Method for detecting audio ticks in a noisy environment

• US 9,235,259 B2
• Filed: 11/26/2010
• Issued: 01/12/2016
• Est. Priority Date: 11/27/2009
• Status: Active Grant

• Current Assignee: K.Mizra LLC
• Original Assignee: Nederlandse Organisatie Voor Toegepast-Natuurwetenschappelijk Onderzoek Tno
• Inventors: van Staalduinen, Mark, Klos, Victor Bastiaan, Doets, Peter Jan Otto
• Primary Examiner(s): Nguyen, Duc
• Assistant Examiner(s): Mohammed, Assad
• Application Number:
  US13/512,139
• Publication Number:
  US 20120288103A1
• Time in Patent Office: 1,873 Days
• US Class Current: 1/1
• CPC Class Codes: G06F 1/1626 with a single-body enclosure integrating a flat display, e.g. Personal Digital Assistants [PDAs]
• CPC Class Codes: G06F 2200/1636 Sensing arrangement for detection of a tap gesture on the housing
• CPC Class Codes: G06F 3/002 Specific input/output arrangements not covered by G06F3/01 - G06F3/16 other optical apparatus G02B27/00
• CPC Class Codes: G06F 3/017 Gesture based interaction, e.g. based on a set of recognized hand gestures interaction based on gestures traced on a digitiser G06F3/04883
• CPC Class Codes: G10L 25/48 specially adapted for particular use

Abstract

A method for detecting, in an audio signal including a stream of audio samples (y(n), (y′(n)), ticks in a noisy environment, comprises the steps of applying the signal to a coarse tick detection processor arranged to decide whether it is likely that the signal includes a tick and, only then, enabling a fine tick detection processor to decide, by more thorough processing the audio signal, whether the audio signal represents an audio tick. The coarse processing step preferably includes buffering audio samples (y(n), (y′(n)), determining the maximum and minimum values of each sample, forming their local range, delaying the local range with one or more samples and comparing the difference between the actual local range and the delayed local range with a threshold value (T_c), and, if the difference exceeds the threshold value, outputting a trigger (n_c) for enabling the fine detection processor The fine processing step preferably includes buffering audio samples (y(n), (y′(n)), computing, for each buffer content (B_p(n,m)), a Fast Fourier Transformed buffer content (E_c(ω,m)), determining the difference (F_c(ω,m)) between the resulting frequency components, comparing the difference (F_c(ω,m)) to the content of a previously trained fingerprint set (F*(ω,m)) and calculating a correlation coefficient (p(F_c,F*)) between them, comparing the correlation coefficient to a threshold value (T_P) and, if the correlation coefficient exceeds the threshold value, outputting a signal indication the presence of a tick in the audio signal.

Claims

1. Method for detecting ticks on a device with which an audio signal is received in a noisy environment, from the audio signal which includes a stream of audio samples (y(n), (y′

(n)), the method comprising the steps of;

applying the audio signal to a coarse tick detection processor arranged to decide whether the audio signal likely includes a tick; and

only if the course tick processor decides that the audio signal likely includes a tick, enabling a fine tick detection processor to decide, by more thorough processing of said audio signal, whether the audio signal represents a tick on the device, wherein the fine tick detection is performed using a comparison with a previously trained set of reference properties that are characteristic for the device with which the audio signal is received.

2. Method according claim 1, wherein the stream of audio samples is derived from a microphone of a mobile communication device that captures sound generated by tapping the mobile communication device against other objects and noise from an environment of the mobile communication device.

3. Method according claim 1, wherein said comparison comprises computing a correlation coefficient between the reference properties and information derived from the stream of audio samples.

4. Method according claim 1, wherein the coarse processing step includes buffering a series of audio samples (y(n), (y′

(n)) in a buffer, determining maximum and minimum values from this buffer, computing a local range between the maximum and minimum values, delaying said local range with one or more samples, and comparing a difference between an actual local range and the delayed local range with a threshold value (T_c), and, if the difference exceeds the threshold value, outputting a trigger (n_c) for enabling said fine detection processor.

5. Method according to claim 1, wherein the fine processing step includes buffering the audio samples (y(n), (y′

(n)) from a time interval that has been decided to likely include the tick by the coarse tick detection into buffer contents that include a buffer content with said tick and a buffer content that precedes said tick, computing, for each buffer content (B_p(n,m)), a Fast Fourier Transformed buffer content (E_c(w,m)), obtaining frequency components from the Fast Fourier Transformed buffer content (E_c(w,m)), determining a difference (F_c(w,m)) between frequency components for the buffer contents with said tick and preceding said tick, comparing said difference (F_c(w,m)) to contents of a previously trained reference set (F*(w,m)) and calculating a correlation coefficient (ρ

(F_c,F*)) between them, comparing said correlation coefficient to a threshold value (T_p) and, if the correlation coefficient exceeds the threshold value, outputting a signal indicating the presence of a tick in the audio signal.

6. Method according to claim 5, comprising a training step, wherein a reference set of a touch is formed that is specific for a mobile device from which the stream of audio samples is derived using a microphone of a mobile communication device.

7. Method according to claim 1, wherein the audio signal is preprocessed before being processed in the coarse and/or fine processing steps respectively, including determining and outputting a difference (y′

(n)) between each audio sample (y(n)) and a previous audio sample (y(n−

x)).

8. A non-tangible computer readable storage medium comprising a program of instructions for making a programmable computer to carry out one or more of the method steps according to claim 1.

15. Consumer product, adapted for carrying out the method according to claim 1, and/or comprising the system according to claim 9.

16. Consumer product according to claim 15, which is a mobile telephone device or a PDA.

17. Server for identifying proximate devices, adapted for carrying out the method according to claim 1 and/or comprising the system according to claim 9.

18. The method according to claim 1, comprising executing, before said applying the audio signal to the coarse tick detection processor and enabling the fine tick detection processor, the preceding steps of:

signalling when a reference touch or respective reference touches are generated on the device,

determining reference properties of a reference audio signal obtained from the device at a time of the signaled touch or respective touches;

storing the reference properties, or averages of the reference properties obtained with the respective touches, as the reference properties that are characteristic for the device; and

wherein the fine tick detection is performed using a comparison with the reference properties that are characteristic for the device.

9. System for detecting, in an audio signal which includes a stream of audio samples (y(n), (y′

(n)), ticks on a device with which the audio signal is received in a noisy environment, the system comprising a coarse tick detection processor and a fine tick processor, wherein the coarse tick processor is arranged to decide whether the signal likely includes a tick; and

, only if the course tick processor decides that the audio signal likely includes a tick, to enable the fine tick processor to decide, by more thorough processing of said audio signal, whether the audio signal represents a tick on the device, wherein the fine tick detection is performed using a comparison with a previously trained set of reference properties that are characteristic for the device with which the audio signal is received.

10. System according claim 9, comprising a mobile telecommunication device comprising a microphone, coupled to the coarse tick detection processor and the fine tick processor to supply the stream of audio samples.

11. System according to claim 9, wherein the coarse tick processor is adapted for buffering a series of the audio samples (y(n), (y′

(n)) in a buffer, for determining maximum and minimum values from the buffer, computing a local range between the maximum and the minimum, for delaying said local range with one or more samples and for comparing a difference between the actual local range and the delayed local range with a threshold value (T_c), and for, if the difference exceeds the threshold value, outputting a trigger (n_c) for enabling said fine tick processor.

12. System according to claim 9, wherein the fine tick processor is adapted for buffering the audio samples (y(n), (y′

(n)) from a time interval that includes the tick that has been decided by the coarse tick detection into buffer contents that include a buffer content with said tick and a buffer content that precedes said tick, for computing, for each buffer content (B_p(n,m)), a Fast Fourier Transformed buffer content (E_c(w,m)), for obtaining frequency components from the Fast Fourier Transformed buffer content (E_c(w,m)), for determining a difference (F_c(w,m)) between the frequency components for the buffer contents with said tick and preceding said tick, for comparing said difference (F_c(w,m)) to contents of a previously trained reference set (F*(w,m)) and calculating a correlation coefficient (ρ

(F_c,F*)) between them, for comparing said correlation coefficient to a threshold value (T_p) and for, if the correlation coefficient exceeds the threshold value, outputting a signal indicating a presence of the tick in the audio signal.

13. System according to claim 9, including a preprocessor adapted for preprocessing the audio signal before being processed in the coarse and/or fine processors respectively, the preprocessor being arranged for determining and outputting a difference (y′

(n)) between each audio sample (y(n)) and a previous audio sample (y(n−

x)).

14. System according to claim 9, comprising a mobile communication device with a microphone coupled to the coarse tick detection processor and the fine tick processor, the fine tick processor being configured to trigger mutual identification of, or opening of a communication channel between, a pair of mobile communication devices that includes the mobile communication device, in response to fine tick detection.

19. The system according to claim 9, the system being configured to receive signals when a reference touch or respective reference touches are generated on the device, determine reference properties of a reference audio signal obtained from the device at a time of a signaled touch or respective touches and store the reference properties, or averages of the reference properties obtained with the respective touches, as the reference properties that are characteristic for the device, the fine tick processor being configured to perform the comparison with the reference properties that are characteristic for the device.

20. A method for detecting ticks in a noisy environment, from an audio signal received by a device, the audio signal including a stream of audio samples (y(n), (y′

(n)), the method comprising the steps of;

applying the audio signal to a coarse tick detection processor adapted to decide whether the audio signal likely includes a tick; and

only if the course tick processor decides that the audio signal likely includes a tick, enabling a fine tick detection processor to decide, by more thorough processing of said audio signal, whether the audio signal represents a tick, wherein the fine tick detection is performed using a comparison with predetermined reference properties that are characteristic for the device with which the audio signal is received, suppressing influence of background signals, by

computing Fourier transforms of the audio samples in a plurality of buffers;

determining differences between energies of respective frequency components of the Fourier transforms for successive buffers; and

comparing the differences to contents of a previously trained reference set.

21. The method according to claim 20, wherein the Fourier transforms of the audio samples are computed in buffers B_p(n_c,m), wherein B_p(n_c,m)={y′

(n_c′

(m)−

N_p), . . . , y′

(n_c′

(m)−

2),y′

(n_c′

(m)−

1)},

wherein n_c is a sample number corresponding to a coarse touch, m is an index of a buffer, N_p a number of samples in the buffer and n_c′

(m)=n_c+mN_p(1−

p), with p in the interval [0,1]

said determining the difference comprising determining a difference F_c(ω

,m′

)=E_c(ω

,m′

)−

E(ω

,m′

−

1) between the energies E_c(ω

,m′

) and E_c(ω

,m′

−

1) of respective frequency components at frequencies ω

of the Fourier transforms for successive buffer numbers m′

, m′

−

1; and

said comparing comprising comparing the differences (F_c(ω

,m)) to the content of the previously trained reference set (F*(ω

,m)).

22. A system for detecting ticks from a noisy environment in an audio signal received by a device, wherein the audio signal includes a stream of audio samples (y(n), (y′

(n)) of the noisy environment, the system comprising a coarse tick detection processor and a fine tick detection processor, wherein the coarse tick detection processor is adapted to decide whether the audio signal likely includes a tick; and

, only if the course tick processor decides that the audio signal likely includes a tick, to enable the fine tick detection processor to decide, by more thorough processing of said audio signal, whether the audio signal represents a tick, wherein the fine tick detection processor is configured to perform a comparison with predetermined reference properties that are characteristic for the device with which the audio signal is received, suppressing influence of background signals, by

computing Fourier transforms of the audio samples in a plurality of buffers;

determining differences between energies of respective frequency components of the Fourier transforms for successive buffers; and

comparing the differences to contents of a previously trained reference set.

23. The system according to claim 22, wherein the fine tick detection processor is configured to compute said Fourier transforms of the audio samples in buffers B_p(n_c,m) B_p (n_c,m)={y′

(n_c′

(m)−

N_p), . . . , y′

(n_c′

(m)−

2),y′

(n_c′

(m)−

1)}, wherein n_c is a sample number corresponding to a coarse touch, m is an index of a buffer B_p(n_c,m), N_p is a number of samples in the buffer and n_c′

(m)=n_c+mN_p(1−

p), with p in the interval [0,1]; and

wherein the fine tick detection processor is configured to;

determine a difference F_c(ω

,m′

)=E_c(ω

,m′

)−

E(ω

,m′

−

1) between the energies E_c(w,m′

) and E_c(w,m′

−

1) of respective frequency components at frequencies w of the Fourier transforms for successive buffer numbers m′

, m′

−

1; and

comparing the differences (F_c(w,m)) to the content of the previously trained reference set (F*(w,m)).