How to structure a SwiftUI Push To Talk App backend?

Question

I would like to build a Push To Talk mobile app on iOS platform using SwiftUI. I read the documentation for PTT framework and I think I understood the role of APNs. The big question mark for me is how I actually encode, send and then decode the audio in real-time. What do I need to implement? Does anyone have experience on that?

I manually configured and joined a channel, but I miss the part of sending the audio:

func channelManager(_ channelManager: PTChannelManager, channelUUID: UUID, didBeginTransmittingFrom source: PTChannelTransmitRequestSource) {        
    // Begin reconnecting to the app’s PTT services backend infrastructure
    // and signal that the user is beginning to transmit.    
}

// The audio session is in an active state and ready to use.    
func channelManager(_ channelManager: PTChannelManager, didActivate audioSession: AVAudioSession) {        
    // Configure the audio session and begin recording. 
    // HOW ??
}