Run Transcript Alignment

To run transcript alignment you must have installed a speech-to-text language pack that matches the language of the speech. You must also enable the speech-to-text module. For information about how to install speech-to-text language packs, see Install Speech-to-Text Language Packs. For information about how to enable and disable modules, see Specify Modules to Enable.

To run transcript alignment, follow these steps.

To run transcript alignment

  1. Prepare a transcript of the speech in your media file. The transcript should be a plain text file.
  2. Send the transcript and the media file to the AlignAudioTranscript action. For example:

    curl http://localhost:14000 -F action=AlignAudioTranscript
                                -F audiodata=@audio.wav
                                -F textdata=@transcript.txt
                                -F languagepack=ENUS
                                -F samplefrequency=16000
    

    Media Server returns a token. You can use the token with the QueueInfo action to retrieve the results. For more information about the AlignAudioTranscript action, refer to the Media Server Reference.