtering the speech signal using a cascade of filter stages, and

wherein said applying a first of the plurality of gain factor values to a first frequency subband of the speech signal comprises applying the gain factor value to a first filter stage of the cascade, and

wherein said applying a second of the plurality of gain factor values to a second frequency subband of the speech signal comprises applying the gain factor value to a second filter stage of the cascade.

11. The method of processing a speech signal according to claim 1, wherein said method comprises:

using an echo canceller to cancel echoes from the multichannel sensed audio signal; and

using the processed speech signal to train the echo canceller.

12. The method of processing a speech signal according to claim 1, wherein said method comprises:

based on information from the noise reference, performing a noise reduction operation on the source signal to obtain the speech signal; and

performing a voice activity detection operation based on a relation between the source signal and the speech signal,

wherein said producing a processed speech signal is based on a result of said voice activity detection operation.

13. An apparatus for processing a speech signal, said apparatus comprising:

means for performing a spatially selective processing operation on a multichannel sensed audio signal to produce a source signal and a noise reference; and

means for performing a spectral contrast enhancement operation on the speech signal to produce a processed speech signal,

wherein said means for performing a spectral contrast enhancement operation includes;

means for calculating a plurality of noise subband power estimates based on information from the noise reference;

means for generating an enhancement vector based on information from the speech signal; and

means for producing the processed speech signal based on the plurality of noise subband power estimates, information from the speech signal, and information from the enhancement vector,

wherein each of a plurality of frequency subbands of the processed speech signal is based on a corresponding frequency subband of the speech signal.

14. The apparatus for processing a speech signal according to claim 13, wherein said spatially selective processing operation includes concentrating energy of a directional component of the multichannel sensed audio signal into the source signal.

15. The apparatus for processing a speech signal according to claim 13, wherein said apparatus comprises means for decoding a signal that is received wirelessly by the apparatus to obtain a decoded speech signal, and

wherein the speech signal is based on information from the decoded speech signal.

16. The apparatus for processing a speech signal according to claim 13, wherein the speech signal is based on the multichannel sensed audio signal.

17. The apparatus for processing a speech signal according to claim 13, wherein said means for performing a spatially selective processing operation is configured to determine a relation between phase angles of channels of the multichannel sensed audio signal at each of a plurality of different frequencies.

18. The apparatus for processing a speech signal according to claim 13, wherein said means for generating an enhancement vector is configured to smooth a spectrum of the speech signal to obtain a first smoothed signal and to smooth the first smoothed signal to obtain a second smoothed signal, and

wherein the enhancement vector is based on a ratio of the first and second smoothed signals.

19. The apparatus for processing a speech signal according to claim 13, wherein said means for generating an enhancement vector is configured to perform an operation that reduces a difference between magnitudes of spectral peaks of the speech signal, and

wherein the enhancement vector is based on a result of said operation.

20. The apparatus for processing a speech signal according to claim 13, wherein said means for producing a processed speech signal comprises:

means for calculating a plurality of gain factor values such that each of the plurality of gain factor values is based on information from a corresponding frequency subband of the enhancement vector;

means for applying a first of the plurality of gain factor values to a first frequency subband of the speech signal to obtain a first subband of the processed speech signal; and

means for applying a second of the plurality of gain factor values to a second frequency subband of the speech signal to obtain a second subband of the processed speech signal,

wherein the first of the plurality of gain factor values differs from the second of the plurality of gain factor values.

21. The apparatus for processing a speech signal according to claim 20, wherein each of the plurality of gain factor values is based on a corresponding one of the plurality of noise subband power estimates.

22. The apparatus for processing a speech signal according to claim 20, wherein said means for producing a processed speech signal includes a cascade of filter stages arranged to filter the speech signal, and

wherein said means for applying a first of the plurality of gain factor values to a first frequency subband of the speech signal is configured to apply the gain factor value to a first filter stage of the cascade, and

wherein said means for applying a second of the plurality of gain factor values to a second frequency subband of the speech signal is configured to apply the gain factor value to a second filter stage of the cascade.

23. The apparatus for processing a speech signal according to claim 13, wherein said apparatus comprises means for cancelling echoes from the multichannel sensed audio signal, and

wherein said means for cancelling echoes is configured and arranged to be trained by the processed speech signal.

24. The apparatus for processing a speech signal according to claim 13, wherein said apparatus comprises:

means for performing a noise reduction operation, based on information from the noise reference, on the source signal to obtain the speech signal; and

means for performing a voice activity detection operation based on a relation between the source signal and the speech signal,

wherein said means for producing a processed speech signal is configured to produce the processed speech signal based on a result of said voice activity detection operation.

25. An apparatus for processing a speech signal, said apparatus comprising:

a spatially selective processing filter configured to perform a spatially selective processing operation on a multichannel sensed audio signal to produce a source signal and a noise reference; and

a spectral contrast enhancer configured to perform a spectral contrast enhancement operation on the speech signal to produce a processed speech signal,

wherein said spectral contrast enhancer includes;

a power estimate calculator configured to calculate a plurality of noise subband power estimates based on information from the noise reference; and

an enhancement vector generator configured to generate an enhancement vector based on information from the speech signal, and

wherein said spectral contrast enhancer is configured to produce the processed speech signal based on the plurality of noise subband power estimates, information from the speech signal, and information from the enhancement vector, and

wherein each of a plurality of frequency subbands of the processed speech signal is based on a corresponding frequency subband of the speech signal.

26. The apparatus for processing a speech signal according to claim 25, wherein said spatially selective processing operation includes concentrating energy of a directional component of the multichannel sensed audio signal into the source signal.

27. The apparatus for processing a speech signal according to claim 25, wherein said apparatus comprises a decoder configured to decode a signal that is received wirelessly by the apparatus to obtain a decoded speech signal, and

wherein the speech signal is based on information from the decoded speech signal.

28. The apparatus for processing a speech signal according to claim 25, wherein the speech signal is based on the multichannel sensed audio signal.

29. The apparatus for processing a speech signal according to claim 25, wherein said spatially selective processing operation includes determining a relation between phase angles of channels of the multichannel sensed audio signal at each of a plurality of different frequencies.

30. The apparatus for processing a speech signal according to claim 25, wherein said enhancement vector generator is configured to smooth a spectrum of the speech signal to obtain a first smoothed signal and to smooth the first smoothed signal to obtain a second smoothed signal, and

wherein the enhancement vector is based on a ratio of the first and second smoothed signals.

31. The apparatus for processing a speech signal according to claim 25, wherein said enhancement vector generator is configured to perform an operation that reduces a difference between magnitudes of spectral peaks of the speech signal, and

wherein the enhancement vector is based on a result of said operation.

32. The apparatus for processing a speech signal according to claim 25, wherein said spectral contrast enhancer includes:

a gain factor calculator configured to calculate a plurality of gain factor values such that each of the plurality of gain factor values is based on information from a corresponding frequency subband of the enhancement vector; and

a gain control element configured to apply a first of the plurality of gain factor values to a first frequency subband of the speech signal to obtain a first subband of the processed speech signal, and

wherein said gain control element is configured to apply a second of the plurality of gain factor values to a second frequency subband of the speech signal to obtain a second subband of the processed speech signal,

wherein the first of the plurality of gain factor values differs from the second of the plurality of gain factor values.

33. The apparatus for processing a speech signal according to claim 32, wherein each of the plurality of gain factor values is based on a corresponding one of the plurality of noise subband power estimates.

34. The apparatus for processing a speech signal according to claim 32, wherein said gain control element includes a cascade of filter stages arranged to filter the speech signal, and

wherein said gain control element is configured to apply the first of the plurality of gain factor values to the first frequency subband of the speech signal by applying the gain factor value to a first filter stage of the cascade, and

wherein said gain control element is configured to apply the second of the plurality of gain factor values to the second frequency subband of the speech signal by applying the gain factor value to a second filter stage of the cascade.

35. The apparatus for processing a speech signal according to claim 25, wherein said apparatus comprises an echo canceller configured to cancel echoes from the multichannel sensed audio signal, and

wherein said echo canceller is configured and arranged to be trained by the processed speech signal.

36. The apparatus for processing a speech signal according to claim 25, wherein said apparatus comprises:

a noise reduction stage configured to perform a noise reduction operation, based on information from the noise reference, on the source signal to obtain the speech signal; and

a voice activity detector configured to perform a voice activity detection operation based on a relation between the source signal and the speech signal,

wherein said spectral contrast enhancer is configured to produce the processed speech signal based on a result of said voice activity detection operation.

37. A computer-readable medium comprising instructions which when executed by at least one processor cause the at least one processor to perform a method of processing a multichannel audio signal, said instructions comprising:

instructions which when executed by a processor cause the processor to perform a spatially selective processing operation on a multichannel sensed audio signal to produce a source signal and a noise reference; and

instructions which when executed by a processor cause the processor to perform a spectral contrast enhancement operation on the speech signal to produce a processed speech signal,

wherein said instructions which when executed by a processor cause the processor to perform a spectral contrast enhancement operation include;

instructions which when executed by a processor cause the processor to calculate a plurality of noise subband power estimates based on information from the noise reference;

instructions which when executed by a processor cause the processor to generate an enhancement vector based on information from the speech signal; and

instructions which when executed by a processor cause the processor to produce a processed speech signal based on the plurality of noise subband power estimates, information from the speech signal, and information from the enhancement vector,

wherein each of a plurality of frequency subbands of the processed speech signal is based on a corresponding frequency subband of the speech signal.

38. The computer-readable medium according to claim 37, wherein said instructions which when executed by a processor cause the processor to perform a spatially selective processing operation include instructions which when executed by a processor cause the processor to concentrate energy of a directional component of the multichannel sensed audio signal into the source signal.

39. The computer-readable medium according to claim 37, wherein said medium comprises instructions which when executed by a processor cause the processor to decode a signal that is received wirelessly by a device that includes said medium to obtain a decoded speech signal, and

wherein the speech signal is based on information from the decoded speech signal.

40. The computer-readable medium according to claim 37, wherein the speech signal is based on the multichannel sensed audio signal.

41. The computer-readable medium according to claim 37, wherein said instructions which when executed by a processor cause the processor to perform a spatially selective processing operation include instructions which when executed by a processor cause the processor to determine a relation between phase angles of channels of the multichannel sensed audio signal at each of a plurality of different frequencies.

42. The computer-readable medium according to claim 37, wherein said instructions which when executed by a processor cause the processor to generate an enhancement vector comprise instructions which when executed by a processor cause the processor to smooth a spectrum of the speech signal to obtain a first smoothed signal and instructions which when executed by a processor cause the processor to smooth the first smoothed signal to obtain a second smoothed signal, and

wherein the enhancement vector is based on a ratio of the first and second smoothed signals.

43. The computer-readable medium according to claim 37, wherein said instructions which when executed by a processor cause the processor to generate an enhancement vector comprise instructions which when executed by a processor cause the processor to reduce a difference between magnitudes of spectral peaks of the speech signal, and

wherein the enhancement vector is based on a result of said reducing.

44. The computer-readable medium according to claim 37, wherein said instructions which when executed by a processor cause the processor to produce a processed speech signal comprise:

instructions which when executed by a processor cause the processor to calculate a plurality of gain factor values such that each of the plurality of gain factor values is based on information from a corresponding frequency subband of the enhancement vector;

instructions which when executed by a processor cause the processor to apply a first of the plurality of gain factor values to a first frequency subband of the speech signal to obtain a first subband of the processed speech signal; and

instructions which when executed by a processor cause the processor to apply a second of the plurality of gain factor values to a second frequency subband of the speech signal to obtain a second subband of the processed speech signal,

wherein the first of the plurality of gain factor values differs from the second of the plurality of gain factor values.

45. The computer-readable medium according to claim 44, wherein each of the plurality of gain factor values is based on a corresponding one of the plurality of noise subband power estimates.

46. The computer-readable medium according to claim 44, wherein said instructions which when executed by a processor cause the processor to produce a processed speech signal include instructions which when executed by a processor cause the processor to filter the speech signal using a cascade of filter stages, and

wherein said instructions which when executed by a processor cause the processor to apply a first of the plurality of gain factor values to a first frequency subband of the speech signal comprise instructions which when executed by a processor cause the processor to apply the gain factor value to a first filter stage of the cascade, and

wherein said instructions which when executed by a processor cause the processor to apply a second of the plurality of gain factor values to a second frequency subband of the speech signal comprise instructions which when executed by a processor cause the processor to apply the gain factor value to a second filter stage of the cascade.

47. The computer-readable medium according to claim 37, wherein said medium comprises:

instructions which when executed by a processor cause the processor to cancel echoes from the multichannel sensed audio signal; and

wherein said instructions which when executed by a processor cause the processor to cancel echoes are configured and arranged to be trained by the processed speech signal.

48. The computer-readable medium according to claim 37, wherein said medium comprises:

instructions which when executed by a processor cause the processor to perform a noise reduction operation, based on information from the noise reference, on the source signal to obtain the speech signal; and

instructions which when executed by a processor cause the processor to perform a voice activity detection operation based on a relation between the source signal and the speech signal,

wherein said instructions which when executed by a processor cause the processor to produce a processed speech signal are configured to produce the processed speech signal based on a result of said voice activity detection operation.

49. A method of processing a speech signal, said method comprising performing each of the following acts within a device that is configured to process audio signals:

smoothing a spectrum of the speech signal to obtain a first smoothed signal;

smoothing the first smoothed signal to obtain a second smoothed signal; and

producing a contrast-enhanced speech signal that is based on a ratio of the first and second smoothed signals.

50. The method of processing a speech signal according to claim 49, wherein said producing a contrast-enhanced speech signal comprises, for each of a plurality of subbands of the speech signal, controlling a gain of the subband based on information from a corresponding subband of the ratio of the first and second smoothed signals.