Mass-scale, user-independent, device-independent, voice messaging system

• US 8,654,933 B2
• Filed: 10/31/2007
• Issued: 02/18/2014
• Est. Priority Date: 02/10/2006
• Status: Expired due to Fees

• Current Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Original Assignee: Nuance Communications, Inc. (Microsoft Corporation)
• Inventors: Doulton, Daniel Michael
• Primary Examiner(s): Gauthier, Gerald
• Assistant Examiner(s): King, Simon
• Application Number:
  US11/932,146
• Publication Number:
  US 20080133219A1
• Time in Patent Office: 2,302 Days
• Field of Search: 704/235, 704/8, 704/9, 704/231, 704/246, 704/215, 375/224, 379/88.22, 379/88.19, 379/93.24, 379/142.01, 379/88.01, 701/208, 707/100
• US Class Current: 379/88.01
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence
• CPC Class Codes: H04M 2201/60 Medium conversion
• CPC Class Codes: H04M 3/4936 Speech interaction details speech recognition per se G10L15/00
• CPC Class Codes: H04M 3/5183 Call or contact centers with computer-telephony arrangements
• CPC Class Codes: H04M 3/53333 Message receiving aspects

Abstract

A mass-scale, user-independent, device-independent, voice messaging system that converts unstructured voice messages into text for display on a screen is disclosed. The system comprises (i) computer implemented sub-systems and also (ii) a network connection to human operators providing transcription and quality control; the system being adapted to optimize the effectiveness of the human operators by further comprising 3 core sub-systems, namely (i) a pre-processing front end that determines an appropriate conversion strategy; (ii) one or more conversion resources; and (iii) a quality control sub-system.

Claims

1. A voice messaging system for converting an audio message from a caller into text, the voice messaging system comprising:

at least one automatic speech recognition (ASR) system to automatically recognize at least some of the audio message;

a computer implemented preprocessing front-end to process the audio message from the caller and to detect if the audio message contains no voice content, wherein;

if the preprocessing front-end detects that the audio message contains no voice content, the preprocessing front-end does not provide the audio message to the ASR component; and

if the preprocessing front-end detects that the audio message contains voice content, the front-end provides the audio message to the ASR component, and

wherein the computer implemented preprocessing front-end comprises a computer implemented speech quality detector to determine at least one measure of speech quality of the voice content of the audio message, and wherein the speech quality detector detects drop-outs, estimates noise levels and/or calculates an overall measure of voice quality using an adaptive threshold to reject lowest quality messages.

2. The system of claim 1, wherein when the preprocessing front-end indicates that the audio message contains no voice content, the preprocessing front-end generates text indicating that the caller called but did not leave a meaningful message.

3. The voice messaging system of claim 2, wherein the voice messaging system transmits the text indicating that the caller called but did not leave a meaningful message to an intended recipient of the audio message.

4. The system of claim 1, wherein the preprocessing front-end further comprises a computer implemented inadvertent call detector to detect inadvertent calls.

5. The system of claim 4, wherein when the inadvertent call detector indicates that a call was inadvertent, the audio message is not provided to the ASR component.

6. The system of claim 1, wherein the audio message corresponds to a voicemail intended for a user of a mobile telephone, and wherein the voice content is converted to text, at least in part by the ASR component, and sent to the mobile telephone.

7. The system of claim 1, wherein the audio message corresponds to a voice message intended for an instant messaging service, and wherein the voice content is converted to text, at least in part by the ASR component, and sent to an instant messaging service for a display on a screen.

8. The system of claim 1, wherein the audio message corresponds to a voice message intended for a web service, and wherein the voice content is converted to text, at least in part by the ASR component, and sent to a server for display as part of the web service.

9. The system of claim 1, wherein the audio message corresponds to a voice message, and wherein the voice content is converted to text format, at least in part by the ASR component, and sent as a text message.

10. The system of claim 1, wherein the audio message corresponds to a voice message, and wherein the voice content is converted to text format, at least in part by the ASR component, and sent as an email message.

11. The system of claim 1, wherein the audio message corresponds to a voice message spoken by an originator, and wherein the voice content is converted to text format, at least in part by the ASR component, and sent by email and/or text to the originator of the voice message.

12. The voice messaging system of claim 1, wherein the audio message is transcribed, at least in part, by a human operator to produce the text.

13. A method comprising:

receiving an audio message from a caller;

determining at least one measure of speech quality of the voice content of the audio message including detecting drop-outs, estimating noise levels and/or calculating an overall measure of voice quality using an adaptive threshold to reject lowest quality messages;

processing the audio message to determine if the audio message contains voice content;

providing the audio message to an automatic speech recognition component to convert, at least in part, the voice content to text if the audio message is determined to contain voice content;

not providing the audio message to an automatic speech recognition component if the audio message is determined to contain no voice content.

14. The method of claim 13, further comprising generating a text indicating that the caller initiated a call to a recipient, but did not leave a meaningful message.

15. The method of claim 14, further comprising sending the text indicating that the caller initiated a call but did not leave a meaningful message to the recipient.

16. The method of claim 13, wherein the voice content is converted to text and sent to a recipient as a text message and/or an email message.

17. The method of claim 16, wherein the voice content is converted to text, at least in part, by the automatic speech recognition component.

18. The method of claim 17, wherein the voice content is converted to text, at least in part, by a human operator.

19. An apparatus comprising:

at least one input to receive an audio message from a caller; and

at least one processor capable of receiving the audio message, the at least one processor configured to;

determine at least one measure of speech quality of the voice content of the audio message including detecting drop-outs, estimating noise levels and/or calculating an overall measure of voice quality using an adaptive threshold to reject lowest quality messages;

process the audio message to determine if the audio message contains voice content;

provide the audio message to an automatic speech recognition component to convert, at least in part, the voice content to text if the audio message is determined to contain voice content; and

not provide the audio message to an automatic speech recognition component if the audio message is determined to contain no voice content.

20. The apparatus of claim 19, wherein the at least one processor is configured to generate a text indicating that the caller initiated a call to a recipient, but did not leave a meaningful message.

21. The apparatus of claim 20, wherein the at least one processor is configured to send the text indicating that the caller initiated a call but did not leave a meaningful message to the recipient.

22. The apparatus of claim 19, wherein the voice content is converted to text, and wherein the at least one processor is configured to send the text to a recipient as a text message and/or an email message.

23. The apparatus of claim 22, wherein the voice content is converted to text, at least in part, by the automatic speech recognition component.

24. The apparatus of claim 23, wherein the voice content is converted to text, at least in part, by a human operator.