[Dictation] Transcription issues. #50
Description
Hello,
I am trying to implement a voice chat feature using dictation experience. I have found that the string output from the partial transcription event is quite accurate. However, the string output from the full transcription event is missing about 70% of the initial part of the actual audio. This issue occurs very frequently. Given that the full transcription event is called at appropriate times, it doesn't seem to be related to the Keep Alive Time.
Is there a way to solve this issue without implementing incremental transcription manually?