The 80% volume is not a bug, it was designed that way. If you want to change the set level for now you can go into the integration options and change the TTS volume to your desired level. It’s still a fixed level across all the zones though. I think I will definitely change from this fixed TTS volume to a snapshot method to be more flexible going forward.
For the delay problem, can you tell me when the delay comes? Is it before the text is spoken or after? Does the system eventually return to normal? Do you happen to be changing the volume or other settings while the TTS call is processing?