Granted Patent B2

US 9,332,319 · App. 12/890,744 · Granted May 3, 2016

Abstract

Claims (52)

1. A method of converting speech to text, the method comprising:

analyzing multimedia content using one or more computing devices to determine the presence of dosed captioning data;

upon detecting closed captioning data, causing at least one of the one or more computing devices to begin:

i) indexing the closed captioning data as associated with the multimedia content;

upon failure to detect closed captioning data in the multimedia content causing at least one of the one or more computing devices to begin:

i) extracting audio data from multimedia content, the audio data including speech data;

ii) performing a plurality of different speech to text conversions on the speech data to create a plurality of transcripts of the speech data, the plurality of different speech to text conversions include speech to text conversion processes from different software vendors, wherein at least one of the plurality of transcripts is different from a remainder of the plurality of transcripts, wherein at least one of the speech to text conversions uses a context-sensitive speech to text dictionary selected according to the subject matter of the multimedia content;

iii) selecting text from among the plurality of transcripts to form an amalgamated transcript; and

iv) indexing the amalgamated transcript as associated with the multimedia content, wherein indexing the amalgamated transcript includes storing metadata associating text in the amalgamated transcript to timestamps associated with the multimedia content.

2. The method of claim 1 , further comprising:

analyzing the multimedia content to detect subtitle information; and

upon detecting subtitle information, indexing the subtitle information as associated with the multimedia content.

3. The method of claim 1 , further comprising associating the indexed closed captioning data or amalgamated transcript with metadata describing the multimedia content.

4. The method of claim 1 , wherein indexing the amalgamated transcript comprises associating indexed portions of the multimedia content with a query engine.

5. The method of claim 1 , wherein each of the plurality of speech to text conversions is performed by a corresponding speech to text conversion program.

6. The method of claim 5 , further comprising training the speech to text program.

7. The method of claim 6 , wherein training the speech to text conversion program includes training the speech to text conversion program using speech patterns specific to a presenter whose speech data is included in the multimedia content.

8. The method of claim 6 , wherein training the speech to text conversion program includes training the speech to text conversion program using a speech to text dictionary in a training mode of the speech to text conversion program.

9. The method of claim 8 , wherein the speech to text dictionary is a context-sensitive dictionary selected according to the subject matter of the multimedia content.

10. The method of claim 1 , further comprising selecting a subject-specific speech to text dictionary related to the multimedia content.

11. The method of claim 1 , further comprising receiving a script alongside the multimedia content and including at least a portion of the script in a transcript indexed to the multimedia content.

12. The method of claim 1 wherein the one or more computing devices comprise a single computing device.

13. A system for converting speech to text, the system comprising:

one or more computing systems each including a programmable circuit and a memory, the one or more computing systems executing program instructions, which, when executed, cause the one or more computing systems to:

i) analyze multimedia content to determine the presence of closed captioning data;

ii) upon detecting closed captioning data: index the closed captioning data as associated with the multimedia content;

iii) upon failure to detect closed captioning data in the multimedia content: extract audio data from multimedia content, the audio data including speech data; perform a plurality of different speech to text conversions on the speech data to create a plurality of transcripts of the speech data, the plurality of different speech to text conversions include speech to text conversion processes from different software vendors, wherein at least one of the plurality of transcripts is different from a remainder of the plurality of transcripts, and wherein at least one of the speech to text conversions uses a context-sensitive speech to text dictionary selected according to the subject matter of the multimedia content; select text from among the plurality of transcripts to form an amalgamated transcript; and index the amalgamated transcript as associated with the multimedia content, wherein indexing the amalgamated transcript includes storing metadata associating text in the amalgamated transcript to timestamps associated with the multimedia content.

14. The system of claim 13 , wherein the closed captioning data or amalgamated transcript is associated with metadata describing the multimedia content.

15. A system for converting speech to text, the system comprising:

an analysis module operating on one or more computing systems and configured to analyze multimedia content to determine the presence of closed captioning data;

an audio extraction module operating on one or more computing systems, the audio extraction module configured to extract audio data from multimedia content, the audio data including speech data;

a plurality of speech to text conversion programs operating on one or more computing systems, each of the plurality of speech to text conversion programs operating on the speech data to create a plurality of transcripts of the speech data, the plurality of speech to text conversion programs include programs from different software vendors, and at least one of the plurality of speech to text conversion programs uses a context-sensitive speech to text dictionary selected according to the subject matter of the multimedia content, wherein at least one of the plurality of transcripts is different from a remainder of the plurality of transcripts;

a transcript selection module configured to select text from one or more of the plurality of transcripts to form an amalgamated transcript; and

an indexing module operating on the one or more computing systems and configured to,

upon detecting closed captioning data:

index the closed captioning data as associated with the multimedia content; and

upon failure to detect closed captioning data in the speech data:

index the amalgamated transcript as associated with the multimedia content, wherein indexing the amalgamated transcript includes storing metadata associating text in the amalgamated transcript to timestamps associated with the multimedia content.

16. The system of claim 15 , wherein the analysis module is further configured to analyze the multimedia content to detect subtitle information, and the indexing module is further configured to, upon detecting subtitle information, index the subtitle information as associated with the multimedia content.

17. The system of claim 15 , further comprising a training module operating on the one or more computing systems and configured to train the one or more speech to text programs using speech patterns specific to a presenter whose speech data is included in the multimedia content.

18. The system of claim 15 , wherein the training module selects a subject-specific speech to text dictionary related to the multimedia content.

19. The system of claim 15 , wherein the closed captioning data or amalgamated transcript is associated with metadata describing the multimedia content.

20. The system of claim 15 , wherein the indexing module is configured to store metadata associating text in the amalgamated transcript to timestamps associated with the multimedia content.

21. A method of converting speech to text, the method comprising:

using one or more processors to train one or more speech to text programs using a context-sensitive speech to text dictionary selected according to the subject matter of the multimedia content;

analyzing the extracted speech data using at least one of the one or more processors to determine the presence of closed captioning data;

upon detecting closed captioning data: using at least one of the one or more processors to begin indexing the closed captioning data as associated with the multimedia content;

upon failure to detect closed captioning data in the speech data, causing at least one of the one or more processors to begin:

i) extracting audio data from multimedia content, the audio data including speech data; performing a plurality of different speech to text conversions on the speech data using the one or more speech to text programs to create a plurality of transcripts of the speech data, wherein each of the plurality of transcripts is different from a remainder of the plurality of transcripts and the speech to text programs each associated with different software vendors;

ii) selecting text from one or more of the plurality of transcripts to form an amalgamated transcript; and

iii) indexing the amalgamated transcript as associated with the multimedia content by storing metadata associating text in the amalgamated transcript to timestamps associated with the multimedia content.

22. The method of claim 21 wherein the one or more processors comprise a single processor.

Assignments (12)