Reconciliation of transcripts
The method includes identifying a plurality of transcripts of an audio event. The method further identifying a difference between two or more of the plurality of transcripts of the audio event. The method further includes determining a confidence level for the two or more transcripts that include the identified difference, wherein the confidence level indicates a measure of quality of the transcript. The method further includes selecting a difference from the two or more transcripts that include the identified difference based on the determined confidence level. The method further includes generating a transcript based on the selected difference.
1. A method for generating a transcript, the method comprising:
identifying, by one or more computer processors, at least a first transcript and a second transcript of an audio event, wherein the first transcript and the second transcript are generated from the same audio;
determining, by one or more computer processors, the overall confidence level for the at least the first transcript and the second transcript, wherein the overall confidence level indicates a measure of quality of the transcript;
identifying, by one or more computer processors, a difference in a first portion of the first transcript and a corresponding first portion of the second transcript of the audio event, wherein the different first portions have individual confidence levels from an overall confidence level of each of the at least first transcript and the second transcript;
identifying, by one or more computer processors, a difference in a second portion of the first transcript and a corresponding second portion of the second transcript of the audio event, wherein the different second portions have individual confidence levels from an overall confidence level of each of the at least first transcript and the second transcript;
selecting, by one or more computer processors, a first portion of the first transcript, wherein the first portion of the first transcript has a highest individual confidence level for the first portion of the at least a first transcript and a second transcript of the audio event;
selecting, by one or more computer processors, a second portion in the second transcript, wherein the second portion of the second transcript has a highest individual confidence level for the second portion of the at least a first transcript and a second transcript of the audio event; and
generating, by one or more computer processors, a transcript to include the selected first portion of the first transcript and the selected second portion of the second transcript.