AI-103: Developing AI Apps and Agents on AzureDemo

By Microsoft · Question Mode

QuestionQ104

Implement text analysis solutions

You are creating a text-to-speech solution using Azure Speech in Foundry Tools that reads instructions aloud from a script contained in a text file.

You notice that the solution frequently mispronounces technical terms.

You need to stop these incorrect pronunciations from occurring. The solution must minimize development effort.

What should you do?

  • A From Speech Studio, train a custom neural voice
  • B Use Speech Synthesis Markup Language (SSML) to specify phonemes.
  • C Use Speech Synthesis Markup Language (SSML) to apply say as rules.
  • D Use Speech Synthesis Markup Language (SSML) to adjust the prosody of the voice.
  • E From Azure OpenAI use the Whisper model.
Explanation

The SSML phoneme element allows you to specify the exact pronunciation of a word using a phonetic alphabet such as IPA or SAPI. This directly addresses mispronunciation of specific technical terms with minimal effort, since you only need to annotate the problematic words in the SSML markup rather than retrain a model. In contrast, say-as is meant for interpreting content types (like dates or numbers) rather than correcting phonetic pronunciation, prosody adjusts pitch/rate/volume rather than pronunciation, custom neural voice training requires substantial data collection and model training effort, and Whisper is a speech-to-text model, not applicable to a text-to-speech pronunciation issue.

Learn more

Community Discussion

No comments yet. Be the first to start the discussion!