Method and apparatus for determining the direction of a sound source

• US 5,764,779 A
• Filed: 08/16/1994
• Issued: 06/09/1998
• Est. Priority Date: 08/25/1993
• Status: Expired due to Term

• Current Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Original Assignee: Canon Kabushiki Kaisha (Canon Inc.)
• Inventors: Haranishi, Masaki
• Primary Examiner(s): Coles, Sr., Edward L.
• Assistant Examiner(s): BRINICH, STEPHEN M
• Application Number:
  US08/291,224
• Time in Patent Office: 1,393 Days
• Field of Search: 381/38, 381/110, 381/71, 395/2.16, 395/2.26, 395/2.23, 395/2.24, 395/2.4, 395/2.42, 395/2.44
• US Class Current: 381/71.1
• 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 3/005 for combining the signals of two or more microphones specially adapted for hearing aids H04R25/407

Abstract

FIG. 1 illustrates the construction of an audio signal processor according to the present invention. An audio signal enters from a microphone 1 and a bandpass filter 2 extracts the frequency band of the human voice. The resulting signal is amplified by an amplifier 3. A level detecting circuit 4 detects the level of the amplified signal. A level discriminating circuit 5 determines if the level is greater than the value. If yes, the circuit 5 outputs a switch-on signal. Or else, the circuit 5 outputs a switch-off signal. An A/D converter 6 converts the analog signal entering from the circuit 5 into a digital signal. A pitch detecting circuit 7 detects the pitch of the digital signal. A pitch discriminating circuit 8 determines if the pitch of the signal agrees with a prescribed pitch. If yes, the circuit 8 outputs the switch-on signal to the voice-output control switch 13. On the basis of the switch-on or switch-off signal, the switch 13 generates an on/off control signal, which causes a voltage-controlled amplifier 9 to amplify and output the voice signal and output to a mixer 10. Voice processing circuits S and S'"'"' are identical in construction. Microphones 1, 1'"'"' of the processing circuits S, S'"'"' are connected to a mixer 10. The latter mixes the audio outputted by the microphones 1, 1'"'"'. An amplifier 11 amplifies the mixed voice signals. A speaker 12 outputs the audio.

Claims

1. A signal processing method comprising:

an input step of entering an audio signal including a periodic audio signal with a predetermined period;

a level detecting step of detecting the level of the audio signal entered at said input step and generating a level signal;

a period detecting step of detecting a period of the audio signal entered at said input step; and

a selecting step of selecting corresponding image pick-up means and inputting an image to said selected pick-up means if the level signal is greater than a prescribed threshold value, the period detected at said period detecting step falls within a prescribed range,

wherein said selecting step selects the corresponding image pick-up means and inputs the image to said selected pick-up means if the level signal is greater than the prescribed threshold value, a centroid of autocorrelation values corresponding to respective periods detected at said period detecting step within a time duration T falls within a prescribed centroid range and an autocorrelation value corresponding to said centroid is greater than a prescribed threshold value.

2. The method according to claim 1, wherein the prescribed centroid range is approximately 3˜

15 msec.

3. A signal processing method comprising:

an input step of entering an audio signal including a periodic audio signal with a predetermined period;

a level detecting step of detecting the level of the audio signal entered at said input step and generating a level signal;

a period detecting step of detecting a period of the audio signal entered at said input step; and

a selecting step of selecting corresponding image pick-up means and inputting an image to said selected pick-up means if the level signal is greater than a prescribed threshold value, the period detected at said period detecting step falls within a prescribed range,

wherein said period detecting step includes;

a step of partitioning the audio signal entered at said input step into audio signals each of a time duration T;

a step of further partitioning each of the partitioned audio signals into audio signals each of a time duration t; and

a frame period detecting step of detecting periodicity of the audio signals of time duration t.

4. The method according to claim 3, wherein said frame period detecting step includes calculating an autocorrelation function corresponding to the audio signals of time duration t and selecting a period corresponding to a maximum autocorrelation value, which is greater than the threshold value, from among autocorrelation values of said autocorrelation function.

5. The method according to claim 3, wherein said frame period detecting step includes generating a linear prediction equation, which is for approximating the audio signal of the time duration t, based upon the audio signal of the time duration t;

calculating an autocorrelation function relating to a residual signal between the audio signal of the time duration t and a predicted audio signal based upon the linear prediction equation; and

selecting a period corresponding to a maximum autocorrelation value, which is greater than the threshold value, from among autocorrelation values of said autocorrelation function.

6. A signal processing method comprising:

an input step of entering an audio signal including a periodic audio signal with a predetermined period;

a level detecting step of detecting the level of the audio signal entered at said input step and generating a level signal;

a period detecting step of detecting a period of the audio signal entered at said input step; and

an audio control output step of outputting a sound corresponding to the audio signal entered at said input step if the level signal is greater than a prescribed threshold value and the period detected at said period detecting step falls within a prescribed range,

wherein said selecting step outputs a sound corresponding to the audio signal entered at said input step if the level signal is greater than the prescribed threshold value, a centroid of autocorrelation values corresponding to respective periods detected at said period detecting step within a time duration T falls within a prescribed centroid range and an auto correlation value corresponding to said centroid is greater than a prescribed threshold value.

7. The method according to claim 6, wherein the prescribed centroid range is approximately 3˜

15 msec.

8. A signal processing method comprising:

an input step of entering an audio signal including a periodic audio signal with a predetermined period;

a level detecting step of detecting the level of the audio signal entered at said input step and generating a level signal;

a period detecting step of detecting a period of the audio signal entered at said input step; and

an audio control output step of outputting a sound corresponding to the audio signal entered at said input step if the level signal is greater than a prescribed threshold value and the period detected at said period detecting step falls within a prescribed range,

wherein said period detecting step includes;

a step of partitioning the audio signal entered at said input step into audio signals each of a time duration T;

a step of further partitioning each of the partitioned audio signals into audio signals each of a time duration t; and

a frame period detecting step of detecting periodicity of the audio signals of time duration t.

9. The method according to claim 8, wherein said frame period detecting step includes calculating an autocorrelation function corresponding to the audio signals of time duration t and selecting a period corresponding to a maximum autocorrelation value, which is greater than the threshold value, from among autocorrelation values of said autocorrelation function.

10. The method according to claim 8, wherein said frame period detecting step includes generating a linear prediction equation, which is for approximating the audio signal of the time duration t, based upon the audio signal of the time duration t;

calculating an autocorrelation function relating to a residual signal between the audio signal of the time duration t and a predicted audio signal based upon the linear prediction equation; and

selecting a period corresponding to a maximum autocorrelation value, which is greater than the threshold value, from among autocorrelation values of said autocorrelation function.

11. A signal processing apparatus comprising:

input means for entering an audio signal including a periodic audio signal with a predetermined period;

level detecting means for detecting a level of the audio signal entered by said input means and generating a level signal;

period detecting means for detecting a period of the audio signal entered by said input means; and

selecting means for selecting corresponding image pick-up means and inputting an image to said selected pick-up means if the level signal is greater than a prescribed threshold value and the period detected by said period detecting means falls within a prescribed range,

wherein said selecting means selects the corresponding image pick-up means and inputs the image to said selected pick-up means if the level signal is greater than the prescribed threshold value, a centroid of autocorrelation values corresponding to respective periods detected by said period detecting means within a time duration T falls within a prescribed centroid range and an autocorrelation value corresponding to said centroid is greater than a prescribed threshold value.

12. The apparatus according to claim 11, wherein said frame period detecting means includes:

means for generating a linear prediction equation, which is for approximating the audio signal of the time duration t, based upon the audio signal of the time duration t;

means for calculating an autocorrelation function relating to a residual signal between the audio signal of the time duration t and a predicted audio signal based upon the linear prediction equation; and

means for selecting a period corresponding to a maximum autocorrelation value, which is greater than the threshold value, from among autocorrelation values of said autocorrelation function.

13. The apparatus according to claim 11, wherein the prescribed centroid range is approximately 3˜

15 msec.

14. A signal processing apparatus comprising:

input means for entering an audio signal including a periodic audio signal with a predetermined period;

level detecting means for detecting a level of an audio signal entered by input means and generating a level signal;

period detecting means for detecting a period of the audio signal entered by said input means; and

selecting means for selecting corresponding image pick-up means and inputting an image to said selected pick-up means if the level signal is greater than a prescribed threshold value and the period detected by said period detecting means falls within a prescribed range,

wherein said period detecting means includes;

means for partitioning the audio signal entered by said input means into audio signals each of a time duration T;

means for further partitioning each of the partitioned audio signals into audio signals each of a time duration t; and

frame period detecting means for detecting periodicity of the audio signals of time duration t.

15. The apparatus according to claim 14, wherein said frame period detecting means calculates an autocorrelation function corresponding to the audio signals of time duration t and selects a period corresponding to a maximum autocorrelation value, which is greater than the threshold value, from among autocorrelation values of said autocorrelation function.

16. A signal processing apparatus comprising:

input means for entering an audio signal including a periodic audio signal with a predetermined period;

level detecting means for detecting a level of the audio signal entered by said input means and generating a level signal;

period detecting means for detecting a period of the audio signal entered by said input means; and

audio control output means for outputting a sound corresponding to the audio signal entered by said input means if the level signal is greater than a prescribed threshold value and the period detected by said period detecting means falls within a prescribed range,

wherein said selecting means outputs a sound corresponding to the audio signal entered by said input means if the level signal is greater than the prescribed threshold value, a centroid of autocorrelation values corresponding to respective periods detected by said period detecting means within a time duration T falls within a prescribed centroid range and an autocorrelation value corresponding to said centroid is greater than a prescribed threshold value.

17. The apparatus according to claim 16, wherein the prescribed centroid range is approximately 3˜

15 msec.

18. A signal processing apparatus comprising:

input means for entering an audio signal including a periodic audio signal with a predetermined period;

level detecting means for detecting a level of the audio signal entered by said input means and generating a level signal;

period detecting means for detecting a period of the audio signal entered by said input means; and

audio control output means for outputting a sound corresponding to the audio signal entered by said input means if the level signal is greater than a prescribed threshold value and the period detected by said period detecting means falls within a prescribed range,

wherein said period detecting means includes;

means for partitioning the audio signal entered by said input means into audio signals each of a time duration T;

means for further partitioning each of the partitioned audio signals into audio signals each of a time duration t; and

frame period detecting means for detecting periodicity of the audio signals of time duration t.

19. The apparatus according to claim 18, wherein said frame period detecting means calculates an autocorrelation function corresponding to the audio signals of time duration t and selects a period corresponding to a maximum autocorrelation value, which is greater than the threshold value, from among autocorrelation values of said autocorrelation function.

20. The apparatus according to claim 18, wherein said frame period detecting means includes:

means for generating a linear prediction equation, which is for approximating the audio signal of the time duration t, based upon the audio signal of the time duration t;

means for calculating an autocorrelation function relating to a residual signal between the audio signal of the time duration t and a predicted audio signal based upon the linear prediction equation; and

means for selecting a period corresponding to a maximum autocorrelation value, which is greater than the threshold value, from among autocorrelation values of said autocorrelation function.