Method and apparatus of audio switching

• US 8,335,686 B2
• Filed: 04/27/2005
• Issued: 12/18/2012
• Est. Priority Date: 05/14/2004
• Status: Active Grant

• Current Assignee: Huawei Technologies Co., Ltd. (Huawei Investment & Holding Co., Ltd.)
• Original Assignee: Huawei Technologies Co., Ltd. (Huawei Investment & Holding Co., Ltd.)
• Inventors: Yu, Shuian
• Primary Examiner(s): Desir, Pierre-Louis
• Assistant Examiner(s): BAKER, MATTHEW H
• Application Number:
  US11/579,500
• Publication Number:
  US 20080040117A1
• Time in Patent Office: 2,792 Days
• Field of Search: 348/14, 370260-266, 370/147, 370/148, 370/447, 370/401, 379/202
• US Class Current: 704/228
• CPC Class Codes: H04M 2201/18 Comparators
• CPC Class Codes: H04M 3/56 Arrangements for connecting several subscribers to a common circuit, i.e. affording conference facilities video conference systems H04N7/15
• CPC Class Codes: H04M 3/569 using the instant speaker's algorithm speech detection per se G10L25/78

Abstract

A method for speech switching, including: extracting mute flags from encoded speech data transmitted by each of the terminals respectively in order to determine one or more non-mute terminals; decoding the encoded speech data of each non-mute terminal respectively; calculating speech energy of each non-mute terminal according to the decoded speech data of each non-mute terminal; comparing the speech energy of the non-mute terminals and selecting one or more terminals with relatively large speech energy; performing linear superposition with different combination methods to the decoded speech data of the selected terminals, and encoding the decoded speech data which is obtained from the linear superposition of the different combination methods, and transmitting the data to the corresponding terminals respectively. An apparatus for speech switching is also disclosed. With this invention, resource consumption during procedure of speech switching may be reduced, the effect of speech communication may be improved.

Claims

1. A method, implemented by a speech-switching device connected to a communication network, for speech switching between multiple terminals, comprising steps of:

(1) extracting, by the speech-switching device, a mute flag from encoded speech data transmitted by each of the terminals respectively, and determining one or more non-mute terminals;

(2) non-fully decoding, by the speech-switching device, the encoded speech data of each non-mute terminal respectively;

(3) calculating, by the speech-switching device, a speech energy value of each non-mute terminal according to the non-fully decoded speech data of each non-mute terminal;

(4) comparing, by the speech-switching device, the speech energy values of the non-mute terminals with each other and selecting one or more terminals with a largest speech energy;

(5) fully decoding, by the speech-switching device, the encoded speech data transmitted by each remaining selected terminal; and

(6) performing, by the speech-switching device, a linear superposition of different combination modes to the decoded speech data of the selected terminals, encoding, by the speech-switching device, the decoded speech data obtained after the linear superposition of the different combination modes, and transmitting the encoded data to the corresponding terminals respectively;

wherein when the terminals adopt speech coding/decoding protocol G.728, the non-fully decoding comprises;

extracting corresponding code vectors from excitation codebooks according to codebook indexes contained in the encoded speech data transmitted by the terminals;

performing a logarithmic gain prediction to the extracted code vectors; and

performing a decoding synthesis filtering process to the code vectors after the logarithmic gain prediction in order to obtain the corresponding non-fully decoded speech data.

2. The method of claim 1, wherein

an order of the logarithmic gain prediction is set to 4;

and

an order of the decoding synthesis filtering process is set to 8.

3. The method of claim 1, further comprising:

pre-determining the number of terminals with the largest speech energy to be selected before step (1);

if the number of the non-mute terminals determined based on the mute flags extracted in step (1) is larger than or equal to the pre-determined number, the pre-determined number of terminals with the largest speech energy are selected in step (4); and

if the number of the non-mute terminals determined based on the mute flags extracted in step (1) is smaller than the pre-determined number, all the non-mute terminals are selected as terminals with the largest speech energy in step (4).

4. The method of claim 1, wherein in step (3), by using the non-fully decoded speech data of the non-mute terminals, the speech energy of each non-mute terminal is calculated by the following formula:

5. The method of claim 1, wherein in step (3), according to the non-fully decoded speech data of the non-mute terminals, the speech energy of each non-mute terminal is calculated by the following formula:

6. The method of claim 1, wherein step (6) further comprises:

(51) for the selected terminals, isolating the decoded speech data of one selected terminal at a time in order, performing a linear superposition process to the decoded speech data of other selected terminals, encoding the superposed speech data and transmitting the data to the isolated terminal; and

(52) performing a uniform linear superposition process to the decoded speech data of all the selected terminals, encoding the superposed speech data and transmitting the data to the unselected terminals.

7. The method of claim 6, wherein step (52) further comprises:

for the terminals adopting a same speech encoding/decoding protocol among the unselected terminals, encoding the superposed speech data according to the protocol and transmitting the data to these unselected terminals adopting the same speech encoding/decoding protocol.

8. The method of claim 6, wherein the linear superposition of the decoded speech data of the selected terminals is implemented by the following formula:

S=λ

₁×

S₁+λ

₂×

S₂+ . . . +λ

_n×

S_n

in which, λ

₁, λ

₂ . . . λ

_n are weighting factors, and λ

₁+λ

₂+ . . . λ

_n=1;

S is the decoded speech data after linear superposition to be received by the terminals, S₁ is the decoded speech data of the first terminal, S₂ is the decoded speech data of the second terminal, . . . and S_n is the decoded speech data of the nth terminal.

9. The method of claim 1, wherein between step (4) and step (5), the method further comprises a step of extracting mute flag from the decoded speech data of each of the selected terminals respectively and filtering off the mute terminals from the selected terminals.

10. An apparatus for speech switching between multiple terminals, comprising:

a first mute flag extracting unit, configured to extract a mute flag from encoded speech data transmitted by each terminal respectively, and determine one or more non-mute terminals;

a first decoding unit, connected with the first mute flag extracting unit, and configured to non-fully decode the encoded speech data of the non-mute terminals after the mute flags are extracted;

a speech energy calculating unit, connected with the first decoding unit, and configured to calculate the speech energy value of each non-mute terminal according to the decoded speech data of each non-mute terminal;

a speech energy comparing unit, connected with the speech energy calculating unit, and configured to compare the speech energy values of the non-mute terminals with each other and select one or more terminals with a largest speech energy;

a second decoding unit configured to fully decode the encoded speech data transmitted by the terminals with the largest speech energy, wherein the speech energy comparing unit is connected with a linear superposition unit through the second decoding unit;

a linear superposition unit, connected with the speech energy comparing unit, and configured to perform a linear superposition in different combination methods to the decoded speech data of the selected terminals; and

an encoding unit, connected with the linear superposition unit, and configured to encode the decoded speech data after linear superposition and transmit the data to the corresponding terminals;

when the terminals adopt speech encoding/decoding protocol G.728, the first decoding unit further comprises;

an excitation codebook searcher, configured to extract the corresponding code vectors from excitation codebook according to the codebook indexes contained in the encoded speech data transmitted by the terminals;

a logarithmic gain predictor, configured to perform a logarithmic gain prediction to the extracted code vectors; and

a decoding synthesis filter, configured to perform a decoding synthesis filtering process to the code vectors after the logarithmic gain prediction in order to obtain the corresponding non-fully decoded speech data.

11. The apparatus of claim 10, wherein an order of the logarithmic gain prediction is set to 4;

and an order of the decoding synthesis filtering process is set to 8.

12. The apparatus of claim 10, wherein the apparatus further comprises:

a second mute flag extracting unit configured to extract the mute flags from the decoded speech data of each selected terminal and filter off the mute terminals from the selected terminals, wherein the second decoding unit is connected with the linear superposition unit through the second mute flag extracting unit.