Transcription of Speech

• US 20150066505A1
• Filed: 03/26/2013
• Published: 03/05/2015
• Est. Priority Date: 03/30/2012
• Status: Active Grant

• Current Assignee: Jpal Limited
• Original Assignee: Jpal Limited
• Inventors: Baker, John Richard
• Granted Patent: US 9,786,283 B2
• Time in Patent Office: Days
• US Class Current: 704/235
• CPC Class Codes: G06F 16/685 using automatically derived transcript of audio data, e.g. lyrics speech recognition G10L15/00
• CPC Class Codes: G06F 16/7844 using original textual content or text extracted from visual content or transcript of audio data
• CPC Class Codes: G10L 15/08 Speech classification or search
• CPC Class Codes: G10L 15/26 Speech to text systems G10L15/08 takes precedence
• CPC Class Codes: G10L 2015/088 Word spotting
• CPC Class Codes: G10L 25/78 Detection of presence or absence of voice signals switching of direction of transmission by voice frequency in two-way loud-speaking telephone systems H04M9/10
• CPC Class Codes: G10L 25/87 Detection of discrete points within a voice signal

Abstract

A speech media transcription system comprises a playback device arranged to play back speech delimited in segments. The system is programmed to provide, for a segment being transcribed, an adaptive estimate of the proportion of the segment that has not been transcribed by a transcriber. The device is arranged to play back that proportion of the segment, optionally after having already played back the entire segment. Additionally, a segmentation engine is arranged to divide speech media into a plurality of segments by identifying speech as such and using timing information but without using a machine conversion of the speech media into text or a representation of text.

Claims

1. A speech media transcription system comprising a playback device arranged to play back speech delimited in segments, the system being programmed to provide, for a segment being transcribed, an adaptive estimate of the proportion of the segment that has not been transcribed by a transcriber, the device being arranged to play back said proportion of the segment.

2. A transcription system according to claim 1, wherein the playback device is arranged to play back an entire segment prior to playing back said proportion estimated not to have been transcribed.

3. A transcription system according to claim 1, wherein the adaptive estimate is made on the basis of at least one of:

an amount of text likely to be present in the segment, a speaking speed of the speaker, a text input rate of the transcriber, a working memory of the transcriber or a current activity of the transcriber.

4. A system according to claim 1, wherein where analysis of a segment indicates that it contains no speech, the segment is omitted from playback or played at high speed.

5. A system according to claim 1, arranged to monitor a text input rate of the transcriber and to play back said speech at a speed depending on said text input rate.

6. A system according to claim 1, comprising a facility for automatically transferring a typed punctuation mark to a previously typed segment.

7. A system according to claim 1, comprising a facility for automatically replaying said speech for review and correction of a transcription.

8. A system according to claim 1, comprising a facility for allowing identification by a reviewer of segments of text considered erroneous.

9. A method of transcribing speech media, comprising providing a system according to claim 1, allowing a transcriber to transcribe part of a segment of the speech media, replaying said proportion of the segment estimated not to have been transcribed, and subsequently determining that said segment has been completely transcribed.

10. A segmentation engine arranged to divide speech media into a plurality of segments by identifying speech as such and using timing information but without using a machine conversion of the speech media into text or a representation of text.

11. A segmentation engine according to claim 10, programmed to commence a segment, to measure time elapsed, to determine gaps between words in the segment, and ending the segment at a gap when said time elapsed in the segment is at least 1 second, preferably at least 2 seconds.

12. A segmentation engine according to claim 11, wherein the segment is ended after an elapsed time of at most 15 seconds, preferably at most 10 seconds.

13. A segmentation engine according to claim 10, wherein timing information for the segmenting is based on cues in the speech media, selected from the group consisting of:

cues derived from an Automatic Speech Recognition system, cues derived from a Voice Activity Detection system and cues resulting from audio or video analysis of the speech media.

14. A segmentation engine according to claim 10, programmed to insert an additional segment into a gap between two segments, to provide contiguous segments.

15. A segmentation engine according to claim 10, programmed to eliminate a gap between two segments, to provide contiguous segments.

16. A segmentation engine according to claim 10, programmed to allow a user to split a segment into two or more segments.

17. A segmentation engine according to claim 10, programmed to allow a user to merge two or more successive segments into one segment.

18. A segmentation engine according to claim 10, programmed to preserve metadata from a segment in a subsequent segment in the absence of an indication that said metadata is inapplicable to said subsequent segment.

19. A method of segmenting speech media, comprising providing a segmentation engine according to claim 10, analysing timing information to determine gaps between words, and ending segments at selected ones of said gaps between words.