An audio blob that contains the next part of the audio that you want to transcribe. The maximum audio chunk size is 32 KB.