Adaptive microphone sampling rate techniques

• US 9,406,313 B2
• Filed: 03/21/2014
• Issued: 08/02/2016
• Est. Priority Date: 03/21/2014
• Status: Active Grant

• Current Assignee: Intel Corporation
• Original Assignee: Intel Corporation
• Inventors: Ryan, Joseph F., Tschanz, James W., Beltman, Willem M.
• Primary Examiner(s): ALBERTALLI, BRIAN LOUIS
• Application Number:
  US14/222,284
• Publication Number:
  US 20150269954A1
• Time in Patent Office: 865 Days
• US Class Current: 1/1
• CPC Class Codes: G10L 15/28 Constructional details of speech recognition systems
• CPC Class Codes: G10L 25/24 the extracted parameters being the cepstrum
• CPC Class Codes: G10L 25/48 specially adapted for particular use
• CPC Class Codes: G10L 25/78 Detection of presence or absence of voice signals switching of direction of transmission by voice frequency in two-way loud-speaking telephone systems H04M9/10
• CPC Class Codes: H04R 2499/11 Transducers incorporated or for use in hand-held devices, e.g. mobile phones, PDA's, camera's
• CPC Class Codes: H04R 29/004 for microphones H04R29/007 takes precedence
• CPC Class Codes: H04R 3/00 Circuits for transducers , loudspeakers or microphones

Abstract

An apparatus for adjusting a microphone sampling rate, the apparatus including an input to receive an audio signal from a microphone and a front-end processing module. The front-end processing module is to generate a plurality of frames from the audio signal received by the microphone, determine a noise profile using the plurality of frames, and adjust a sampling rate of the microphone based on the determined noise profile.

Claims

1. An apparatus for adjusting a microphone sampling rate, the apparatus comprising:

an input to receive an audio signal from a microphone; and

a front-end processing module to;

generate a plurality of frames from the audio signal received by the microphone;

determine a spectral energy representation of each of the plurality of frames, the spectral energy representation including N-points when the microphone sampling rate of is a first sampling rate, the spectral energy representation including M-points when the sampling rate is a second sampling rate that is less than the first sampling rate, where M is less than N;

map and scale at least some of the M-points to the N-points when using the second sampling rate, prior to calculating at least one Mel-frequency cepstral coefficient (MFCC);

determine a noise profile using the plurality of frames;

adjust the sampling rate of the microphone based on the determined noise profile;

calculate the at least one MFCC for each of the plurality of frames; and

detect a presence of a human voice based on the at least one MFCC.

2. The apparatus of claim 1, wherein the front-end processing module is to detect a presence of a human voice based on the at least one MFCC is configured to:

detect a presence of the human voice when the at least one MFCC is above a threshold value.

3. The apparatus of claim 1, wherein to adjust the sampling rate of the microphone based on the determined noise profile, the front-end processing module is to adjust the sampling rate of the microphone from the first sampling rate to the second sampling rate based on the determined noise profile.

4. The apparatus of claim 3, wherein the front-end processing module is to adjust the sampling rate of the microphone from the first sampling rate to the second sampling rate when the determined noise profile is below a threshold.

5. The apparatus of claim 1, wherein the front-end processing module includes a Fast Fourier Transform (FFT) module to perform an FFT on each of the plurality of frames,

wherein the spectral energy representation that includes N-points when the sampling rate is a first sampling rate includes a 512-point FFT;

wherein the spectral energy representation that includes M-points when the sampling rate is a second sampling rate includes a 256-point FFT; and

wherein the front-end processing module configured to map and scale at least some of the M-points to the N-points when using the second sampling rate, prior to calculating the at least one MFCC is to;

map points 1 through 128 of the 256-point FFT to points 1 through 128 of the 512-point FFT, respectively, prior to calculating the at least one MFCC, and

map points 129 through 256 of the 256-point FFT to points 385 through 512 of the 512-point FFT, respectively, prior to calculating the at least one MFCC.

6. The apparatus of claim 1, wherein the first sampling rate is about 8 kilohertz (kHz) and the second sampling rate is about 4 kHz.

7. The apparatus of claim 1, wherein to determine the noise profile using the plurality of frames, the front-end processing module is to determine the noise profile based on a running average.

8. The apparatus of claim 1, wherein the front-end processing module is to:

determine the noise profile using at least one of a plurality of coefficients; and

adjust the at least one of the plurality of coefficients based on the determined sampling rate.

9. A method for adjusting a microphone sampling rate, the method comprising:

receiving an audio signal from a microphone;

generating a plurality of frames from the audio signal received by the microphone;

determining a spectral energy representation of each of the plurality of frames, wherein the spectral energy representation includes N-points when the microphone sampling rate is a first sampling rate, wherein the spectral energy representation includes M-points when the sampling rate is a second sampling rate that is less than the first sampling rate, where M is less than N;

mapping and scaling at least some of the M-points to the N-points when using the second sampling rate, prior to calculating at least one Mel-frequency cepstral coefficient (MFCC);

determining a noise profile using the plurality of frames;

adjusting the sampling rate of the microphone based on the determined noise profile;

calculating the at least one MFCC for each of the plurality of frames; and

detecting a presence of a human voice based on the at least one MFCC.

10. The method of claim 9, wherein detecting a presence of a human voice based on the at least one MFCC includes:

detecting a presence of a human voice when the at least one MFCC is above a threshold value.

11. The method of claim 9, wherein adjusting the sampling rate of the microphone based on the determined noise profile includes:

adjusting the sampling rate of the microphone from the first sampling rate to the second sampling rate based on the determined noise profile.

12. The method of claim 11, comprising:

adjusting the sampling rate of the microphone from the first sampling rate to the second sampling rate when the determined noise profile is below a threshold.

13. The method of claim 9, further comprising:

performing a Fast Fourier Transform (FFT) on each of the plurality of frames,

wherein the spectral energy representation that includes N-points when the sampling rate is a first sampling rate includes a 512-point FFT;

wherein the spectral energy representation that includes M-points when the sampling rate is a second sampling rate includes a 256-point FFT; and

wherein mapping and scaling at least some of the M-points to the N-points when using the second sampling rate, prior to calculating the at least one MFCC includes;

mapping and scaling points 1 through 128 of the 256-point FFT to points 1 through 128 of the 512-point FFT, respectively, prior to calculating the at least one MFCC, and

mapping and scaling points 129 through 256 of the 256-point FFT to points 385 through 512 of the 512-point FFT, respectively, prior to calculating the at least one MFCC.

14. The method of claim 9, wherein the first sampling rate is about 8 kilohertz (kHz) and the second sampling rate is about 4 kHz.

15. The method of claim 9, wherein determining a noise profile using the plurality of frames includes determining the noise profile based on a running average.

16. The method of claim 9, further comprising:

determining the noise profile using at least one of a plurality of coefficients; and

adjusting the at least one of the plurality of coefficients based on the determined sampling rate.

17. A machine-readable medium including instructions for adjusting a microphone sampling rate, which when executed by a processor, cause the processor to perform operations comprising:

receiving an audio signal from a microphone;

generating a plurality of frames from the audio signal received by the microphone;

determining a spectral energy representation of each of the plurality of frames, wherein the spectral energy representation includes N-points when a sampling rate of the microphone is a first sampling rate, wherein the spectral energy representation includes M-points when the sampling rate is a second sampling rate that is less than the first sampling rate, where M is less than N;

mapping and scaling at least some of the M-points to the N-points when using the second sampling rate, prior to calculating at least one Mel-frequency cepstral coefficient (MFCC);

determining a noise profile using the plurality of frames;

adjusting the sampling rate of the microphone based on the determined noise profile;

calculating the at least one MFCC for each of the plurality of frames; and

detecting a presence of a human voice based on the at least one MFCC.

18. The machine-readable medium of claim 17, wherein detecting a presence of a human voice based on the at least one MFCC includes:

detecting a presence of a human voice when the at least one MFCC is above a threshold value.

19. The machine-readable medium of claim 17, wherein adjusting the sampling rate of the microphone based on the determined noise profile includes:

adjusting the sampling rate of the microphone from the first sampling rate to the second sampling rate based on the determined noise profile.

20. The machine-readable medium of claim 19, the operations comprising:

adjusting the sampling rate of the microphone from the first sampling rate to the second sampling rate when the determined noise profile is below a threshold.