Function createAudioMessage

createAudioMessage(chunk: AudioChunk, role?: AudioMessageRole): ModelMessage
Build the canonical audio ModelMessage from an AudioChunk.

The content is [textPart?, fileAudioPart]: a raw-PCM16 AI-SDK file part (mediaType: "audio/pcm16"), preceded by a transcript text part when the chunk carries one. Audio travels cleanly in any role — no forceUserRole.
Parameters
- chunk: AudioChunk
  The audio chunk to wrap.
- role: AudioMessageRole = "user"
  The message role. Defaults to "user".
Returns ModelMessage
- Defined in work/scenario/scenario/javascript/src/voice/messages.ts:99