Sound event processing with echo analysis

• US 7,822,192 B2
• Filed: 03/30/2005
• Issued: 10/26/2010
• Est. Priority Date: 03/30/2005
• Status: Active Grant

• Current Assignee: Applied Voice & Speech Tech Incorporated
• Original Assignee: Applied Voice & Speech Technologies Incorporated (Open Text Corporation)
• Inventors: Shmalko, Michael V., Gierach, Karl Daniel
• Primary Examiner(s): Matar; Ahmad F
• Assistant Examiner(s): Asanbayev; Oleg
• Application Number:
  US11/096,004
• Publication Number:
  US 20060227964A1
• Time in Patent Office: 2,036 Days
• Field of Search: 379/221.11, 379 671- 8828, 37920201-20709, 37937301-37305, 379/372, 37940601-40609, 370282-290
• US Class Current: 379/406.01
• CPC Class Codes: H04M 2201/40 using speech recognition
• CPC Class Codes: H04M 3/002 Applications of echo suppressors or cancellers in telephonic connections in two-way loud-speaking telephone systems H04M9/08, echo suppressors or cancellers per se H04B3/20
• 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 3/568 audio processing specific to telephonic conferencing, e.g. spatial distribution, mixing of participants echo suppression in two-way loud-speaking telephone systems H04M9/02; sound field processing per se H04S7/30
• CPC Class Codes: H04M 9/082 using echo cancellers echo cancellers per se H04B3/23

Abstract

Apparatus, methods, and machine-readable articles of manufacture enable a telephone system to distinguish between actual events and echoes. Events may include conferees'"'"' voice and DTMF commands. In selected embodiments, the system categorizes an event received on one channel as a real event if the energy of the event'"'"'s waveform is greater than concurrent energy of every other channel of the same conference. Otherwise, the event is stored. When the system receives a new event of the same type on a channel of the same conference during a predetermined period, it compares the energies of the new and stored events. If the energy of the new event exceeds that of the stored event, the new event becomes the stored event, and the old stored event is ignored. At the end of the period, the stored event is categorized as a real event and forwarded to appropriate application code.

Claims

1. A method of processing events in a telephone conference in which a plurality of telephone devices are connected via a plurality of channels, a channel per telephone device, the method comprising:

detecting real and echo commands in audio data received from each telephone device of the plurality of telephone devices during the conference, thereby obtaining detected commands;

categorizing each detected command in a real command category or in an echo command category based at least in part on one or more characteristics of said each detected command;

when said each detected command is categorized in the real command category, executing code based on type of said each detected command and the telephone device from which said each detected command originated, the code affecting state of a telephone system that connects the plurality of telephone devices in the conference; and

when said each detected command is categorized in the echo command category, ignoring said each detected command.

2. A method according to claim 1, wherein the step of categorizing comprises:

categorizing said each detected command based at least in part on energy level of the audio data from which said each detected command was detected.

3. A method according to claim 1, further comprising:

buffering in a separate buffer the audio data received from each channel of the plurality of channels during the conference, one channel per buffer;

wherein the step of categorizing comprises;

categorizing said each detected command based at least in part on comparison of energy level of the audio data from which said each detected command was detected to energy level of the audio data in a buffer other than the buffer corresponding to the audio data from which said each detected command was detected.

4. A method according to claim 1, further comprising:

buffering in a separate buffer the audio data received from each channel of the plurality of channels during the conference, one channel per buffer;

wherein the step of categorizing comprises;

categorizing said each detected command based at least in part on comparison of energy level of the audio data from which said each detected command was detected to energy level of the audio data in each buffer other than the buffer corresponding to the audio data from which said each detected command was detected.

5. A method according to claim 1, wherein the one or more characteristics comprise command type and command average energy.

6. A method according to claim 1, further comprising:

buffering in a separate buffer the audio data received from said each telephone device during the conference, one telephone device per buffer;

wherein the step of categorizing comprises;

(1) categorizing said each detected command in the echo command category if another detected command of the same type as said each detected command was categorized in the real command category within a first predetermined time period of detection of said each detected command; and

(2) categorizing said each detected command in the real command category if (i) said each detected command has not been categorized in the echo command category in step (1), (ii) energy level of the audio data from which said each detected command was detected is greater than energy level of the audio data buffered during said each detected command in each buffer other than the buffer corresponding to the audio data from which said each detected command was detected, and (iii) the energy level of the audio data from which said each detected command was detected is not less than energy levels of the audio data from which every detected command of the same type as said each detected command is detected within a second predetermined time period following detection of said each detected command.

7. A method according to claim 6, further comprising:

varying at least one of the first predetermined time period and the second predetermined time period in response to computational load of the telephone system.

21. A method according to claim 5, wherein the plurality of telephone devices comprises at least three telephone devices.

22. A method according to claim 1, wherein the type of said each detected command is selected from a group consisting of (i) command types that cause a telephone system that connects the plurality of telephone devices in the telephone conference to change current or future connections, (ii) command types that cause a telephone system that connects the plurality of telephone devices in the telephone conference to receive and store data, and (iii) command types that cause a telephone system that connects the plurality of telephone devices in the telephone conference to send data.

23. A method according to claim 1, wherein said code corresponds to a tone command or a voice command of a participant in the conference.

8. A method of processing events in a telephone conference in which a plurality of telephone devices are connected through a plurality of channels, a channel per telephone device, the method comprising:

buffering in a separate circular buffer audio data received through each channel;

detecting real and echo commands in the audio data received through each channel during the conference, thereby obtaining detected commands;

determining types of the detected commands;

using one or more command characteristics to categorize a first detected command detected from the first audio data received through a first channel of the plurality of channels in a real command category or in an echo command category;

if the first detected command is categorized in the real command category, executing code based at least in part on type of the first detected command, the code affecting state of a telephone system that connects the plurality of telephone devices in the conference; and

when the first detected command is categorized in the echo command category, ignoring the first detected command.

9. A method according to claim 8, wherein the one or more command characteristics include average energy level in the buffer of the first channel.

10. A method according to claim 8, wherein the one or more command characteristics include average energy levels in each of the buffers.

11. A method according to claim 8, wherein the step of using comprises:

(1) categorizing the first detected command in the echo command category if another detected command of the same type as the first detected command was categorized in the real command category within a first predetermined time period of detection of the first detected command;

(2) categorizing the first detected command in the real command category if (i) no other detected command of the same type as the first detected command was categorized in the real command category within the first predetermined time period of detection of the first detected command, (ii) average energy level of the audio data in the buffer of the first channel is not less than average energy level of the audio data in buffer of each other channel, and (iii) the average energy level of the audio data in the buffer of the first channel is not less than average energy level of the audio data in buffer of each other channel that causes detection of a command of the same type as the first detected command within a second predetermined time period of detection of the first detected command; and

(3) categorizing the first detected command in the echo category if the first detected command was not categorized in the real command category in step (2).

12. A method according to claim 11, wherein the first predetermined period is between 100 milliseconds and 3 seconds.

13. A method of processing events in a telephone conference in which a plurality of telephone devices are connected through a plurality of channels, a channel per telephone device, the method comprising:

detecting commands in audio data received through each channel during the conference, thereby obtaining detected commands;

determining types of the detected commands;

using one or more command characteristics to categorize the detected commands in a real command category or in an echo command category;

in response to categorization of a detected command in the real command category, executing code based at least in part on type of the detected command, the code affecting state of a telephone system that connects the plurality of telephone devices in the conference; and

ignoring detected commands categorized in the echo command category.

14. A method according to claim 13, wherein command types comprise voice commands.

15. A method according to claim 13, wherein command types comprise tone commands.

16. A method according to claim 13, wherein command types comprise disconnect signaling.

17. A method of processing events in a telephone conference in which a plurality of telephone devices are connected through a plurality of channels, a channel per telephone device, the method comprising:

detecting commands in audio data received through each channel during the conference, thereby obtaining detected commands;

step for determining types of the detected commands;

step for using one or more command characteristics to categorize the detected commands as real or echo commands;

step for varying response of a telephone system that connects the plurality of telephone devices in the conference based on categorization of each detected command of a plurality of detected commands.

18. A telephone system comprising:

a plurality of computer telephony devices;

a plurality of trunk interfaces capable of connecting to a plurality of telephone devices through a plurality of channels, a channel per telephone device, each trunk interface of the plurality of trunk interfaces being coupled to a computer telephony device of the plurality of computer telephony devices;

an audio bus board coupled to each computer telephony device of the plurality of computer telephony devices; and

at least one processor coupled to a memory embedded with code, wherein the at least one processor, under control of the code, is configured to perform steps comprising;

connecting two or more of the trunk interfaces in a telephone conference;

buffering in separate circular buffers the audio data received through each of the two or more trunk interfaces;

detecting commands in the audio data received through each of the two or more trunk interfaces during the conference, thereby obtaining detected commands;

determining types of the detected commands;

using one or more command characteristics to categorize the detected commands, each detected command being categorized in a real command category or in an echo command category;

in response to categorization of a detected command in the real command category, executing instructions based at least in part on type of the detected command, the instructions changing state of the telephone system; and

discarding detected commands categorized in the echo command category.

19. A telephone system according to claim 18, wherein the audio bus board comprises a time division multiplexed communications bus with a plurality of TDM channels.

20. A telephone system according to claim 18, wherein the at least one processor, under control of the code, is further configured to categorize a first detected command based at least in part on average energy levels of the audio data in the circular buffers.