How to Extract Text from Audio Files

About the MSAudioToText Component

The MSAudioToText pipeline stage:

• Extracts the text and words with the time stamp from audio files

• Uses the Microsoft Azure Speech to Text library

How to Add the MSAudioToText to AutoClassifier

Use the following steps to add the MSAudioToText component to an AutoClassifier pipeline stage.

1. Navigate to the AutoClassifier Pipelines component page.
   
   
   
2. Click New Componentand select MS Audio To Text from the component list:
   
   
   
3. Name your new MSAudioToText component and click Add
   
   
   
4. Click Apply to save your changes.
5. Ensure your new MS Audio To Text component is placed in the list of existing pipeline stages.

How to Configure the MS Audio To Text Component

1. Open your MS Audio To Text component.
   
2. Subscription Key: Enter the key from Cognitive Services resource.
   
3. Region: Enter the region from Cognitive Services resource.
   
4. Source languages:
   1. Represents the audio file languages.
   2. The stage supports maximum 4 different languages in the same time.
   3. The languages must be added in Locale (BCP-47) format.
      1. Example: en-US;it-IT;fr-FR;de-DE;
         
5. Click Apply.

Output Property	Type
MSAudioToText	string
StartTimeWords	string - Multi