Method for talk/listen determination and multipoint conferencing system using such method

• US 5,768,263 A
• Filed: 10/20/1995
• Issued: 06/16/1998
• Est. Priority Date: 10/20/1995
• Status: Expired due to Term

• Current Assignee: Cisco Technology, Inc. (Cisco Systems, Inc.)
• Original Assignee: Vtel Corporation
• Inventors: Tischler, Paul V., Clements, Bill
• Primary Examiner(s): Kizou, Hassan
• Assistant Examiner(s): BNIMOUSSA, ABDELHAMID
• Application Number:
  US08/546,276
• Time in Patent Office: 970 Days
• Field of Search: 370/260, 370/263, 370/264, 370/266, 370/270, 370/465, 379/157, 379/158, 379/202
• US Class Current: 370/263
• CPC Class Codes: H04M 2201/40 using speech recognition
• CPC Class Codes: H04M 3/567 Multimedia conference systems
• CPC Class Codes: H04M 3/569 using the instant speaker's algorithm speech detection per se G10L25/78

Abstract

A method for determining a talk/listen state using a dynamic threshold is provided that can be implemented in a multipoint control unit (12). An audio sample is received and represents sound measured during a sample time interval. An audio level is generated from the audio sample The audio level represents an average power level of the audio sample. A dynamic threshold is set by processing the audio level. The audio level is compared to the dynamic threshold. A talk/listen state is determined to be a talking state if the audio level is above the dynamic threshold and to be a listening state if the audio level is below the dynamic threshold. The previous steps are repeated for each audio sample in a sequential stream of received audio samples such that the dynamic threshold is maintained and used to determine the talk/listen state.

Claims

1. A method for determining a talk/listen state using a dynamic threshold, comprising:

receiving an audio sample representing sound measured during a sample time interval;

generating an audio level from the audio sample, the audio level representing an average power level of the audio sample;

setting a dynamic threshold by processing the audio level to set and maintain a dynamic level and using the dynamic level to determine a value for the dynamic threshold;

comparing the audio level to the dynamic threshold;

determining a talk/listen state to be a talking state if the audio level is above the dynamic threshold and to be a listening state if the audio level is below the dynamic threshold; and

repeating the previous steps for each audio sample in a sequential stream of received audio samples such that the dynamic threshold is maintained and used to determine the talk/listen state.

2. The method of claim 1, wherein the method is implemented in a multipoint control unit for a multipoint conference and receiving an audio sample comprises receiving an audio sample from a conference site.

3. The method of claim 1, wherein receiving an audio sample comprises receiving an audio data packet that is part of a data stream including audio and video information.

4. The method of claim 3, wherein receiving the audio data packet comprises receiving an audio data packet representing sound measured over an interval of approximately twenty milliseconds.

5. The method of claim 1 wherein generating an audio level comprises assigning to the average power level a value from zero to a maximum value.

6. The method of claim 5, wherein assigning to the average power level a value comprises assigning a value from zero to 255.

7. The method of claim 1, wherein the dynamic level is a background level, and wherein setting a dynamic threshold comprises processing the audio level to set and maintain the background level and using the background level to determine a value for the dynamic threshold.

8. The method of claim 7, wherein setting a dynamic threshold further comprises processing the audio level to set and maintain a foreground level and using the background level and the foreground level to determine a value for the dynamic threshold.

9. The method of claim 8, wherein setting a dynamic threshold further comprises processing the audio level to set and maintain a long term background level and using the background level, the foreground level and the long term background level to determine a value for the dynamic threshold.

10. The method of claim 9, wherein setting a dynamic threshold comprises setting the dynamic threshold as a weighted sum of the background level, the foreground level and the long term background level.

11. The method of claim 10, wherein setting a dynamic threshold comprises a weighted sum with a ratio of 1:

2;

4 with respect to the long term background level, the background level and the foreground level.

12. The method of claim 1, wherein the dynamic level is a foreground level, and wherein setting a dynamic threshold comprises processing the audio level to set and maintain the foreground level and using the foreground level to determine a value for the dynamic threshold.

13. The method of claim 1, wherein the dynamic level is a long term background level, and wherein setting a dynamic threshold comprises processing the audio level to set and maintain the long term background level and using the long term background level to determine a value for the dynamic threshold.

14. A method for determining a talk/listen state using a dynamic threshold, comprising:

receiving an audio data packet representing sound measured during a packet time interval;

generating an audio level from the audio data packet, the audio level representing an average power level of the audio data packet;

setting a long term background level;

setting a foreground level by weighting the foreground level with the audio level;

setting a background level by weighting the background level with the audio level;

setting a dynamic threshold equal to a weighted sum of the long term background level, the foreground level and the background level;

comparing the audio level to the dynamic threshold;

determining a talk/listen state to be a talking state if the audio level is above the dynamic threshold and to be a listening state if the audio level is below the dynamic threshold; and

repeating the previous steps for each audio data packet in a sequential stream of received audio data packets such that the dynamic threshold and the long term background, background and foreground levels are maintained and used to determine the talk/listen state.

15. The method of claim 14, wherein the method is implemented in a multipoint control unit for a multipoint conference and receiving an audio data packet comprises receiving an audio data packet from conference equipment located at a conference site.

16. The method of claim 14, further comprising signaling a change in the talk/listen state if the talk/listen state is determined to change from listening to talking for a current audio data packet.

17. The method of claim 14, further comprising signaling a change in the talk/listen state if the talk/listen state is determined to change from talking to listening for a plurality of current audio data packets.

18. The method of claim 14, wherein receiving the audio data packet comprises receiving an audio data packet representing sound measured over an interval of approximately twenty milliseconds.

19. The method of claim 14, wherein generating an audio level comprises assigning to the average power level a value from zero to a maximum value.

20. The method of claim 14, wherein setting a dynamic threshold comprises a weighted sum with a ratio of 1:

2;

4 with respect to the long term background level, the background level and the foreground level.

21. A multipoint control unit, comprising:

a plurality of conference units, each conference unit operable to receive and process each audio sample in a sequential stream of audio samples received from an associated conference site;

each conference unit further operable to set and maintain a dynamic level, to maintain a dynamic threshold using the dynamic level, to compare an audio level of each audio sample to the dynamic threshold and to provide a notification signal indicating a talk/listen state of the associated conference site; and

a control unit coupled to the plurality of conference units and receiving each notification signal, the control unit operable to control voice activated switching of video between the associated conference sites using the notification signals.

22. The multipoint control unit of claim 21, wherein each conference unit comprises:

an audio memory operable to store compressed and expanded audio samples;

an audio processor coupled to the audio memory, the audio processor operable to expand compressed audio samples received from an associated conference site; and

a talk/listen determination unit coupled to the audio memory, the talk/listen determination unit operable to access the expanded audio samples in order to process the audio samples.

23. The multipoint control unit of claim 21, wherein the dynamic level is a background level, and wherein each conference unit is further operable to set and maintain the background level and use the background level to determine a value for the dynamic threshold.

24. The multipoint control unit of claim 23, wherein each conference unit is further operable to set and maintain a foreground level and use the background level and the foreground level to determine a value for the dynamic threshold.

25. The multipoint control unit of claim 24, wherein each conference unit is further operable to set and maintain a long term background level and use the background level, the foreground level and the long term background level to determine a value for the dynamic threshold.

26. The multipoint control unit of claim 25, wherein each conference unit is operable to set the dynamic threshold as a weighted sum of the background level, the foreground level and the long term background level.

27. The multipoint control unit of claim 26, wherein the weighted sum is a ratio of 1:

2;

4 with respect to the long term background level, the background level and the foreground level.

28. The multipoint control unit of claim 21, wherein the dynamic level is a foreground level, and wherein each conference unit is further operable to set and maintain the foreground level and use the foreground level to determine a value for the dynamic threshold.

29. The multipoint control unit of claim 21, wherein the dynamic level is a background level, and wherein each conference unit is further operable to set and maintain the long term background level and use the long term background level to determine a value for the dynamic threshold.

30. A multipoint conference system, comprising:

a plurality of sets of conference equipment, each set of conference equipment located at a conference site and operable to transmit a sequential stream of audio samples, where each audio sample represents sound measured from the conference site for a sampled interval of time;

a multipoint control unit coupled to the plurality of sets of conference equipment and receiving each sequential stream of audio samples, the multipoint conference system operable to set and maintain a dynamic level associated with each site, and set a separate dynamic threshold associated with each conference site by processing each dynamic level and audio sample in the sequential stream received from each set of equipment; and

the multipoint control unit further operable to compare an audio level of each audio sample to the dynamic threshold to determine a talk/listen state of each conference site and to control voice activated switching of video between the conference sites using the determined talk/listen states of the conference sites.

31. The multipoint conference system of claim 30, wherein the dynamic level is a background level, and wherein the multipoint control unit is further operable to set and maintain the background level and use the background level to determine a value for the dynamic threshold.

32. The multipoint conference system of claim 31, wherein the multipoint control unit is further operable to set and maintain a foreground level and use the background level and the foreground level to determine a value for the dynamic threshold.

33. The multipoint conference system of claim 32, wherein the multipoint control unit is further operable to set and maintain a long term background level and use the background level, the foreground level and the long term background level to determine a value for the dynamic threshold.

34. The multipoint conference system of claim 33, wherein the multipoint control unit is operable to set the dynamic threshold as a weighted sum of the background level, the foreground level and the long term background level.

35. The multipoint conference system of claim 34, wherein the weighted sum is a ratio of 1:

2;

4 with respect to the long term background level, the background level and the foreground level.

36. The multipoint conference system of claim 30, wherein the dynamic level is a foreground level, and wherein the multipoint control unit is further operable to set and maintain the foreground level and use the foreground level to determine a value for the dynamic threshold.

37. The multipoint conference system of claim 30, wherein the dynamic level is a long term background level, and wherein the multipoint control unit is further operable to set and maintain the long term background level and use the long term background level to determine a value for the dynamic threshold.