Run a Data Obfuscation Task

This section describes how to prepare your data and use it to run a data obfuscation task.

To run a data obfuscation task

  1. Normalize the transcription files (see Run Text Normalization).
  2. Create a language model based on the normalized transcription files (see Build the Language Model).
  3. Run speech-to-text on the audio data, using the language model created in the previous step to optimize performance (see Speech-to-Text).
  4. Run the scorer task on the text, using the speech-to-text output produced in the previous step (see Run the Scorer). This process produces both a score file (which might indicate whether there are any issues with the transcript compared to what is actually said in the audio) and an aligned .ctm file. The .ctm file is used as input in the training process.
  5. Run audio analysis on all the files, to obtain audio quality information (see Run Audio Analysis). The audio analysis files are then passed to the dataObfuscation module.

  6. Send an AddTask action to HPE IDOL Speech Server, and set the following parameters:

    Type The task name.
    Am The acoustic model to use for processing.
    AudioAnalysisDir The location of the audio analysis output directory.
    DataList A list of the files to use for processing.
    DnnFile The DNN acoustic model file to use for processing.
    OtdPath The directory to write the obfuscated training data files to.
    Pgf The pronunciation generation (.pgf) file included in the language pack.
    PlhPath The path to the directory containing the acoustic feature (.plh) files specified in the DataList.
    TxtPath The path to the directory containing the transcript (.ctm) files specified in the DataList.

For example:

http://localhost:13000/action=AddTask&Type=DataObfuscation&Am=ver-ENUK-tel-6.2-8k.am&Pgf=ver-ENUK-tel-6.2.pgf&DnnFile=ver-ENUK-tel-6.2-8k.dnn&DataList=ListManager/ObfuscList&PlhPath=T:\data\PLH&TxtPath=T:\data\transcripts&OtdPath=T:\data\OTD&AudioAnalysisDir=C:\AudioAnalysis

This action uses port 13000 to instruct HPE IDOL Speech Server, which is located on the local machine, to perform data obfuscation on the acoustic feature files and transcript files that you specify in the ListManager/ObfuscList list, using the ver-ENUK-tel-6.2-8k.am acoustic model file, the ver-ENUK-tel-6.2.pgf file, and the ver-ENUK-tel-6.2-8k.dnn DNN file for processing. HPE IDOL Speech Server writes the obfuscated training files to the T:\data\OTD directory.

This action returns a token. You can use the token to:

Related Topics


_HP_HTML5_bannerTitle.htm