languageBehavior

Defines how the transcription model detects the audio language. Default value is 'automatic single language'.