Hi everybody! Couple questions.

Question about Vmix call and audio feedback.

We set up a preset that lets us record a video podcast where both the host and the guest "call in" via vmix call. This goes to a third person who is operating vmix to record. They see and hear each other fine. That all works. The operator is able to record everything just fine. Everything in our tests had no issues with audio. Today when we recorded the "host" caller was using his Yeti USB mic (as selected when the vmix call started) and then used Apple airpods as his headphones. He could hear a delayed version of himself in his earphones. This doesn't happen using the same preset in our tests but we're using headsets with separate mic. (Im not sure what you call it, but essentially a "gaming headsetup" with headphones and a boom style mouth-piece microphone)
Is there an obvious error here? Does vmix call allow two microphones from one user? I thought maybe his airpods were picking up his voice or something. The feedback wasn't recorded in the audio track that we multicorded.

Secondly, we setup an additional audio input that's routed to the master that the vmix operator uses. Basically so that the "host" and "caller" can hear him if he needs to relay instructions. This works perfectly in our tests. But, again, but the actual host and caller called into vmix call they could NOT hear this operator input. The "audio" button was green and we could see levels etc.

Any thoughts?

Please be gentle I'm relatively new to vmix. Thanks!

#1. It looks like the "caller" was actually using his computer speakers. So does that mean the host had audio was flowing back through the callers mic, causing him to hear himself as a delay? Or are you referring to something else?

#2. Busses are definitely tricky for me. I have master, bus A and bus B. I have output 1 set to master which is a multiview. Output 2 to bus A which is the "host". and output 3 to bus B which is the caller". This setup has tested to work and lets me have the host and caller call in, and I record a multiview on output 1 with a mixed audio track, output 2 as the "host" with one side of the conversation and output 3 has the "caller" with the other side of the conversation. 100% worked in testing. Attached are some screengrabs for clarity.