Is there some specific configuration or setting required to enable the audio-based active speaker detection for video-active-change?
The video-active-change event excludes the current user, so in a two-person session, when this event is triggered, the userId in the payload will always be the other user.
If you need to include the current user, you can listen to the active-speaker event and take the first user in the payload; this event will include the current user.
Thanks
Vic