I don't see this much different than what commonly happens with humans when we hear our named called when it was some environmental noise.

As for the silence, I wonder why the the model even receives it. I would think a lot of that would be compressed out of existence to save bandwidth.