were you able to solve this ? I am stuck in a similar situation where i can see the audio buffer in the backend getting transmitted to aws transcribe but getting empty results array.