Audio signal processing for speech communication

• US 20050027526A1
• Filed: 09/03/2004
• Published: 02/03/2005
• Est. Priority Date: 05/07/2001
• Status: Active Grant

• Current Assignee: Intel Corporation
• Original Assignee: Intel Corporation
• Inventors: Erell, Adoram, Kleinstein, Avi
• Granted Patent: US 7,149,685 B2
• Time in Patent Office: Days
• US Class Current: 704/237
• CPC Class Codes: G10L 2021/02085 Periodic noise
• CPC Class Codes: H03G 3/32 the control being dependent upon ambient noise level or sound level

Abstract

A device receives a signal that includes human-interpretable audio information. The device detects sound locally and analyzes it to determine if an intermittent component is present. If the intermittent component is present, the received signal is altered so that the audio information is more easily human-interpretable when the signal is performed. The device can be a portable telephone. The intermittent component can be detected, for example, in music.

Claims

1. A method comprising:

at a first device, receiving from a remote device a signal that comprises human-interpretable audio information;

detecting sound at the first device;

analyzing the detected sound to determine if an intermittent component is present; and

if the intermittent component is present, altering the signal so that the audio information is more easily human-interpretable when the signal is converted to sound.

2. The method of claim 1, in which the human-interpretable audio information comprises speech.

3. The method of claim 1, in which the analyzing discriminates the intermittent component from speech.

4. The method of claim 1, in which the detecting comprises storing information about the levels of the detected sound during a series of time intervals, and

the analyzing comprises evaluating the stored information to determine if the intermittent component is present.

5. The method of claim 4, in which the analyzing comprises comparing a first subset of the levels to a second subset of the levels, each subset including levels that correspond to regularly spaced time intervals, the first and second subsets having different regular spacings.

6. The method of claim 4, in which the analyzing comprises determining if a parameter of an auto-correlation function applied to the stored information satisfies a criterion.

7. The method of claim 4, in which the stored information abbreviates at least some information for high level and low levels of the signal during the time intervals.

8. The method of claim 1, in which the first device uses a wireless connection to receive the signal.

9. The method of claim 1, in which the dynamic range of the signal is altered to render the audio information more easily human-interpretable when the signal is performed.

10. The method of claim 9, in which the dynamic range of the signal is altered as a function of an estimate of non-periodic noise when the intermittent component is not detected and an estimate of periodic noise when the intermittent component is detected.

11. The method of claim 10, further comprising analyzing levels of the signal for an interval corresponding to at least one period of the intermittent component to generate the estimate of periodic noise.

12. The method of claim 11, in which the intervals correspond to approximately one period of the intermittent component.

13. The method of claim 1, in which the signal is altered to increase the amplitude of sound generated from the signal when the signal is converted to sound.

14. The method of claim 1, in which the audio component comprises music.

15. The method of claim 1, further comprising rendering the altered signal as sound.

16. A method comprising:

at a device, receiving a signal that comprises audio information;

detecting sound at the device;

storing values related to energy or amplitude of the detected sound, each of the values corresponding to an interval of the detected sound;

analyzing the values to determine if an intermittent component is present;

altering the signal if the intermittent component is determined to be present; and

rendering the altered signal as sound.

17. The method of claim 16, in which the audio signal is received in a digital format.

18. The method of claim 16, in which the altering includes compressing the dynamic range of the signal.

19. The method of claim 18, in which the dynamic range of the signal is altered as a function of an estimate of non-periodic noise when the intermittent component is not detected and an estimate of periodic noise when the intermittent component is detected.

20. The method of claim 16, in which the analyzing includes comparing a first subset of the levels to a second subset of the levels, each subset including levels that correspond to regularly spaced time intervals, the first and second subsets having different regular spacings.

21. The method of claim 16, in which each of the values is modified prior to storage to remove information for bits of the value that correspond to high values and for bits of the value that correspond to low values.

22. The method of claim 21, in which each of the values is stored in a packed format and unpacked for the analyzing.

23. A device comprising:

a receiver, configured to receive a signal representing audio information from a wireless transmission;

a signal modulator, configured to alter the signal in response to a noise estimate;

a sound input unit that detects sound at the device; and

a detector, configured to analyze sound, detected by the sound input unit, for an intermittent component of regular periodicity, generate a noise estimate for the intermittent component when the intermittent component is detected, and communicate the noise estimate to the signal modulator.

24. The device of claim 23, wherein the sound input unit comprises a microphone.

25. The device of claim 23, in which the signal modulator alters the dynamic range of the signal as a function of the noise estimate.

26. The device of claim 23, in which the signal detector comprises

a memory store for storing information about levels of the signal during a series of time intervals;

and

a processor configured to analyze the stored information for the intermittent component.

27. The device of claim 26, in which the processor is configured to (1) compare a first subset of the levels to a second subset of the levels, each subset including levels that correspond to regularly spaced time intervals, the first and second subsets having different regular spacings;

and (2) generate the noise estimate for the intermittent component if the first and second subset satisfy a rule.

28. The device of claim 26, in which the processor is configured to apply an auto-correlation function to the stored information, and generate the noise estimate for the intermittent component if a parameter of the auto-correlation function satisfies a criterion.

29. The device of claim 26, in which the information about levels of the signal is stored in a packed format.

30. The device of claim 29, in which the information about levels of the signal is modified such that only information about a middle range of signal levels is retained.

31. A device comprising:

a processor, configured to modulate a received signal from a remote location, the received signal representing sound at the remote location, if the processor receives an indication that an intermittent component is present in sound detected at the device.

32. The device of claim 31, in which the processor alters the dynamic range of the received signal as a function of a noise estimate for the sound at the device.

33. The device of claim 32, in which the dynamic range of the signal is altered as a function of an estimate of non-periodic noise when the indication of the intermittent component is not received and an estimate of periodic noise when the indication of the intermittent component is received.

34. The device of claim 31, in which the processor increases the amplitude of the sound in the received signal if the indication of the intermittent component is received.