Phonetic Phrase Search in a Single Step

Use the following procedure to perform phonetic phrase search in a single operation.

NOTE:

HPE recommends that you use this task for single searches only, because the procedure does not produce a phoneme time track file. If you use the WavPhraseSearch task to perform multiple searches on a file, HPE IDOL Speech Server must process the audio file each time. To perform multiple searches, HPE recommends that you process the audio file to produce a phoneme time track file, and then perform searches separately.

To perform phonetic phrase search in a single operation

  1. To search for more than one phrase, create a list that contains the phrases.

    For more information about HPE IDOL Speech Server's list manager, see Create and Manage Lists.

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

    Type The task name. Set to WavPhraseSearch.
    DnnFile

    The DNN acoustic model file to use for processing.

    NOTE:

    The use of DNNs improves both precision and recall, but increases processing time. If faster processing time is more important to you than accuracy, set DnnFile to none to disable DNN acoustic modeling.

    File

    The audio file to process.

    To restrict processing to a section of the audio file, set the start and end times in the wav module (for information about how to configure the wav module, see the HPE IDOL Speech Server Reference).

    Lang The phonetic phrase search language pack to use.
    Out The file to write the results to.

    Set one of the following parameters.

    Phrase The word or phrase to search for.
    PhraseList

    The list that specifies multiple phrases to search for.

For example:

http://localhost:13000/action=AddTask&Type=WavPhraseSearch&File=C:/Data/Speech.wav&DnnFile=ver-ENUK-tel-6.2-8k.dnn&Lang=ENUK-pm&Out=SearchResults.ctm&Phrase=financial

This action uses port 13000 to instruct HPE IDOL Speech Server, which is located on the local machine, to search the Speech.wav file for the phrase ‘financial’, , using the ver-ENUK-tel-6.2-8k.dnn DNN acoustic model file for processing, and to write the search results to the SearchResults.ctm file.

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


_HP_HTML5_bannerTitle.htm