Method and system for sharing speech processing resources over a communication network

• US 8,706,501 B2
• Filed: 12/09/2004
• Issued: 04/22/2014
• Est. Priority Date: 12/09/2004
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Inventors: Mandalia, Baiju D., Mansey, Pradeep P.
• Primary Examiner(s): Shah, Paras D
• Application Number:
  US11/007,985
• Publication Number:
  US 20060129406A1
• Time in Patent Office: 3,421 Days
• Field of Search: 704/270.1, 455/412.1, 455/436, 455/445, 455/455, 379/219, 379/221.01, 709/201, 709/203, 709/217, 709/219
• US Class Current: 704/270.1
• CPC Class Codes: G10L 13/00 Speech synthesis; Text to speech systems
• CPC Class Codes: G10L 15/30 Distributed recognition, e.g. in client-server systems, for mobile phones or network applications

Abstract

A method and system (40) for sharing speech processing resources (54) over a (communication network (21) for handling multiple client types (100, 101, etc.) and multiple media protocol types. The system can include a router (400) coupled to the communication network, a speech response system (500) coupled to the router, and a server (600) coupled to the speech response system and the router. The server can include at least one processor programmed to determine a media protocol and a client type of a client used for speech communication with the server, adapt at least one among encoding or decoding for the speech communication based on the media protocol and the client type, and dynamically and adaptively configure of the speech processing resources based on the media protocol and the client type.

Claims

1. A method for providing speech processing services comprising acts of:

receiving, by a server comprising at least one processor, a speech processing request from one of a plurality of clients;

identifying one media protocol type of the received speech processing request from among a plurality of media protocol types, wherein the plurality of media protocol types include a real-time protocol for data received over a public switched telephone network, a real-time protocol for high-quality dictation data received over a network, a real-time voice over internet protocol, and a non-real time protocol; and

adapting, at least in part, a speech resource by setting decoding logic for the speech resource in accordance with the identified media protocol type identified in the speech processing request.

2. The method of claim 1, wherein identifying comprises identifying a complexity of the received speech processing request and adapting the resource comprises allocating processing resources and memory for the speech resource responsive to a level of complexity identified in the speech processing request.

3. The method of claim 1, wherein identifying comprises identifying a language of the speech processing request.

4. The method of claim 1, wherein the received speech processing request includes a header having at least two of a client type, the media-protocol type, an application type, and a language type.

5. The method of claim 1, further comprising processing the speech processing request with the adapted speech resource.

6. The method of claim 1, wherein adapting comprises setting a language property for the speech resource responsive to a language identified in the speech processing request.

7. The method of claim 1, wherein adapting comprises determining a format for speech processing results delivered from the speech resource in accordance with a client type identified in the speech processing request.

8. The method of claim 1, wherein adapting comprises configuring a media-protocol application programming interface of the resource in accordance with a client type identified in the speech processing request and an application complexity identified in the speech processing request.

9. The method of claim 1, further comprising negotiating, by the server, properties of the data exchanged between the one of the plurality clients and the server.

10. The method of claim 9, wherein the negotiating comprises specifying a simplified voice quality format comprising either a first form for voice data delivered via a public switched telephone network or a second form for high quality audio data.

11. The method of claim 9, wherein the negotiating comprises specifying a voice packetization format.

12. At least one non-transitory computer-readable storage medium comprising computer-readable instructions that when executed by at least one processor cause the at least one processor to execute acts of:

receiving, by a server comprising at least one processor, a speech processing request from one of a plurality of clients;

identifying one media protocol type of the received speech processing request from among a plurality of media protocol types, wherein the plurality of media protocol types include a real-time protocol for data received over a public switched telephone network, a real-time protocol for high-quality dictation data received over a network, a real-time voice over internet protocol, and a non-real time protocol; and

adapting, at least in part, a speech resource by setting decoding logic for the speech resource in accordance with the identified media protocol type identified in the speech processing request.

13. The at least one non-transitory computer-readable storage medium of claim 12, wherein identifying comprises identifying a complexity of the received speech processing request and adapting the resource comprises allocating processing resources and memory for the speech resource responsive to a level of complexity identified in the speech processing request.

14. The at least one non-transitory computer-readable storage medium of claim 12, wherein identifying comprises identifying a language of the speech processing request.

15. The at least one non-transitory computer-readable storage medium of claim 12, wherein the received speech processing request includes a header having at least two of a client type, the media-protocol type, an application type, and a language type.

16. The at least one non-transitory computer-readable storage medium of claim 12, further comprising computer-readable instructions that when executed by the at least one processor cause the at least one processor to process the speech processing request with the adapted speech resource.

17. The at least one non-transitory computer-readable storage medium of claim 12, wherein adapting comprises setting a language property for the speech resource responsive to a language identified in the speech processing request.

18. The at least one non-transitory computer-readable storage medium of claim 12, wherein adapting comprises determining a format for speech processing results delivered from the speech resource in accordance with a client type identified in the speech processing request.

19. The at least one non-transitory computer-readable storage medium of claim 12, wherein adapting further comprises configuring a media-protocol application programming interface of the resource in accordance with a client type identified in the speech processing request and an application complexity identified in the speech processing request.

20. The at least one non-transitory computer-readable storage medium of claim 12, further comprising instructions for causing the at least one processor to negotiate properties of the data exchanged between the one of the plurality clients and the at least one processor.

21. The at least one non-transitory computer-readable storage medium of claim 20, wherein the negotiating comprises specifying a simplified voice quality format comprising either a first form for voice data delivered via a public switched telephone network or a second form for high quality audio data.

22. The at least one non-transitory computer-readable storage medium of claim 20, wherein the negotiating comprises specifying a voice packetization format.

23. A system for providing speech processing comprising:

a server comprising at least one processor, the server configured to receive speech processing requests from a plurality of clients, and upon receiving a speech processing request from one of the plurality of clients, to identify one media-protocol type of the received speech processing request from among a plurality of media protocol types, wherein the plurality of media protocol types include a real-time protocol for data received over a public switched telephone network, a real-time protocol for high-quality dictation data received over a network, a real-time voice over internet protocol, and a non-real time protocol; and

a speech resource configured to be adapted, at least in part, by the server by setting decoding logic for the speech resource in accordance with the identified media protocol type identified in the speech processing request.

24. The system of claim 23, wherein the server is further configured to identify a complexity of the received speech processing request and allocate processing resources and memory for the speech resource responsive to a level of complexity identified in the speech processing request.

25. The system of claim 23, wherein the server is further configured to identify a language of the speech processing request.

26. The system of claim 23, wherein the speech processing request includes a header having at least two of a client type, the media-protocol type, an application type, and a language type.

27. The system of claim 23, wherein the server is further configured to process the speech processing request with the adapted speech resource.

28. The system of claim 23, wherein the server is further configured to set a language property for the speech resource responsive to a language identified in the speech processing request.

29. The system of claim 23, wherein the server is further configured to set a format for speech processing results delivered from the speech resource in accordance with a client type identified in the speech processing request.

30. The system of claim 23, wherein the server is further configured to configure a media-protocol application programming interface of the speech resource in accordance with a client type identified in the speech processing request and an application complexity identified in the speech processing request.

31. The system of claim 23, wherein the server is further configured to negotiate properties of the data exchanged between the one of the plurality of clients and the server.

32. The system of claim 31, wherein the negotiating comprises specifying a simplified voice quality format comprising either a first form for voice data delivered via a public switched telephone network or a second form for high quality audio data.

33. The system of claim 31, wherein the negotiating comprises specifying a voice packetization format.

34. The system of claim 23, further comprising a plurality of speech resources, wherein each of the plurality of speech resources is substantially similar to the speech resource and each of the plurality of speech resources is configured to be adapted by the server to be able to process the received speech processing request.

35. The system of claim 23, further comprising a router configured to select the server for processing a speech processing request based upon a geographic proximity of the server to the one of the plurality of clients.