Create a Speaker Template from Multiple Feature Files

A two-step process allows you to perform speaker training with multiple audio files:

  1. Create a set of Speaker ID feature files by using the SpkIdFeature task (see Create Speaker ID Feature Files for more information).
  2. Use these feature files to create the speaker template file.

Assuming you have already created a set of feature files, you can train the new speaker template from this set by using the SpkIdTrain task.

First you must create a list file that contains a list of all the feature files. You can do this manually, or by using the List Manager.

Each element of the list should be the file name relative to the Speaker ID directory, and must include the file extension.

For example:

BrownSpeech1.atv
BrownSpeech2.atv
BrownSpeech3.atv
BrownSpeech4.atv

Note: If the files are not stored in the Speaker ID directory, you can use the DataPath parameter to specify the location.

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

After you have created the list, send an AddTask action to IDOL Speech Server, and set the following parameters:

Type The task name. Set to SpkIdTrain.
DataList

A list file that lists the feature files to use.

Out The name of the speaker template file to create.

For example:

http://localhost:15000/action=AddTask&Type=SpkIdTrain&DataList=ListManager/BrownList&Out=Brown.atf

This action uses port 15000 to instruct IDOL Speech Server, which is located on the local machine, to create the Brown.atf template using the feature files listed in the BrownList list file.

You can set additional parameters. For details of the optional parameters, see the IDOL Speech Server Reference.

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


_HP_HTML5_bannerTitle.htm