System and method for speech enhancement using a coherent to diffuse sound ratio

• US 10,242,690 B2
• Filed: 12/12/2014
• Issued: 03/26/2019
• Est. Priority Date: 12/12/2014
• Status: Active Grant

• Current Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Original Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Inventors: Wolff, Tobias, Matheja, Timo, Buck, Markus
• Primary Examiner(s): Roberts, Shaun
• Application Number:
  US15/535,245
• Publication Number:
  US 20170330580A1
• Time in Patent Office: 1,565 Days
• Field of Search: 704226
• CPC Class Codes: G10L 2021/02166 Microphone arrays; Beamforming
• CPC Class Codes: G10L 21/0208 Noise filtering
• CPC Class Codes: G10L 21/0216 characterised by the method used for estimating noise
• CPC Class Codes: H04B 1/62 for providing a predistortion of the signal in the transmitter and corresponding correction in the receiver, e.g. for improving the signal/noise ratio
• CPC Class Codes: H04R 2430/21 Direction finding using differential microphone array [DMA]
• CPC Class Codes: H04R 2430/23 Direction finding using a sum-delay beam-former
• 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

Embodiments of the present disclosure may include a system and method for speech enhancement using the coherent to diffuse sound ratio. Embodiments may include receiving an audio signal at one or more microphones and controlling one or more adaptive filters of a beamformer using a coherent to diffuse ratio (“CDR”).

Claims

1. A computer-implemented method comprising:

receiving an audio signal at one or more microphones; and

controlling one or more adaptive filters of a beamformer using a coherent to diffuse ratio (“

CDR”

) by determining speech and speech pause based upon the CDR, and adapting one of a blocking filter of the one or more adaptive filters during the speech and an interference cancellation filter of the one or more adaptive filters during the speech pause.

2. The computer-implemented method of claim 1, wherein the CDR is estimated and mapped to a control variable.

3. The computer-implemented method of claim 1, wherein the CDR is combined with at least one of a signal to noise ratio (“

SNR”

) and a signal power ratio (“

SPR”

).

4. The computer-implemented method of claim 1, wherein controlling occurs in a generalized sidelobe canceller (“

GSC”

) type beamformer.

5. The computer-implemented method of claim 1, further comprising:

controlling an update associated with the interference cancellation filter.

6. The computer-implemented method of claim 1, wherein the CDR is configured to suppress late reverberation prior to direction of arrival (“

DOA”

) estimation.

7. The computer-implemented method of claim 1, wherein the CDR is used as a spectral weight in a generalized cross correlation.

8. The computer-implemented method of claim 1, wherein the CDR is used to detect at least one signal component for T-60 estimation.

9. The computer-implemented method of claim 1, further comprising:

estimating a power spectral density associated with one of a post-beamforming noise and a microphone signal used in a spatial postfilter.

10. The computer-implemented method of claim 1, further comprising:

generating a CDR-based overestimation factor for a noise power spectral density.

11. The computer-implemented method of claim 9, wherein estimating includes suppressing a coherent sound part in the power spectral density using a spectral CDR that refers to one of an output of the blocking matrix and an error of the blocking filter.

12. A system comprising:

one or more processors and one or more microphones configured to receive an audio signal, the one or more processors configured to control one or more adaptive filters of a beamformer using a coherent to diffuse ratio (“

CDR”

) by determining speech and speech pause based upon the CDR, and adapting one of a blocking filter of the one or more adaptive filters during the speech and an interference cancellation filter of the one or more adaptive filters during the speech pause.

13. The system of claim 12, wherein the CDR is estimated and mapped to a control variable.

14. The system of claim 12, wherein the CDR is combined with at least one of a signal to noise ratio (“

SNR”

) and a signal power ratio (“

SPR”

).

15. The system of claim 12, wherein controlling occurs in a generalized sidelobe canceller (“

GSC”

) type beamformer.

16. The system of claim 12, further comprising:

controlling an update associated with the interference cancellation filter.

17. The system of claim 12, wherein the CDR is configured to suppress late reverberation prior to direction of arrival (“

DOA”

) estimation.

18. The system of claim 12, wherein the CDR is used as a spectral weight in a generalized cross correlation.

19. The system of claim 12, wherein the CDR is used to detect at least one signal component for T-60 estimation.

20. The system of claim 12, further comprising:

estimating a power spectral density associated with one of a post-beamforming noise and a microphone signal used in a spatial postfilter.

21. The system of claim 12, further comprising:

generating a CDR-based overestimation factor for a noise power spectral density.

22. The system of claim 20, wherein estimating includes suppressing a coherent sound part in the power spectral density using a spectral CDR that refers to one of an output of the blocking matrix and an error of the blocking filter.