Method for converging a G.729 Annex B compliant voice activity detection circuit

• US 20020184015A1
• Filed: 06/01/2001
• Published: 12/05/2002
• Est. Priority Date: 06/01/2001
• Status: Active Grant

• Current Assignee: Telogy Networks, Inc. (Texas Instruments, Inc.)
• Original Assignee: Telogy Networks, Inc. (Texas Instruments, Inc.)
• Inventors: Li, Dunling, Sisli, Gokhan, Thomas, Daniel C.
• Granted Patent: US 7,031,916 B2
• Time in Patent Office: Days
• US Class Current: 704/233
• CPC Class Codes: G10L 2021/02168 the estimation exclusively taking place during speech pauses
• CPC Class Codes: G10L 2025/783 based on threshold decision
• 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

Abstract

A method of initializing an ITU Recommendation G.729 Annex B voice activity detection (VAD) device is disclosed, having the steps of (1) extracting a set of parameters from a signal that characterize the signal; (2) calculating an energy measure of the signal from the set of parameters; (3) comparing the energy measure with a reference value; (4) determining an initial value for an average of a noise characteristic of the signal; and (5) counting the number of times the energy measure equals or exceeds the reference level.

Also disclosed is a method of converging an ITU Recommendation G.729 Annex B voice activity detection (VAD) device, having the steps of: (1) determining a noise identification threshold value; (2) comparing a number of energy measures of a signal to the noise threshold value; (3) determining a first value representing an average of the number of energy measures, when the energy measure is less than the noise threshold, wherein only the energy measures of the number of energy measures having values less than the noise threshold value are used to determine the first value; (4) determining a second value representing an average of the number of energy measures; and (5) substituting the first value for the second value when a specific event occurs, indicating the divergence of the two values.

Claims

1. A method of initializing an ITU Recommendation G.729 Annex B voice activity detection (VAD) device, comprising the steps of:

extracting a set of parameters from a signal that characterize said signal;

calculating an energy measure of said signal from said set of parameters;

comparing said energy measure with a reference value;

determining an initial value for an average of a noise characteristic of said signal; and

counting the number of times said energy measure equals or exceeds said reference level.

2. The method according to claim 1, further comprising the step of:

performing the sequential process of steps recited in claim 1 repeatedly, in their listed order, until said number of times equals thirty-two.

3. A method of initializing an ITU Recommendation G.729 Annex B voice activity detection (VAD) device, comprising the steps of:

extracting a set of parameters characterizing a signal from a digital representation of said signal within a data frame, wherein said parameters are the autocorrelation coefficients, which are derived in accordance with said Recommendation G.729, and are denoted by {R(i)}_i=0^p;

calculating a full-band frame energy by multiplying a value of ten times a base ten logarithm of a quotient obtained by dividing a first autocorrelation coefficient R(0), of said autocorrelation coefficients, by a constant value of 240;

comparing said full-band frame energy with a reference level;

updating initial values for averages of the noise characteristics in accordance with said Recommendation G.729 Annex B; and

changing the value of a frame counter during said initialization only if said full-band frame energy equals or exceeds said reference level.

4. The method according to claim 3, further comprising the step of:

performing the sequential process of steps recited in claim 3 repeatedly, in their listed order, until said frame count has been changed thirty-two times.

5. A method of converging an ITU Recommendation G.729 Annex B voice activity detection (VAD) device, comprising the steps of:

determining a noise identification threshold value;

comparing a number of energy measures of a signal to said noise threshold value;

determining a first value representing an average of said number of energy measures, when said energy measure is less than said noise threshold, wherein only the energy measures of said number of energy measures having values less than said noise threshold value are used to determine said first value;

determining a second value representing an average of said number of energy measures; and

substituting said first value for said second value when a specific event occurs.

6. The method according to claim 5, wherein:

said specific event is an increasing divergence between said first and second values with time.

7. The method according to claim 5, wherein:

said specific event is the expiration of a period of time.

8. The method according to claim 5, further comprising the step of:

counting the number of consecutive times said energy measures of said number of energy measures equal or exceed a reference value, wherein

only the energy measures of said number of energy measures having values less than said reference value are used to determine said second value, and

said specific event is a predetermined number of consecutive times said energy measures of said number of energy measures equal or exceed said reference value.

11. The method according to claim 5, further comprising the steps of:

determining the lesser of two values T₁ and T₂,

multiplying said lesser value of T₁ and T₂ by two to obtain a product;

comparing said product to a value of −

21 dBm;

assigning the lesser value of −

21 dBm and said product to said noise threshold value for an updating period, τ

_p.

13. The method according to claim 11, further comprising the steps of:

measuring the maximum block energy occurring during said updating period, τ

_p, and assigning said measured maximum block energy to E_max;

measuring the minimum block energy occurring during said updating period, τ

_p, and assigning said measured maximum block energy to E_min;

calculating said value of T₁ given by the equation T₁=E_max+(E_max−

E_min)/32; and

calculating said value of T₂ given by the equation T₂=4* E_min.

14. The method according to claim 12, further comprising the steps of:

measuring the maximum block energy occurring during said updating period, τ

_p, and assigning said measured maximum block energy to E_max;

measuring the minimum block energy occurring during said updating period, τ

_p, and assigning said measured maximum block energy to E_min;

calculating said value of T₁ given by the equation T₁=E_min+(E_max−

E_min)/32; and

calculating said value of T₂ given by the equation T₂4* E_min.

16. The method according to claim 12, further comprising the step of:

updating said noise threshold value about every 1.28 seconds during a communication link.

9. A method of converging an ITU Recommendation G.729 Annex B voice activity detection (VAD) device, comprising the steps of:

determining a noise identification threshold value;

comparing a number of energy measures of a signal to said noise threshold value;

determining a differential spectral distance, Δ

SD, between a current spectral state of said signal and a value representing an average of a number of prior spectral states of said signal;

updating a first set of values representing averages of said signal'"'"'s noise characteristics, when said energy measure is less than said noise threshold;

updating a second set of values representing averages of said signal'"'"'s noise characteristics, when said energy measure is less than a reference value and said differential spectral distance has a value less than about 0.0637; and

substituting said first value for said second value when a specific event occurs.

10. The method according to claim 9, further comprising the step of:

counting the number of consecutive times said energy measures of said number of energy measures equal or exceed said reference value, wherein

said specific event is a predetermined number of consecutive times said energy measures of said number of energy measures equal or exceed said reference value.

12. The method according to claim 9, further comprising the steps of:

determining the lesser of two values T₁ and T₂,

multiplying said lesser value of T₁ and T₂ by two to obtain a product;

comparing said product to a value of −

21 dBm;

assigning the lesser value of −

21 dBm and said product to said noise threshold value for an updating period, τ

_p.

15. A method of converging an ITU Recommendation G.729 Annex B voice activity detection (VAD) device, comprising the steps of:

measuring the maximum block energy occurring during an updating period, τ

_p, and assigning said measured maximum block energy to E_max;

measuring the minimum block energy occurring during said updating period, τ

_p, and assigning said measured maximum block energy to E_min;

calculating a value of T₁ given by the equation T₁=E_min+(E_max−

E_min)/32;

calculating a value of T₂ given by the equation T₂=4* E_min;

determining the lesser value of said values T₁ and T₂,

multiplying said lesser value of T₁ and T₂ by two to obtain a product;

comparing said product to a value of −

21 dBm;

assigning the lesser value of −

21 dBm and said product to a noise threshold value;

comparing a number of energy measures of a signal to said noise threshold value;

determining a differential spectral distance, Δ

SD, between a current spectral state of said signal and a value representing an average of a number of prior spectral states of said signal;

updating a first set of values representing averages of said signal'"'"'s noise characteristics, when said energy measure is less than said noise threshold;

updating a second set of values representing averages of said signal'"'"'s noise characteristics, when said energy measure is less than a reference value and said differential spectral distance has a value less than about 0.0637;

counting the number of consecutive times said energy measures of said number of energy measures equal or exceed said reference value; and

substituting said first value for said second value when said number of consecutive times exceeds a predetermined value.