Jigasi active speaker for transcription

I need to know who is the active speaker in Jigasi while passing the voice data to a custom ASR service.

But I couldn’t find the right place in the Jigasi codebase to add a listener to the DominantSpeakerEndpointChangeEvent.

Can anyone help, please?

@damencho

Would you please?

Please don’t mention people that are not already posting in same thread.

That information is received from the bridge in the client through the datachannel or websocket, unfortunately jigasi does not have that connection so it does not have and that information … the bridge uses the audio level rtp header to calculate that … So probably and jigasi can do the same … But I have no answer for how easy is that or how to do it … need to spend some time checking the code …

1 Like

Thank you.

but if the dominant speaker data passes through the bridge data channel, then how does the mod_speakerstats_component.lua work (in prosody plugins)?

I mean there should be another mirroring event, isn’t it?

Speaker stats receive an iq message from every participant when it becomes a dominant speaker …