I use Loopback to pull in sound from other places so it can be muxed into the conference from my side. It would be nice for Spellar to support a Loopback input as well as a Mic input so it can distinguish between other audio and my voice. This is along the lines of what the "Exclude other people from the transcript" is about.