The transcript checker performs a high-level alignment of a text transcript and the corresponding speech-to-text output, considering large blocks of text rather than individual words or phonemes. This process allows you to quickly determine if there is a mismatch between broad regions of the input files, and to provide approximate timestamps to the detailed alignment process to improve its performance.
Mode | Input | Output | Description |
---|---|---|---|
w
|
Accepts a word label stream. |
Example:
output ← transcheck (_, w)
BlockSize | MatchThreshold | OutputTranscriptLabel |
DiagFile | MaxWordCount | WarningFile |
DiagFileLabel | MinBlockSize | WarningFileLabel |
InputTranscript* | OutputTranscript* |
*You must set the value of these configuration parameters by setting the corresponding action parameters.
|