because audio is optional. you can read the trascribed notes if you have audio playback off. ideally the transcription is covered by the publishers signature, so they have a chance to edit it. you could in theory make them compatible by having the device do transcriptions on untranscribed notes that are incoming... but that would use a lot of cpu

But the event still contains the audio. It's only optional to the user. I told him this can all be fixed via client kung fu but he didn't want that.