METHODS AND APPARATUS TO PROVIDE SPEECH PRIVACY

• US 20140095153A1
• Filed: 09/28/2012
• Published: 04/03/2014
• Est. Priority Date: 09/28/2012
• Status: Active Grant

• Current Assignee: Intel Corporation
• Original Assignee: Intel Corporation
• Inventors: de la Guardia Gonzales, Rafael
• Granted Patent: US 9,123,349 B2
• Time in Patent Office: Days
• US Class Current: 704/201
• CPC Class Codes: G10L 21/06 Transformation of speech into a non-audible representation, e.g. speech visualisation or speech processing for tactile aids G10L15/26 takes precedence
• CPC Class Codes: G10L 25/48 specially adapted for particular use

Abstract

Methods and apparatus to provide speech privacy are disclosed. An example method includes forming a sampling block based on a first received audio sample, the sampling block representing speech of a user, creating, with a processor, a mask based on the sampling block, the mask to reduce the intelligibility of the speech of the user, wherein the mask is created by converting the sampling block from a time domain to a frequency domain to form a frequency domain sampling block, identifying a first peak within the frequency domain sampling block, demodulating the frequency domain sampling block at the first peak to form a first envelope of the sampling block, distorting the first envelope to form a first distorted envelope, and emitting an acoustic representation of the mask via a speaker.

Claims

1. A method to provide speech privacy, comprising:

forming a sampling block based on a first received audio sample, the sampling block representing speech of a user;

creating, with a processor, a mask based on the sampling block, the mask to reduce the intelligibility of the speech of the user, wherein the mask is created by;

converting the sampling block from a time domain to a frequency domain to form a frequency domain sampling block;

identifying a first peak within the frequency domain sampling block;

demodulating the frequency domain sampling block at the first peak to form a first envelope of the sampling block;

distorting the first envelope to form a first distorted envelope; and

emitting an acoustic representation of the mask via a speaker.

2. The method of claim 1, further comprising subtracting the mask from a second received audio sample to form a third audio sample, the second audio sample representing the speech of the user plus the mask.

3. The method of claim 2, further comprising transmitting the third audio sample to a calling partner.

4. The method of claim 2, further comprising storing the third audio sample in a memory.

5. The method of claim 1, further comprising storing the mask in a memory.

6. The method of claim 1, further comprising:

identifying a second peak within the frequency domain sampling block;

demodulating the frequency domain sampling block at the second peak to form a second envelope of the sampling block;

distorting the second envelope to form a second distorted envelope; and

combining the first distorted envelope and the second distorted envelope.

7. The method of claim 6, wherein:

distorting the first envelope comprises adding a first phase shift to the first envelope; and

distorting the second envelope comprises adding a second phase shift to the second envelope.

8. The method of claim 7, wherein the first phase shift is different from the second phase shift.

9. The method of claim 1, wherein converting the sampling block is implemented using a short time Fourier transform.

10. The method of claim 1, wherein the first peak represents a first harmonic of the sampling block.

11. The method of claim 1, wherein distorting the first envelope comprises adding a phase shift to the first envelope.

12. A speech privacy apparatus comprising:

an audio receiver to receive speech from a user;

a masker to create an audio mask based on the speech from the user, the audio mask to reduce an intelligibility of the speech of the user, the masker comprising;

a domain converter to convert the speech received from the user into a frequency domain sampling block;

a frequency tracker to identify a first peak within the frequency domain sampling block;

a demodulator to demodulate the frequency domain sampling block at the first peak to form a first envelope;

a distorter to distort the first envelope to form a first distorted envelope; and

a speaker to emit an acoustic representation of the audio mask.

13. The speech privacy apparatus of claim 12, wherein the audio receiver is to receive the speech from the user and the audio mask emitted from the speaker as a second audio sample, and further comprising:

a memory to store the audio mask; and

a de-masker to subtract the audio mask stored in the memory from the second audio sample to form a clean speech sample.

14. The speech privacy apparatus of claim 13, further comprising a network communicator to transmit the clean speech sample to a calling partner.

15. The speech privacy apparatus of claim 13, wherein the de-masker is to store the clean speech sample in the memory.

16. A computer-readable storage medium comprising instructions which, when executed, cause a machine to at least:

form a sampling block based on a first received audio sample, the sampling block representing speech of a user;

create a mask based on the sampling block, the mask to reduce the intelligibility of the speech of the user, wherein the mask is created by;

converting the sampling block from a time domain to a frequency domain to form a frequency domain sampling block;

identifying a first peak within the frequency domain sampling block;

demodulating the frequency domain sampling block at the first peak to form a first envelope of the sampling block;

distorting the first envelope to form a first distorted envelope; and

emit an acoustic representation of the mask via a speaker.

17. The computer-readable storage medium of claim 16, further comprising instruction which, when executed, cause the machine to subtract the mask from a second received audio sample to form a third audio sample, the second audio sample representing the speech of the user plus the mask.

18. The computer-readable storage medium of claim 17, further comprising instruction which, when executed, cause the machine to transmit the third audio sample to a calling partner.

19. The computer-readable storage medium of claim 17, further comprising instruction which, when executed, cause the machine to store the third audio sample in a memory.

20. The computer-readable storage medium of claim 16, further comprising instruction which, when executed, cause the machine to store the mask in a memory.

21. The computer-readable storage medium of claim 16, further comprising instruction which, when executed, cause the machine to at least:

identify a second peak within the frequency domain sampling block;

demodulate the frequency domain sampling block at the second peak to form a second envelope of the sampling block;

distort the second envelope to form a second distorted envelope; and

combine the first distorted envelope and the second distorted envelope.

22. The computer-readable storage medium of claim 21, further comprising instruction which, when executed, cause the machine to at least:

distort the first envelope by adding a first phase shift to the first envelope; and

distort the second envelope by adding a second phase shift to the second envelope.

23. The computer-readable storage medium of claim 22, wherein the first phase shift is different from the second phase shift.

24. The computer-readable storage medium of claim 16, wherein converting the sampling block is implemented using a short time Fourier transform.

25. The computer-readable storage medium of claim 16, wherein the first peak represents a first harmonic of the sampling block.

26. The computer-readable storage medium of claim 16, wherein distorting the first envelope comprises adding a phase shift to the first envelope.