The Scorer
task measures the accuracy of a speech recognition transcript (such as that generated by the SpeechToText
task), when given a reference transcript file. The task reports scores for general word precision, recall, and the F-measure. You can choose the level of detail that the response includes.
Parameter | Description | Required |
---|---|---|
Type | The task name. Set to Scorer . |
Yes |
CtmFile | The .ctm file that was produced by speech-to-text on the audio file. | Yes |
Lang | The language pack to use as a foundation. | |
MatchType | The alignment mode. | |
Out | The .ctm file to write the ‘truth’ transcript including timestamps to. | Yes |
ScoreBeam | The score-based beam-width value to use during dynamic programming alignment. | |
ScoreFile | The file to write scores to. | Yes |
ShowAlignment | Whether the score file includes a section containing line-by-line matching of the reference transcript with the speech-to-text transcript. | |
ShowDetails | Whether to include detailed scoring information in the score file. | |
TxtFile | The normalized reference transcript file. | Yes |
WordBeam | The word region-based beam-width value to use during dynamic programming alignment. |
http://localhost:13000/action=AddTask&Type=Scorer&TxtFile=C:\data\RefTranscript.txt&CtmFile=C:\misc\Stt.ctm&ScoreFile=score&Out=FinalTranscript.ctm
This action uses port 13000
to instruct HPE IDOL Speech Server, which is located on the local machine, to score the Stt.ctm
recognition transcript by comparing it to the RefTranscript.txt
file. The action instructs HPE IDOL Speech Server to write the truth transcript containing timestamps to the FinalTranscript.ctm
file and the scores to the score.ctm
file.
|